Menu
EN

Vyhledávání

Lidé

prof. Ing. Jan Faigl, Ph.D.

Garant a tutor programu Otevřená informatika - bakalářský
Garant a tutor programu Otevřená informatika - magisterský

Všechny publikace

Autonomous Robotic Exploration with Simultaneous Environment and Traversability Models Learning

  • DOI: 10.3389/frobt.2022.910113
  • Odkaz: https://doi.org/10.3389/frobt.2022.910113
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this study, we address generalized autonomous mobile robot exploration of unknown environments where a robotic agent learns a traversability model and builds a spatial model of the environment. The agent can benefit from the model learned online in distinguishing what terrains are easy to traverse and which should be avoided. The proposed solution enables the learning of multiple traversability models, each associated with a particular locomotion gait, a walking pattern of a multi-legged walking robot. We propose to address the simultaneous learning of the environment and traversability models by a decoupled approach. Thus, navigation waypoints are generated using the current spatial and traversability models to gain the information necessary to improve the particular model during the robot's motion in the environment. From the set of possible waypoints, the decision on where to navigate next is made based on the solution of the generalized traveling salesman problem that allows taking into account a planning horizon longer than a single myopic decision. The proposed approach has been verified in simulated scenarios and experimental deployments with a real hexapod walking robot with two locomotion gaits, suitable for different terrains. Based on the achieved results, the proposed method exploits the online learned traversability models and further supports the selection of the most appropriate locomotion gait for the particular terrain types.

Bounding optimal headings in the Dubins Touring Problem

  • Autoři: Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the 37th ACM/SIGAPP Symposium on Applied Computing. New York: ACM, 2022. p. 770-773. ISBN 978-1-4503-8713-2.
  • Rok: 2022
  • DOI: 10.1145/3477314.3507350
  • Odkaz: https://doi.org/10.1145/3477314.3507350
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    The Dubins Touring Problem (DTP) stands to find the shortest curvature-constrained multi-goal path connecting a prescribed sequence of locations. The problem is to determine the optimal vehicle heading angle at each location and thus find the shortest sequence of Dubins paths. The heading angles can be determined by iterative refinement of possible heading intervals for which finer resolution yields a shorter path at the cost of increased computational requirements. In this paper, we introduce a novel method to bound the optimal heading angles by eliminating unpromising intervals that cannot contribute to the optimal solution. The method is employed in the branch-and-bound solution of the DTP, where it significantly reduces the search space in finding the optimal solution.

Continually trained life-long classification

  • DOI: 10.1007/s00521-021-06154-9
  • Odkaz: https://doi.org/10.1007/s00521-021-06154-9
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Two challenges can be found in a life-long classifier that learns continually: the concept drift, when the probability distribution of data is changing in time, and catastrophic forgetting when the earlier learned knowledge is lost. There are many proposed solutions to each challenge, but very little research is done to solve both challenges simultaneously. We show that both the concept drift and catastrophic forgetting are closely related to our proposed description of the life-long continual classification. We describe the process of continual learning as a wrap modification, where a wrap is a manifold that can be trained to cover or uncover a given set of samples. The notion of wraps and their cover/uncover modifiers are theoretical building blocks of a novel general life-long learning scheme, implemented as an ensemble of variational autoencoders. The proposed algorithm is examined on evaluation scenarios for continual learning and compared to state-of-the-art algorithms demonstrating the robustness to catastrophic forgetting and adaptability to concept drift but also showing the new challenges of the life-long classification.

Deep Transfer Learning of Traversability Assessment for Heterogeneous Robots

  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    For autonomous robots operating in an unknown environment, it is important to assess the traversability of the surrounding terrain to improve path planning and decision-making on where to navigate next in a cost-efficient way. Specifically, in mobile robot exploration, terrains and their traversability are unknown prior to the deployment. The robot needs to use its limited resources to learn its terrain traversability model on the go; however, reusing a provided model is still a desirable option. In a team of heterogeneous robots, the models assessing traversability cannot be reused directly since robots might possess different morphology or sensory equipment and thus experience the terrain differently. In this paper, we propose a transfer learning approach for convolutional neural networks assessing the traversability between heterogeneous robots, where the transferred network is retrained using data available for the target robot to accommodate itself to the robot’s traversability. The proposed method is verified in real-world experiments, where the proposed approach provides faster learning convergence and better traversal cost predictions than the baseline.

Gait Adaptation After Leg Amputation of Hexapod Walking Robot Without Sensory Feedback

  • DOI: 10.1007/978-3-031-15934-3_54
  • Odkaz: https://doi.org/10.1007/978-3-031-15934-3_54
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we address the adaptation of the locomotion controller to change of the multi-legged walking robot morphology, such as leg amputation. In nature, the animal compensates for the amputation using its neural locomotion controller that we aim to reproduce with the Central Pattern Generator (CPG). The CPG is a rhythm-generating recurrent neural network used in gait controllers for the rhythmical locomotion of walking robots. The locomotion corresponds to the robot's morphology, and therefore, the locomotion rhythm must adapt if the robot's morphology is changed. The leg amputation can be handled by sensory feedback to compensate for the load distribution imbalances. However, the sensory feedback can be disrupted due to unexpected external events causing the leg to be damaged, thus leading to unexpected motion states. Therefore, we propose dynamic rules for learning a new gait rhythm without the sensory feedback input. The method has been experimentally validated on a real hexapod walking robot to demonstrate its usability for gait adaptation after amputation of one or two legs.

Generating Safe Corridors Roadmap for Urban Air Mobility

  • Autoři: Ing. Jakub Sláma, Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Intelligent Robots and Systems (IROS), 2022 IEEE/RSJ International Conference on. Piscataway: IEEE, 2022. p. 11866-11871. ISSN 2153-0866. ISBN 978-1-6654-7927-1.
  • Rok: 2022
  • DOI: 10.1109/IROS47612.2022.9981326
  • Odkaz: https://doi.org/10.1109/IROS47612.2022.9981326
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Personal air transportation on short distances, so-called Urban Air Mobility (UAM), is a trend in modern aviation that raises new challenges as flying in urban areas at low altitudes induces an additional risk to people and properties on the ground. Risk-aware trajectory planning can mitigate the risk by detouring and flying over less populated and thus less risky areas. Existing risk-aware trajectory planning approaches are computationally demanding single-query methods that are impractical for online usage. Moreover, coordinated planning for multiple aircraft is prohibitively expensive. Therefore, we propose to reduce computational demands by determining low-risk areas called safe corridors and creating a roadmap of safe corridors based on multiple least risky trajectories. The created roadmap can be used in graph-based multi-agent planning methods for coordinated trajectory planning. The proposed method has been evaluated in a realistic urban scenario, suggesting a significant computational burden reduction and less risky trajectories than the current state-of-the-art methods.

GNG-based Clustering of Risk-aware Trajectories into Safe Corridors

  • Autoři: Ing. Jakub Sláma, Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Advances in Self-Organizing Maps, Learning Vector Quantization, Clustering and Data Visualization. Springer, Cham, 2022. p. 87-97. ISSN 2367-3370. ISBN 978-3-031-15443-0.
  • Rok: 2022
  • DOI: 10.1007/978-3-031-15444-7_9
  • Odkaz: https://doi.org/10.1007/978-3-031-15444-7_9
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Personal air transportation on short distances is a promising trend in modern aviation, raising new challenges as flying in low altitudes in highly populated environments induces additional risk to people and properties on the ground. Risk-aware planning can mitigate the risk by preferring flying above low-risk areas such as rivers or brownfields. Finding such trajectories is computationally demanding, but they can be precomputed for areas that are not changing rapidly and form a planning roadmap. The roadmap can be utilized for multi-query trajectory planning using graph-based search. However, a quality roadmap is required to provide a low-risk trajectory for an arbitrary query on a risk-aware trajectory from one location to another. Even though a dense roadmap can achieve the quality, it would be computationally demanding. Therefore, we propose to cluster the found trajectories and create a sparse roadmap of safe corridors that provide similar quality of risk-aware trajectories. In this paper, we report on applying Growing Neural Gas (GNG) in estimating the suitable number of clusters. Based on the empirical evaluation using a realistic urban scenario, the results suggest a significant reduction of the computational burden on risk-aware trajectory planning using the roadmap with the clustered safe corridors.

Learning-based Detection of Leg-Surface Contact using Position Feedback Only

  • DOI: 10.1109/ETFA52439.2022.9921720
  • Odkaz: https://doi.org/10.1109/ETFA52439.2022.9921720
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this work-in-progress report, we present experimental results of lightweight learning-based leg-contact detection methods for a small hexapod walking robot with position feed- back only. The detection of the leg contact with the surface is addressed as anomaly detection using predicted and measured positions of the leg’s joints in the leg swing phase. A polynomial regressor and three-layer neural network are evaluated regarding the prediction error and computational requirements using realistic datasets collected with the real hexapod walking robot.

T*ε—-Bounded-Suboptimal Efficient Motion Planning for Minimum-Time Planar Curvature-Constrained Systems

  • Autoři: Pinsky, D., Váňa, P., prof. Ing. Jan Faigl, Ph.D., Salzman, O.
  • Publikace: IEEE Robotics and Automation Letters. 2022, 7(2), 4102-4109. ISSN 2377-3766.
  • Rok: 2022
  • DOI: 10.1109/LRA.2022.3149307
  • Odkaz: https://doi.org/10.1109/LRA.2022.3149307
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    We consider the problem of finding collision-free paths for curvature-constrained systems in the presence of obstacles while minimizing execution time. Specifically, we focus on the setting where a planar system can travel at some range of speeds with unbounded acceleration. This setting can model many systems, such as fixed-wing drones. Unfortunately, planning for such systems might require evaluating many (local) time-optimal transitions connecting two close-by configurations, which is computationally expensive. Existing methods either pre-compute all such transitions in a preprocessing stage or use heuristics to speed up the search, thus foregoing any guarantees on solution quality. Our key insight is that computing all the time-optimal transitions is both (i) computationally expensive and (ii) unnecessary formany problem instances. We show that by finding bounded-suboptimal solutions (solutionswhose cost is bounded by 1 + epsilon times the cost of the optimal solution for any user-provided epsilon) and not time-optimal solutions, one can dramatically reduce the number of time-optimal transitions used. We demonstrate using empirical evaluation that our planning framework can reduce the runtime by several orders of magnitude compared to the state-of-the-art while still providing guarantees on the quality of the solution.

Terrain Traversal Cost Learning with Knowledge Transfer Between Multi-legged Walking Robot Gaits

  • DOI: 10.1109/ICARSC55462.2022.9784790
  • Odkaz: https://doi.org/10.1109/ICARSC55462.2022.9784790
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    The terrain traversal abilities of multi-legged walking robots are affected by gaits, the walking patterns that enable adaptation to various operational environments. Fast and lowset gaits are suited to flat ground, while cautious and highset gaits enable traversing rough areas. A suitable gait can be selected using prior experience with a particular terrain type. However, experience alone is insufficient in practical setups, where the robot experiences each terrain with only one or just a few gaits and thus would infer novel gait-terrain interactions from insufficient data. Therefore, we use knowledge transfer to address unsampled gait-terrain interactions and infer the traversal cost for every gait. The proposed solution combines gaitterrain cost models using inferred gait-to-gait models projecting the robot experiences between different gaits. We implement the cost models as Gaussian Mixture regressors providing certainty to identify unknown terrains where knowledge transfer is desirable. The presented method has been verified in synthetic showcase scenarios and deployment with a real walking robot. The proposed knowledge transfer demonstrates improved cost prediction and selection of the appropriate gait for specific terrains.

Traveling Salesman Problem with neighborhoods on a sphere in reflectance transformation imaging scenarios

  • DOI: 10.1016/j.eswa.2022.116814
  • Odkaz: https://doi.org/10.1016/j.eswa.2022.116814
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    In this paper, we propose a solution to the non-Euclidean variant of the Traveling Salesman Problem with Neighborhoods on a Sphere (TSPNS). The introduced problem formulation is motivated by practical scenarios of employing unmanned aerial vehicles in the Reflectance Transformation Imaging (RTI). In the RTI, a vehicle is requested to visit a set of sites at a constant distance from the object of interest and cast light from different directions to model the object from the images captured from another fixed location. Even though the problem can be formulated as an instance of the regular traveling salesman problem, we report a significant reduction of the solution cost by exploiting a non-zero tolerance on the light direction and defining the sites as regions on a sphere. The continuous neighborhoods of the TSPNS can be sampled into discrete sets, and the problem can be transformed into the generalized traveling salesman problem. However, finding high-quality solutions requires dense sampling, which increases the computational requirements. Therefore, we propose a practical heuristic solution based on the unsupervised learning of the Growing Self-Organizing Array (GSOA) that quickly finds an initial solution with the competitive quality to the sampling-based method. Furthermore, we propose a fast post-processing optimization to improve the initial solutions of both solvers. Based on the reported results, the proposed GSOA-based solver provides solutions of a similar quality to the transformation approach while it is about two orders of magnitude less computationally demanding.

Traversability Transfer Learning Between Robots with Different Cost Assessment Policies

  • DOI: 10.1007/978-3-030-98260-7_21
  • Odkaz: https://doi.org/10.1007/978-3-030-98260-7_21
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Predicting mobile robots' traversability over terrains is crucial to select safe and efficient paths through rough and unstructured environments. In multi-robot missions, knowledge transfer techniques can enable learning terrain traversability assessment the robots did not experience individually. The knowledge can be incrementally aggregated for homogeneous robots since they can treat foreign knowledge as their own. However, robots with different perceptions might experience the same terrain differently, so it is impossible to aggregate the shared knowledge directly. In this paper, we show how to learn a model that transfers the experience between heterogeneous robots, enabling each robot to use the whole sum of the experience of the multi-robot team. The proposed approach uses correlation to combine individual neural networks that assess the traversability of individual robots. The presented method has been verified in a real-world deployment of multi-legged walking robots with different cost assessment policies.

Vehicle Fault-Tolerant Robust Power Transmission Line Inspection Planning

  • DOI: 10.1109/ETFA52439.2022.9921692
  • Odkaz: https://doi.org/10.1109/ETFA52439.2022.9921692
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    This paper concerns fault-tolerant power transmission line inspection planning as a generalization of the multiple traveling salesmen problem. The addressed inspection planning problem is formulated as a single-depot multiple-vehicle scenario, where the inspection vehicles are constrained by the battery budget limiting their inspection time. The inspection vehicle is assumed to be an autonomous multi-copter with a wide range of possible flight speeds influencing battery consumption. The inspection plan is represented by multiple routes for vehicles providing full coverage over inspection target power lines. On an inspection vehicle mission interruption, which might happen at any time during the execution of the inspection plan, the inspection is re-planned using the remaining vehicles and their remaining battery budgets. Robustness is introduced by choosing a suitable cost function for the initial plan that maximizes the time window for successful re-planning. It enables the remaining vehicles to successfully finish all the inspection targets using their respective remaining battery budgets. A combinatorial metaheuristic algorithm with various cost functions is used for planning and fast re-planning during the inspection.

WiSM: Windowing Surrogate Model for Evaluation of Curvature-Constrained Tours With Dubins Vehicle

  • DOI: 10.1109/TCYB.2020.3000465
  • Odkaz: https://doi.org/10.1109/TCYB.2020.3000465
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Dubins tours represent a solution of the Dubins traveling salesman problem (DTSP) that is a variant of the optimization routing problem to determine a curvature-constrained shortest path to visit a set of locations such that the path is feasible for Dubins vehicle, which moves only forward and has a limited turning radius. The DTSP combines the NP-hard combinatorial optimization to determine the optimal sequence of visits to the locations, as in the regular TSP, with the continuous optimization of the heading angles at the locations, where the optimal heading values depend on the sequence of visits and vice versa. We address the computationally challenging DTSP by fast evaluation of the sequence of visits by the proposed windowing surrogate model (WiSM), which estimates the length of the optimal Dubins path connecting a sequence of locations in a Dubins tour. The estimation is sped up by a regression model trained using close to optimum solutions of small Dubins tours that are generalized for large-scale instances of the addressed DTSP utilizing the sliding-window technique and a cache for already computed results. The reported results support that the proposed WiSM enables fast convergence of a relatively simple evolutionary algorithm to high-quality solutions of the DTSP. We show that with an increasing number of locations, our algorithm scales significantly better than other state-of-the-art DTSP solvers.

Decentralized Topological Mapping for Multi-robot Autonomous Exploration under Low-bandwidth Communication

  • DOI: 10.1109/ECMR50962.2021.9568824
  • Odkaz: https://doi.org/10.1109/ECMR50962.2021.9568824
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    This paper concerns a mapping framework for multi-robot exploration of underground environments with only very limited communication available. We focus on multi-robot map building and coordination to explore large areas with real-time planning to long distances. The considered communication can broadcast only 100 B/s, and therefore, we propose coordination planning using two terrain models. The first model is a dense 3D map built by each robot individually to identify explorable places and generate detailed plans to avoid un-traversable areas. The second model is a global topological map built in a decentralized manner by exchanging tiny 12 B packets between the robots. The feasibility of the proposed approach has been verified in the real-world autonomous exploration mission and various multi-robot scenarios inspired by a virtual cave circuit of the DARPA Subterranean Challenge while adapting two different decentralized coordination strategies.

Design, Construction, and Rough-Terrain Locomotion Control of Novel Hexapod Walking Robot With Four Degrees of Freedom Per Leg

  • DOI: 10.1109/ACCESS.2021.3053492
  • Odkaz: https://doi.org/10.1109/ACCESS.2021.3053492
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Multi-legged walking robots are suitable platforms for unstructured and rough terrains because of their immense locomotion capabilities. These are, however, redeemed by more sophisticated control and energy-demanding motion in comparison to wheeled robots. Particularly, electrically actuated multi-legged walking robots suffer from the adverse ratio between the robot body weight and payload capacity. Moreover, the locomotion speed and endurance ratio is far from what can be achieved with wheeled robots. In this paper, we focus on six-legged walking robots with statically-stable gait. Based on the analysis of existing solutions, we propose a novel construction of the affordable electrically actuated robot with substantial improvements in its motion capabilities, locomotion speed, reliability, and endurance. The proposed design is implemented in a Hexapod Ant Robot (HAntR) that is accompanied by the developed locomotion control approach to improve its rough terrains negotiation capabilities by the active distribution of the robot weight to the legs in the stance phase. Properties of the robot have been experimentally verified in extensive deployments, and based on the experimental benchmarking of the built prototype, HAntR is capable of locomotion for over an hour with the payload of 85% of its weight, and its maximum crawled distance per one second is 87% of its nominal length. HAntR represents significant improvements not only regarding the robots with identical actuators but also in comparison to other existing platforms. Therefore, we consider HAntR represents a step further towards a wide range of future applications and deployments of six-legged walking robots.

Experimental Leg Inverse Dynamics Learning of Multi-legged Walking Robot

  • DOI: 10.1007/978-3-030-70740-8_10
  • Odkaz: https://doi.org/10.1007/978-3-030-70740-8_10
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    Rough terrain locomotion is a domain where multi-legged robots benefit from their relatively complex morphology compared to the wheeled or tracked robots. Efficient rough terrain locomotion requires the legged robot sense contacts with the terrain to adapt its behavior and cope with the terrain irregularities. Usage of inverse dynamics to estimate the leg state and detect the leg contacts with the terrain suffers from computational complexity. Furthermore, it requires a precise analytical model identification that does not cope with adverse changes of the leg parameters such as friction changes due to the joint wear, the increased weight of the leg due to the mud deposits, and possible leg morphology change due to damage. In this paper, we report the experimental study on the locomotion performance with machine learning-based inverse dynamics model learning. Experimental examining three different learning models show that a simplified model is sufficient for leg collision detection learning. Moreover, the learned model is faster for calculation and generalizes better than more complex models when the leg parameters change.

Finding 3D Dubins Paths with Pitch Angle Constraint Using Non-linear Optimization

  • DOI: 10.1109/ECMR50962.2021.9568787
  • Odkaz: https://doi.org/10.1109/ECMR50962.2021.9568787
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper presents a novel non-linear programming formulation to find the shortest 3D Dubins path with a limited pitch angle. Such a path is suitable for fix-wing aircraft because it satisfies both the minimum turning radius and pitch angle constraints, and thus it is a feasible and smooth path in the 3D space. The proposed method utilizes the existing decoupled approach as an initial solution and improves its quality by dividing the path into small segments with constant curvature. The proposed formulation encodes the path using the direction vectors that significantly reduce the needed optimization variables. Therefore, a path with 100 segments can be optimized in about one second using conventional computational resources. Although the decoupled paths are usually within 2 % from the lower bound, the proposed approach further reduces the gap by about 30 %.

Gait Genesis Through Emergent Ordering of RBF Neurons on Central Pattern Generator for Hexapod Walking Robot

  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    The neurally based gait controllers for multi-legged robots are designed to reproduce the plasticity observed in animal locomotion. In animals, gaits are regulated by Central Pattern Generator (CPG), a recurrent neural network producing rhythmical signals prescribing each leg’s action timing, leading to coordinated motion of multiple legs. The biomimetic CPG-RBF architecture, where leg motion timing is encoded by Radial Basis Function (RBF) neurons coupled with CPG, is used in recent gait controllers. However, the RBF neurons coupling is usually parameterized by the supervisor. Therefore, the RBF parameters get outdated when the CPG signal’s wave-form changes. We propose self-supervised dynamics for RBF parameters adapting to a given CPG and producing the required gait rhythm. The method orders the leg activity with respect to inter-leg coordination rules and maps the activity onto CPG states. The proposed dynamics produce rhythmic control for three different hexapod gaits and adapts to the CPG parametric changes.

Gait-free Planning for Hexapod Walking Robot

  • DOI: 10.1109/ECMR50962.2021.9568834
  • Odkaz: https://doi.org/10.1109/ECMR50962.2021.9568834
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    This paper presents a gait-free motion planning approach for quasi-static walking of hexapod walking robots on terrains with limited available footholds. The proposed approach avoids using a prescribed gait pattern allowing an arbitrary sequence of leg swings. Furthermore, it is allowed that some legs do not need to be placed on the terrain for an extended duration. The proposed method is based on a decomposition of the motion planning into: (i) finding a candidate sequence of stances and intermediate configurations representing plausible steps using a graph-search; and (ii) connecting the intermediate configurations by feasible paths satisfying the motion constraints of the walking robot. The individual one-step paths are determined using a Bézier curve-based parametrization that seems to be sufficient for the relatively simple paths of a single step, and the low-capacity parametrization yields natural-looking motion.

Multi-tour Set Traveling Salesman Problem in Planning Power Transmission Line Inspection

  • DOI: 10.1109/LRA.2021.3091695
  • Odkaz: https://doi.org/10.1109/LRA.2021.3091695
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    This paper concerns optimal power transmission line inspection formulated as a proposed generalization of the traveling salesman problem for a multi-route one-depot scenario. The problem is formulated for an inspection vehicle with a limited travel budget. Therefore, the solution can be composed of multiple runs to provide full coverage of the given power lines. Besides, the solution indicates how many vehicles can perform the inspection in a single run. The optimal solution of the problem is solved by the proposed Integer Linear Programming (ILP) formulation, which is, however, very computationally demanding. Therefore, the computational requirements are addressed by the combinatorial metaheuristic. The employed greedy randomized adaptive search procedure is significantly less demanding while providing competitive solutions and scales better with the problem size than the ILP-based approach. The proposed formulation and algorithms are demonstrated in a real-world scenario to inspect power line segments at the electrical substation.

On Building Communication Maps in Subterranean Environments

  • DOI: 10.1007/978-3-030-70740-8_2
  • Odkaz: https://doi.org/10.1007/978-3-030-70740-8_2
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    Communication is of crucial importance for coordinating a team of mobile robotic units. In environments such as underground tunnels, the propagation of wireless signals is affected by nontrivial physical phenomena. Hence, both modeling of the communication properties and the consequent task to estimate where communication is available becomes demanding. A communication map is a tool assessing the characteristic of communication between two arbitrary spatial coordinates. The existing approaches based on interpolation of a priori obtained spatial measurements do not provide precise extrapolation estimates for unvisited locations. Therefore, we propose to address the extrapolation of the signal strength by a position-independent model based on approximating the obstacle occupancy ratio between the signal source and receiver. The proposed approach is compared to the existing attenuation models based on free-space path loss and spatial projection using a natural cave dataset. Based on the reported results, the proposed approach provides more accurate predictions than the existing approaches.

Risk-aware Trajectory Planning in Urban Environments with Safe Emergency Landing Guarantee

  • Autoři: Ing. Jakub Sláma, Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 2021 IEEE 17th International Conference on Automation Science and Engineering (CASE). Vienna: IEEE Industrial Electronic Society, 2021. p. 1606-1612. ISSN 2161-8089. ISBN 978-1-6654-1873-7.
  • Rok: 2021
  • DOI: 10.1109/CASE49439.2021.9551407
  • Odkaz: https://doi.org/10.1109/CASE49439.2021.9551407
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In-flight aircraft failures are never avoidable entirely, inducing a significant risk to people and properties on the ground in an urban environment. Existing risk-aware trajectory planning approaches minimize the risk by determining trajectories that might result in less damage in the case of failure. However, the risk of the loss of thrust can be eliminated by executing a safe emergency landing if a landing site is reachable. Therefore, we propose a novel risk-aware trajectory planning that minimizes the risk to people on the ground while an option of a safe emergency landing in the case of loss of thrust is guaranteed. The proposed method has been empirically evaluated on a realistic urban scenario. Based on the reported results, an improvement in the risk reduction is achieved compared to the shortest and risk-aware only trajectory. The proposed risk-aware planning with safe emergency landing seems to be suitable trajectory planning for urban air mobility.

Self-Learning Event Mistiming Detector Based on Central Pattern Generator

  • DOI: 10.3389/fnbot.2021.629652
  • Odkaz: https://doi.org/10.3389/fnbot.2021.629652
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    A repetitive movement pattern of many animals, a gait, is controlled by the Central Pattern Generator (CPG), providing rhythmic control synchronous to the sensed environment. As a rhythmic signal generator, the CPG can control the motion phase of biomimetic legged robots without feedback. The CPG can also act in sensory synchronization, where it can be utilized as a sensory phase estimator. Direct use of the CPG as the estimator is not common, and there is little research done on its utilization in the phase estimation. Generally, the sensory estimation augments the sensory feedback information, and motion irregularities can reveal from comparing measurements with the estimation. In this work, we study the CPG in the context of phase irregularity detection, where the timing of sensory events is disturbed. We propose a novel self-supervised method for learning mistiming detection, where the neural detector is trained by dynamic Hebbian-like rules during the robot walking. The proposed detector is composed of three neural components: (i) the CPG providing phase estimation, (ii) Radial Basis Function neuron anticipating the sensory event, and (iii) Leaky Integrate-and-Fire neuron detecting the sensory mistiming. The detector is integrated with the CPG-based gait controller. The mistiming detection triggers two reflexes: the elevator reflex, which avoids an obstacle, and the search reflex, which grasps a missing foothold. The proposed controller is deployed and trained on a hexapod walking robot to demonstrate the mistiming detection in real locomotion. The trained system has been examined in the controlled laboratory experiment and real field deployment in the Bull Rock cave system, where the robot utilized mistiming detection to negotiate the unstructured and slippery subterranean environment.

Variable-Speed Traveling Salesman Problem for Vehicles with Curvature Constrained Trajectories

  • Autoři: Kučerová, K., Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Piscataway: IEEE, 2021. p. 4714-4719. ISSN 2153-0866. ISBN 978-1-6654-1714-3.
  • Rok: 2021
  • DOI: 10.1109/IROS51168.2021.9636762
  • Odkaz: https://doi.org/10.1109/IROS51168.2021.9636762
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper presents a novel approach to the multigoal trajectory planning for vehicles with curvature-constrained trajectories such as fixed-wing aircraft. In the existing formulation called the Dubins Traveling Salesman Problem (DTSP), the vehicle speed is assumed to be constant over the whole trajectory, and that does not allow adaptation of the turning radius of the trajectory between the target locations. It does not support optimization of the overall flight time of the multi-goal trajectory by exploiting higher speeds for longer turning radii. Therefore, we propose a novel problem formulation called the Variable-Speed Traveling Salesman Problem (VS-TSP) that employs time-efficient trajectories with variable speed based on a generalization of the Dubins vehicle model, allowing multiple turning radii and change of the forward speed of the vehicle. The VS-TSP allows the vehicle to slow down if high maneuverability is necessary and speed up if high-speed turns with a large radius are beneficial to the overall tour cost. Based on the evaluation results for Cessna 172 aircraft model, the proposed VNS-based algorithm with variable speed provides up to about 20 % faster trajectories than a solution of the DTSP with a single turning radius.

Vision-Based Localization for Multi-rotor Aerial Vehicle in Outdoor Scenarios

  • DOI: 10.1007/978-3-030-70740-8_14
  • Odkaz: https://doi.org/10.1007/978-3-030-70740-8_14
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we report on the experimental evaluation of the embedded visual localization system, the Intel RealSense T265, deployed on a multi-rotor unmanned aerial vehicle. The performed evaluation is targeted to examine the limits of the localization system and discover its weak points. The system has been deployed in outdoor rural scenarios at altitudes up to 20 m. The Absolute trajectory error measures the accuracy of the localization with the reference provided by the differential GPS with centimeter precision. Besides, the localization performance is compared to the state-of-the-art feature-based visual localization ORB-SLAM2 utilizing the Intel RealSense D435 depth camera. In both types of experimental scenarios, with the teleoperated and autonomous vehicle, the identified weak point of the system is a translation drift. However, taking into account all experimental trials, both examined localization systems provide competitive results.

Aerial Reconnaissance and Ground Robot Terrain Learning in Traversal Cost Assessment

  • Autoři: Ing. Miloš Prágr, Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 6th International Workshop on Modelling and Simulation for Autonomous Systems. Wien: Springer, 2020. p. 3-10. ISSN 1611-3349. ISBN 9783030438890.
  • Rok: 2020
  • DOI: 10.1007/978-3-030-43890-6_1
  • Odkaz: https://doi.org/10.1007/978-3-030-43890-6_1
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we report on the developed system for assessment of ground unit terrain traversal cost using aerial reconnaissance of the expected mission environment. The system combines an aerial vehicle with ground robot terrain learning in the traversal cost modeling utilized in the mission planning for ground units. The aerial vehicle is deployed to capture visual data used to build a terrain model that is then used for the extraction of the terrain features of the expected operational area of the ground units. Based on the previous traversal experience of the ground units in similar environments, the learned model of the traversal cost is employed to predict the traversal cost of the new expected operational area to plan a cost-efficient path to visit the desired locations of interest. The particular modules of the system are demonstrated in an experimental scenario combining the deployment of an unmanned aerial vehicle with a multi-legged walking robot used for learning the traversal cost model.

DARPA Subterranean Challenge: Multi-robotic exploration of underground environments

  • DOI: 10.1007/978-3-030-43890-6_22
  • Odkaz: https://doi.org/10.1007/978-3-030-43890-6_22
  • Pracoviště: Centrum umělé inteligence, Vidění pro roboty a autonomní systémy, Multirobotické systémy
  • Anotace:
    The Subterranean Challenge (SubT) is a contest organised by the Defense Advanced Research Projects Agency (DARPA). The contest reflects the requirement of increasing safety and efficiency of underground search-and-rescue missions. In the SubT challenge, teams of mobile robots have to detect, localise and report positions of specific objects in an underground environment. This paper provides a description of the multi-robot heterogeneous exploration system of our CTU-CRAS team, which scored third place in the Tunnel Circuit round, surpassing the performance of all other non-DARPA-funded competitors. In addition to the description of the platforms, algorithms and strategies used, we also discuss the lessons-learned by participating at such contest.

Fast Sequence Rejection for Multi-Goal Planning with Dubins Vehicle

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Váňa, P., Ing. Jan Drchal, Ph.D.,
  • Publikace: 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE Robotics and Automation Society, 2020. p. 6773-6780. ISSN 2153-0866. ISBN 978-1-7281-6212-6.
  • Rok: 2020
  • DOI: 10.1109/IROS45743.2020.9340644
  • Odkaz: https://doi.org/10.1109/IROS45743.2020.9340644
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Multi-goal curvature-constrained planning such as the Dubins Traveling Salesman Problem (DTSP) combines NP-hard combinatorial routing with continuous optimization to determine the optimal vehicle heading angle for each target location. The problem can be addressed as combinatorial routing using a finite set of heading samples at target locations. In such a case, optimal heading samples can be determined for a sequence of targets in polynomial time, and the DTSP can be solved as searching for a sequence with the minimal cost. However, the examination of sequences can be computationally demanding for high numbers of heading samples and target locations. A fast rejection schema is proposed to quickly examine unfavorable sequences using lower bound estimation of Dubins tour cost based on windowing technique that evaluates short subtours of the sequences. Furthermore, the computation using small problem instances can benefit from reusing stored results and thus speed up the search. The reported results indicate that the computational burden is decreased about two orders of magnitude, and the proposed approach supports finding high-quality solutions of routing problems with Dubins vehicle.

Greedy Randomized Adaptive Search Procedure for Close Enough Orienteering Problem

  • DOI: 10.1145/3341105.3374010
  • Odkaz: https://doi.org/10.1145/3341105.3374010
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we address the Close Enough Orienteering Problem (CEOP) that is motivated to find the most rewarding route visiting disk-shaped regions under the given travel budget. The CEOP differs from the regular OP in the continuous optimization of finding the most suitable waypoint locations to collect the reward associated with each region of interest in addition to the selection of the subset of the regions and sequence of their visits as in the OP. We propose to employ the Greedy Randomized Adaptive Search Procedure (GRASP) combinatorial metaheuristic to solve the addressed CEOP, in particular, the GRASP with Segment Remove. The continuous optimization is addressed by the newly introduced heuristic search that is applied in the construction phase and also in the local search phase of the GRASP. The proposed approach has been empirically evaluated using existing benchmarks, and based on the reported comparison with existing algorithms, the proposed GRASP-based approach provides solutions with the competitive quality while its computational requirements are low.

Handheld Localization Device for Indoor Environments

  • Autoři: Ing. Jan Bayer, prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the 2020 4 th International Conference on Automation, Control and Robots. Piscataway: IEEE Service Center, 2020. p. 60-64. ISBN 978-1-7281-9207-9.
  • Rok: 2020
  • DOI: 10.1109/ICACR51161.2020.9265494
  • Odkaz: https://doi.org/10.1109/ICACR51161.2020.9265494
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we address the problem of human safety during the search of underground or unknown indoor environments in scenarios where it is required to share the information about the position of the personnel such as first responders or speleologists. We report on the developed localization device to estimate and transmit the user position and state through a custom communication system. The information provided by the device serves to monitor the user state, enable communication between the user and the base station located outside the environment being searched. The personnel's position in the environment is thus provided to the mission supervisor in the case of an emergency. The device is based on low-cost off-the-shelf cameras for vision-based localization and low-bandwidth communication modules suitable for long-range communication in underground environments. Furthermore, the communication range is extendable by affordable transceivers that allow the dynamical building of independent communication networks in the operational environment. The performance of the communication system has been examined in an experimental scenario where five transceivers sufficiently cover 280 m of cave tunnels. The system can localize a human independently on any infrastructure supported by a report on the experimental deployment in an urban scenario with both indoor and outdoor areas.

Hopfield Neural Network in Solution of the Close Enough Orienteering Problem

  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we report on the Hopfield Neural Network (HNN) for the Orienteering Problem (OP) that is generalized to solve instances of the Close Enough Orienteering Problem (CEOP). In the orienteering problems, we are searching for a limited budget tour to maximize collected rewards by visiting selected target locations. In the CEOP, it is allowed to collect the reward remotely within a non-zero communication range. Thus we can save travel costs by collecting rewards at suitable visiting locations of the disk-shaped neighborhoods of target locations. The proposed approach combines the HNN for the OP with the Second-Order Cone Programming (SOCP) that is employed to determine locally optimal locations of visits to the disk-shaped neighborhoods of the target locations. Regarding the reported evaluation results using standard benchmarks, the proposed SOCP-based approach provides solutions with the improved solution quality compared to the previous HNN-based method with discrete samples of the possible locations of visits.

Minimal 3D Dubins Path with Bounded Curvature and Pitch Angle

  • Autoři: Váňa, P., Alves Neto, A., prof. Ing. Jan Faigl, Ph.D., MacHaret, D.G.
  • Publikace: IEEE International Conference on Robotics and Automation (ICRA). IEEE Xplore, 2020. p. 8497-8503. ISSN 2577-087X. ISBN 978-1-7281-7395-5.
  • Rok: 2020
  • DOI: 10.1109/ICRA40945.2020.9197084
  • Odkaz: https://doi.org/10.1109/ICRA40945.2020.9197084
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we address the problem of finding cost-efficient three-dimensional paths that satisfy the maximum allowed curvature and the pitch angle of the vehicle. For any given initial and final configurations, the problem is decoupled into finding the horizontal and vertical parts of the path separately. Although the individual paths are modeled as two-dimensional Dubins curves using closed-form solutions, the final 3D path is constructed using the proposed local optimization to find a cost-efficient solution. Moreover, based on the decoupled approach, we provide a lower bound estimation of the optimal path that enables us to determine the quality of the found heuristic solution. The proposed solution has been evaluated using existing benchmark instances and compared with state-of-the-art approaches. Based on the reported results and lower bounds, the proposed approach provides paths close to the optimal solution while the computational requirements are in hundreds of microseconds. Besides, the proposed method provides paths with fewer turns than others, which make them easier to be followed by the vehicle's controller.

Natural Criteria for Comparison of Pedestrian Flow Forecasting Models

  • Autoři: Vintr, T., Yan, Z., Eyisoy, K., Kubiš, F., Ing. Jan Blaha, Ing. Jiří Ulrich, Swaminathan, C., Molina, S., Kucner, T.P., Magnusson, M., Cielniak, G., prof. Ing. Jan Faigl, Ph.D., Duckett, T., Lilienthal, A.J., doc. Ing. Tomáš Krajník, Ph.D.,
  • Publikace: 2020 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE Robotics and Automation Society, 2020. p. 11197-11204. ISSN 2153-0866. ISBN 978-1-7281-6212-6.
  • Rok: 2020
  • DOI: 10.1109/IROS45743.2020.9341672
  • Odkaz: https://doi.org/10.1109/IROS45743.2020.9341672
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Models of human behaviour, such as pedestrian flows, are beneficial for safe and efficient operation of mobile robots. We present a new methodology for benchmarking of pedestrian flow models based on the afforded safety of robot navigation in human-populated environments. While previous evaluations of pedestrian flow models focused on their predictive capabilities, we assess their ability to support safe path planning and scheduling. Using real-world datasets gathered continuously over several weeks, we benchmark state-of-the-art pedestrian flow models, including both time-averaged and time-sensitive models. In the evaluation, we use the learned models to plan robot trajectories and then observe the number of times when the robot gets too close to humans, using a predefined social distance threshold. The experiments show that while traditional evaluation criteria based on model fidelity differ only marginally, the introduced criteria vary significantly depending on the model used, providing a natural interpretation of the expected safety of the system. For the time-averaged flow models, the number of encounters increases linearly with the percentage operating time of the robot, as might be reasonably expected. By contrast, for the time-sensitive models, the number of encounters grows sublinearly with the percentage operating time, by planning to avoid congested areas and times.

Neurodynamic Sensory-Motor Phase Binding for Multi-Legged Walking Robots

  • DOI: 10.1109/IJCNN48605.2020.9207507
  • Odkaz: https://doi.org/10.1109/IJCNN48605.2020.9207507
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    Motivated by observations of animal behavior, locomotion of multi-legged walking robots can be controlled by the central pattern generators (CPGs) that produce a repetitive motion pattern. A rhythmic pattern, a gait, is defined by phase relations between all leg joints. In a case of an external influence such as terrain irregularity, some actuator phase can shift and thus disrupt the phase relations between the actuators. The actuator phase relations can be maintained only by synchronizing to the sensors, which output can indicate the motion disruption. However, establishing correct sensory-motor phase relations requires not only the motor phase model but also a model of the sensory phase, which is generally unknown. Although both sensory and motor phases can be modeled by single CPG, the capabilities of such CPG-based controllers are limited because they are not flexible and robust. In this paper, we propose to model the phases of each sensor and motor by separate CPGs. The phase relations between the sensor and motor phases are established by radial basis function (RBF) neurons learned with proposed periodic Grossberg rule for which we present the convergence proof. Based on the reported evaluation results using high-fidelity simulation, the proposed locomotion controller demonstrates the desired plasticity, and it is capable of learning multiple gaits with robust synchronization to terrain changes using sensor inputs.

On finding time-efficient trajectories for fixed-wing aircraft using dubins paths with multiple radii

  • Autoři: Kučerová, K., Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the ACM Symposium on Applied Computing. New York: Association for Computing Machinery, 2020. p. 829-831. ISBN 978-1-4503-6866-7.
  • Rok: 2020
  • DOI: 10.1145/3341105.3374112
  • Odkaz: https://doi.org/10.1145/3341105.3374112
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Trajectory generation for fixed-wing aircraft can be based on Dubins vehicle model that constrains the vehicle to move forward with a limited turning radius and a constant speed. However, such a vehicle model cannot benefit from the physical capabilities of a real fixed-wing aircraft that can adjust its speed. Therefore, we propose to address the limitation of Dubins vehicle model by a generalized model that combines various turning radii, and thus allows increasing the cruise speed whenever possible. The proposed method provides faster trajectories in comparison to the trajectory generated by Dubins vehicle with a single turning radius and a constant cruise speed. The benefit of the proposed method is demonstrated on point-to-point trajectories, for which the parameters are inspired by Cessna 172 aircraft.

Optimal solution of the Generalized Dubins Interval Problem: finding the shortest curvature-constrained path through a set of regions

  • DOI: 10.1007/s10514-020-09932-x
  • Odkaz: https://doi.org/10.1007/s10514-020-09932-x
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    The Generalized Dubins Interval Problem (GDIP) stands to determine the minimal length path connecting two disk-shaped regions where the departure and terminal headings of Dubins vehicle are within the specified angle intervals. The GDIP is a generalization of the existing point-to-point planning problem for Dubins vehicle with a single heading angle per particular location that can be solved optimally using closed-form expression. For the GDIP, both the heading angles and locations need to be chosen from continuous sets which makes the problem challenging because of infinite possibilities how to connect the regions by Dubins path. We provide the optimal solution of the introduced GDIP based on detailed problem analysis. Moreover, we propose to employ the GDIP to provide the first tight lower bound for the Dubins Touring Regions Problem which stands to find the shortest curvature-constrained path through a set of regions in the prescribed order.

Speeded Up Elevation Map for Exploration of Large-Scale Subterranean Environments

  • Autoři: Ing. Jan Bayer, prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 6th International Workshop on Modelling and Simulation for Autonomous Systems. Wien: Springer, 2020. p. 190-202. ISSN 1611-3349. ISBN 9783030438890.
  • Rok: 2020
  • DOI: 10.1007/978-3-030-43890-6_15
  • Odkaz: https://doi.org/10.1007/978-3-030-43890-6_15
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we address a problem of the exploration of large-scale subterranean environments using autonomous ground mobile robots. In particular, we focus on an efficient data representation of the large-scale elevation map, where it is desirable to capture the shape of the terrain to avoid areas not traversable by a robot. Subterranean environments such as mine tunnel systems can be in units of kilometers large, but only a relatively small portion of the environment represents observable parts. Therefore, uniform grid-based elevation maps with resolution in units of centimeters are not memory efficient, and more suitable are hierarchical tree-based structures. However, hierarchical structures suffer from the increased computational requirements of accessing particular grid cells needed in determination of the navigational goals or evaluation of the terrain traversability in planning safe and cost-efficient paths. We propose a speed-up technique to combine the benefits of uniform grid-based and tree-based representations. The proposed elevation map representation keeps the memory footprint low using tree structure but enables fast access to the grid cells corresponding to the robot surroundings. The efficiency of the proposed data representation is demonstrated in an experimental deployment of the autonomous exploration of outdoor and subterranean environments.

Surveillance Planning with Safe Emergency Landing Guarantee for Fixed-wing Aircraft

  • DOI: 10.1016/j.robot.2020.103644
  • Odkaz: https://doi.org/10.1016/j.robot.2020.103644
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we study Emergency Landing Aware Surveillance Planning (ELASP) to determine a cost-efficient trajectory to visit a given set of target locations such that a safe emergency landing is possible at any point of the multi-goal trajectory. The problem is motivated to guarantee a safe mission plan in a case of loss of thrust for which it is desirable to have a safe gliding trajectory to a nearby airport. The problem combines computational challenges of the combinatorial multi-goal planning with demanding motion planning to determine safe landing trajectories for the curvature-constrained aerial vehicle. The crucial property of safe landing is a minimum safe altitude of the vehicle that can be found by trajectory planning to nearby airports using sampling-based motion planning such as RRT*. A trajectory is considered safe if the vehicle is at least at the minimum safe altitude at any point of the trajectory. Thus, a huge number of samples have to be evaluated to guarantee the safety of the trajectory, and an evaluation of all possible multi-goal trajectories is quickly computationally intractable. Therefore, we propose to utilize a roadmap of safe altitudes combined with the estimation of the trajectory lengths to evaluate only the most promising candidate trajectories. Based on the reported results, the proposed approach significantly reduces the computational burden and enables a solution of ELASP instances with tens of locations in units of minutes using standard single-core computational resources.

Terrain Learning Using Time Series of Ground Unit Traversal Cost

  • Autoři: Ing. Miloš Prágr, prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 6th International Workshop on Modelling and Simulation for Autonomous Systems. Wien: Springer, 2020. p. 97-107. ISSN 1611-3349. ISBN 9783030438890.
  • Rok: 2020
  • DOI: 10.1007/978-3-030-43890-6_8
  • Odkaz: https://doi.org/10.1007/978-3-030-43890-6_8
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we concern learning of terrain types based on the traversal experience observed by a hexapod walking robot. The addressed problem is motivated by the navigation of unmanned ground vehicles in long-term autonomous missions in a priory unknown environments such as extraterrestrial exploration. In such deployments, the robotic vehicle needs to learn hard to traverse terrains to improve its autonomous performance and avoid possibly dangerous areas. We propose to utilize Growing Neural Gas for terrain learning to capture the robot experience with traversing the terrain and thus learn a classifier of individual terrain types. The classifier is learned using a real time-series dataset collected by a hexapod walking robot traversing various terrain types. The learned model can be utilized to predict the traversal cost of newly observed terrains to support decisions on where to navigate next.

Transfer of Inter-Robotic Inductive Classifier

  • DOI: 10.1109/ICACR51161.2020.9265509
  • Odkaz: https://doi.org/10.1109/ICACR51161.2020.9265509
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In multi-robot deployments, the robots need to share and integrate their own experience and perform transfer learning. Under the assumption that the robots have the same morphology and carry equivalent sensory equipment, the problem of transfer learning can be considered incremental learning. Thus, the transfer learning problem inherits the challenges of incremental learning, such as catastrophic forgetting and concept drift. In catastrophic forgetting, the model abruptly forgets the previously learned knowledge during the learning process. The concept drift arises with different experiences between consecutively sampled models. However, state-of-the-art robotic transfer learning approaches do not address both challenges at once. In this paper, we propose to use an incremental classifier on a transfer learning problem. The feasibility of the proposed approach is demonstrated in a real deployment. The robot consistently merges two classifiers learned on two different tasks into a classifier that performs well on both tasks.

Unsupervised learning-based solution of the Close Enough Dubins Orienteering Problem

  • Autoři: prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Neural Computing and Applications. 2020, 24(32), 18193-18211. ISSN 0941-0643.
  • Rok: 2020
  • DOI: 10.1007/s00521-019-04222-9
  • Odkaz: https://doi.org/10.1007/s00521-019-04222-9
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper reports on the application of novel unsupervised learning-based method called the Growing Self-Organizing Array (GSOA) to data collection planning with curvature-constrained paths that is motivated by surveillance missions with aerial vehicles. The planning problem is formulated as the Close Enough Dubins Orienteering Problem which combines combinatorial optimization with continuous optimization to determine the most rewarding data collection path that does not exceed the given travel budget and satisfies the motion constraints of the vehicle. The combinatorial optimization consists of selecting a subset of the most rewarding data to be collected and the schedule of data collection. On the other hand, the continuous optimization stands to determine the most suitable waypoint locations from which selected data can be collected together with the determination of the headings at the waypoints for the used Dubins vehicle model. The existing purely combinatorial approaches need to discretize the possible waypoint locations and headings into some finite sets, and the solution is computationally very demanding because the problem size is quickly increased. On the contrary, the employed GSOA performs online sampling of the waypoints and headings during the adaptation of the growing structure that represents the requested curvature-constrained data collection path. Regarding the presented results, the proposed approach provides solutions to orienteering problems with competitive quality, but it is significantly less computationally demanding.

Adaptive locomotion control of hexapod walking robot for traversing rough terrains with position feedback only

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Čížek, P.
  • Publikace: Robotics and Autonomous Systems. 2019, 116 136-147. ISSN 0921-8890.
  • Rok: 2019
  • DOI: 10.1016/j.robot.2019.03.008
  • Odkaz: https://doi.org/10.1016/j.robot.2019.03.008
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Traversing rough terrains is one of the domains where multi-legged walking robots benefit from their relatively more complex kinematics in comparison to wheeled robots. The complexity of walking robots is usually related not only to mechanical parts but also to servomotors and the necessary electronics to efficiently control such a robotic system. Therefore, large, middle, but even small walking robots capable of traversing rough terrains can be very costly because of all the required equipment. On the other hand, using intelligent servomotors with the position control and feedback, affordable hexapod walking robots are becoming increasingly available. However, additional sensors may still be needed to stabilize the robot motion on rough terrains, e.g., inclinometers or inertial measurement units, force or tactile sensors to detect the ground contact point of the leg foot-tip. In this work, we present a minimalistic approach for adaptive locomotion control using only the servomotors position feedback. Adaptive fine-tuning of the proposed controller is supported by a dynamic model of the robot leg accompanied by the model of the internal servomotor controller. The models enable timely detection of the leg contact point with the ground and reduce developed stress and torques applied to the robot construction and servomotors without any additional sensor feedback. The presented results support that the proposed approach reliably detects the ground contact point, and thus enable traversing rough terrains with small, affordable hexapod walking robot.

Analysis of Using Mixed Reality Simulations for Incremental Development of Multi-UAV Systems

  • DOI: 10.1007/s10846-018-0875-8
  • Odkaz: https://doi.org/10.1007/s10846-018-0875-8
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Developing complex robotic systems requires expensive and time-consuming verification and testing which, especially in a case of multi-robot unmanned aerial systems (UASs), aggregates risk of hardware failures and may pose legal issues in experiments where operating more than one unmanned aircraft simultaneously is required. Thus, it is highly favorable to find and resolve most of the eventual design flaws and system bugs in a simulation, where their impacts are significantly lower. On the other hand, as the system development process approaches the final stages, the fidelity of the simulation needs to rise. However, since some phenomena that can significantly influence the system behavior are difficult to be modeled precisely, a partial embodiment of the simulation in the physical world is necessary. In this paper, we present a method for incremental development of complex unmanned aerial systems with the help of mixed reality simulations. The presented methodology is accompanied with a cost analysis to further show its benefits. The generality and versatility of the method is demonstrated in three practical use cases of various aviation systems development: (i) an unmanned system consisting of heterogeneous team of autonomous unmanned aircraft; (ii) a system for verification of collision avoidance methods among fixed wing unmanned aerial vehicles; and (iii) a system for planning collision-free paths for light-sport aircraft.

Autoencoders Covering Space as a Life-Long Classifier

  • DOI: 10.1007/978-3-030-19642-4_27
  • Odkaz: https://doi.org/10.1007/978-3-030-19642-4_27
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    A life-long classifier that learns incrementally has many challenges such as concept drift, when the class changes in time, and catastrophic forgetting when the earlier learned knowledge is lost. Many successful connectionist solutions are based on an idea that new data are learned only in a part of a network that is relevant to the new data. We leverage this idea and propose a novel method for learning an ensemble of specialized autoencoders. We interpret autoencoders as manifolds that can be trained to contain or exclude given points from the input space. This manifold manipulation allows us to implement a classifier that can suppress catastrophic forgetting and adapt to concept drift. The proposed algorithm is evaluated on an incremental version of the XOR problem and on an incremental version of the MNIST classification where we achieved 0.9 accuracy which is a significant improvement over the previously published results

Basic Evaluation Scenarios for Incrementally Trained Classifiers

  • DOI: 10.1007/978-3-030-30484-3_41
  • Odkaz: https://doi.org/10.1007/978-3-030-30484-3_41
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Evaluation of incremental classification algorithms is a complex task because there are many aspects to evaluate. Besides the aspects such as accuracy and generalization that are usually evaluated in the context of classification, we also need to assess how the algorithm handles two main challenges of the incremental learning: the concept drift and the catastrophic forgetting. However, only catastrophic forgetting is evaluated by the current methodology, where the classifier is evaluated in two scenarios for class addition and expansion. We generalize the methodology by proposing two new scenarios of incremental learning for class inclusion and separation that evaluate the handling of the concept drift. We demonstrate the proposed methodology on the evaluation of three different incremental classifiers, where we show that the proposed methodology provides a more complete and finer evaluation.

Benchmarking Incremental Regressors in Traversal Cost Assessment

  • Autoři: Ing. Miloš Prágr, prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Artificial Neural Networks and Machine Learning – ICANN 2019: Theoretical Neural Computation. ICANN 2019. Basel: Springer, 2019. p. 685-697. Lecture Notes in Computer Science. vol. 11727. ISSN 0302-9743. ISBN 978-3-030-30486-7.
  • Rok: 2019
  • DOI: 10.1007/978-3-030-30487-4_52
  • Odkaz: https://doi.org/10.1007/978-3-030-30487-4_52
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    Motivated by the deployment of multi-legged walking robots in traversing various terrain types, we benchmark existing online and unsupervised incremental learning approaches in traversal cost prediction. The traversal cost is defined by the proprioceptive signal of the robot traversal stability that is combined with appearance and geometric properties of the traversed terrains to construct the traversal cost model incrementally. In the motivational deployment, such a model is instantaneously utilized to extrapolate the traversal cost for observed areas that have not yet been visited by the robot to avoid difficult terrains in motion planning. The examined approaches are Incremental Gaussian Mixture Network, Growing Neural Gas, Improved Self-Organizing Incremental Neural Network, Locally Weighted Projection Regression, and Bayesian Committee Machine with Gaussian Process Regressors. The performance is examined using a dataset of the various terrains traversed by a real hexapod walking robot. A part of the presented benchmarking is thus a description of the dataset and also a construction of the reference traversal cost model that is used for comparison of the evaluated regressors. The reference is designed as a compound Gaussian process-based model that is learned separately over the individual terrain types. Based on the evaluation results, the best performance among the examined regressors is provided by Incremental Gaussian Mixture Network, Improved Self-Organizing Incremental Neural Network, and Locally Weighted Projection Regression, while the latter two have the lower computational requirements.

Data collection path planning with spatially correlated measurements using growing self-organizing array

  • DOI: 10.1016/j.asoc.2018.11.005
  • Odkaz: https://doi.org/10.1016/j.asoc.2018.11.005
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Data collection path planning is a problem to determine a cost-efficient path to read the most valuable data from a given set of sensors. The problem can be formulated as a variant of the combinatorial optimization problems that are called the price-collecting traveling salesman problem or the orienteering problem in a case of the explicitly limited travel budget. In these problems, each location is associated with a reward characterizing the importance of the data from the particular sensor location. The used simplifying assumption is to consider the measurements at particular locations independent, which may be valid, e.g., for very distant locations. However, measurements taken from spatially close locations can be correlated, and data collected from one location may also include information about the nearby locations. Then, the particular importance of the data depends on the currently selected sensors to be visited by the data collection path, and the travel cost can be saved by avoiding visitation of the locations that do not provide added value to the collected data. This is a computationally challenging problem because of mutual dependency on the cost of data collection path and the possibly collected rewards along such a path. A novel solution based on unsupervised learning method called the Growing Self-Organizing Array (GSOA) is proposed to address computational challenges of these problems and provide a solution in tens of milliseconds using conventional computational resources. Moreover, the employed GSOA-based approach allows to exploit capability to retrieve data by wireless communication or remote sensing, and thus further save the travel cost.

Data Collection Planning with Non-zero Sensing Distance for a Budget and Curvature Constrained Unmanned Aerial Vehicle

  • DOI: 10.1007/s10514-019-09844-5
  • Odkaz: https://doi.org/10.1007/s10514-019-09844-5
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    Data collection missions are one of the many effective use cases of Unmanned Aerial Vehicles (UAVs), where the UAV is required to visit a predefined set of target locations to retrieve data. However, the flight time of a real UAV is time constrained, and therefore only a limited number of target locations can typically be visited within the mission. In this paper, we address the data collection planning problem called the Dubins Orienteering Problem with Neighborhoods (DOPN), which sets out to determine the sequence of visits to the most rewarding subset of target locations, each with an associated reward, within a given travel budget. The objective of the DOPN is thus to maximize the sum of the rewards collected from the visited target locations using a budget constrained path between predefined starting and ending locations. The variant of the Orienteering Problem (OP) addressed here uses curvature-constrained Dubins vehicle model for planning the data collection missions for UAV. Moreover, in the DOPN, it is also assumed that the data, and thus the reward, may be collected from a close neighborhood sensing distance around the target locations, e.g., taking a snapshot by an onboard camera with a wide field of view, or using a sensor with a long range. We propose a novel approach based on the Variable Neighborhood Search (VNS) metaheuristic for the DOPN, in which combinatorial optimization of the sequence for visiting the target locations is simultaneously addressed with continuous optimization for finding Dubins vehicle waypoints inside the neighborhoods of the visited targets. The proposed VNS-based DOPN algorithm is evaluated in numerous benchmark instances, and the results show that it significantly outperforms the existing methods in both solution quality and computational time. The practical deployability of the proposed approach is experimentally verified in a data collection scenario with a real hexarotor UAV.

Emergency landing aware surveillance planning for fixed-wing planes

  • DOI: 10.1109/ECMR.2019.8870933
  • Odkaz: https://doi.org/10.1109/ECMR.2019.8870933
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we introduce the Emergency Landing Aware Surveillance Planning (ELASP) problem that stands to find the shortest feasible trajectory to visit a given set of locations while considering a loss of thrust may happen to the vehicle at any time. Two main challenges can be identified in ELASP. First, the ELASP is a planning problem to determine a feasible close-loop trajectory visiting all given locations such that the total trajectory length is minimized, which is a variant of the traveling salesman problem. The second challenge arises from the safety constraints to determine the cost-efficient trajectory such that its altitude is sufficiently high to guarantee a gliding emergency landing to a nearby airport from any point of the trajectory. Methods to address these challenges individually already exist, but the proposed approach enables to combine the existing methods to address both challenges at the same time and returns a safe, feasible, and cost-efficient multi-goal trajectory for the curvature-constrained vehicle.

Fast Heuristics for the 3-D Multi-Goal Path Planning Based on the Generalized Traveling Salesman Problem With Neighborhoods

  • DOI: 10.1109/LRA.2019.2900507
  • Odkaz: https://doi.org/10.1109/LRA.2019.2900507
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this letter, we address the multi-goal path planning problem to determine a cost-efficient path to visit a set of three-dimensional regions. The problem is a variant of the traveling salesman problem with neighborhoods (TSPN) where an individual neighborhood consists of multiple regions, and the problem is to determine a shortest multi-goal path to visit at least one region of each neighborhood. Because each neighborhood may consist of several regions, it forms a neighborhood set, and the problem is called the generalized TSPN (GTSPN) in the literature. We propose two heuristic algorithms to provide a feasible solution of the GTSPN quickly. The first algorithm is based on a decoupled approach using a solution of the generalized TSP that is further improved by a quick post-processing procedure. Besides, we propose to employ the existing unsupervised learning technique called the growing self-organizing array to quickly find a feasible solution of the GTSPN that can be further improved by more demanding optimization. The reported results on existing benchmarks for the GTSPN indicate that both proposed heuristics provide better or competitive solutions than the state-of-the-art reference algorithm, but they are up to two orders of magnitude faster.

Incremental Learning of Traversability Cost for Aerial Reconnaissance Support to Ground Units

  • Autoři: Ing. Miloš Prágr, Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Modelling and Simulation for Autonomous Systems. Basel: Springer, 2019. p. 412-421. LNCS. vol. 11472. ISSN 0302-9743. ISBN 978-3-030-14983-3.
  • Rok: 2019
  • DOI: 10.1007/978-3-030-14984-0_30
  • Odkaz: https://doi.org/10.1007/978-3-030-14984-0_30
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we address traversability cost estimation using exteroceptive and proprioceptive data collected by a team of aerial and ground vehicles. The main idea of the proposed approach is to estimate the terrain traversability cost based on the real experience of the multi-legged walking robot with traversing different terrain types. We propose to combine visual features with the real measured traversability cost based on proprioceptive signals of the utilized hexapod walking robot as a ground unit. The estimated traversability cost is augmented by extracted visual features from the onboard robot camera, and the features are utilized to extrapolate the learned traversability model for an aerial scan of new environments to assess their traversability cost. The extrapolated traversability cost can be utilized in the high-level mission planning to avoid areas that are difficult to traverse but not visited by the ground units. The proposed approach has been experimentally verified with a real hexapod walking robot in indoor and outdoor scenarios.

Incremental Traversability Assessment Learning Using Growing Neural Gas Algorithm

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Ing. Miloš Prágr,
  • Publikace: Advances in Self-Organizing Maps, Learning Vector Quantization, Clustering and Data Visualization. Düsseldorf: Springer-VDI-Verlag, 2019. p. 166-176. ISSN 2194-5357. ISBN 978-3-030-19641-7.
  • Rok: 2019
  • DOI: 10.1007/978-3-030-19642-4_17
  • Odkaz: https://doi.org/10.1007/978-3-030-19642-4_17
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we report early results on the deployment of the growing neural gas algorithm in online incremental learning of traversability assessment with a multi-legged walking robot. The addressed problem is to incrementally build a model of the robot experience with traversing the terrain that can be immediately utilized in the traversability cost assessment of seen but not yet visited areas. The main motivation of the studied deployment is to improve the performance of the autonomous mission by avoiding hard to traverse areas and support planning cost-efficient paths based on the continuously collected measurements characterizing the operational environment. We propose to employ the growing neural gas algorithm to incrementally build a model of the terrain characterization from exteroceptive features that are associated with the proprioceptive based estimation of the traversal cost. Based on the reported results, the proposed deployment provides competitive results to the existing approach based on the Incremental Gaussian Mixture Network.

Localization Fusion for Aerial Vehicles in Partially GNSS Denied Environments

  • Autoři: Ing. Jan Bayer, prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Modelling and Simulation for Autonomous Systems. Basel: Springer, 2019. p. 251-262. LNCS. vol. 11472. ISSN 0302-9743. ISBN 978-3-030-14983-3.
  • Rok: 2019
  • DOI: 10.1007/978-3-030-14984-0_20
  • Odkaz: https://doi.org/10.1007/978-3-030-14984-0_20
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we report on early results of the experimental deployment of localization techniques for a multi-rotor Micro Aerial Vehicle (MAV). In particular, we focus on deployment scenarios where the Global Navigation Satellite System (GNSS) does not provide a reliable signal, and thus it is not desirable to rely solely on the GNSS. Therefore, we consider recent advancements in the visual localization, and we employ an onboard RGB-D camera to develop a robust and reliable solution for the MAV localization in partially GNSS denied operational environments. We consider a localization method based on Kalman filter for data fusion of the vision-based localization with the signal from the GNSS. Based on the reported experimental results, the proposed solution supports the localization of the MAV for the temporarily unavailable GNSS, but also improves the position estimation provided by the incremental vision-based localization system while it can run using onboard computational resources of the small vehicle.

Modeling Proprioceptive Sensing for Locomotion Control of Hexapod Walking Robot in Robotic Simulator

  • Autoři: Nguyenová, M., Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Modelling and Simulation for Autonomous Systems. Basel: Springer, 2019. p. 215-225. LNCS. vol. 11472. ISSN 0302-9743. ISBN 978-3-030-14983-3.
  • Rok: 2019
  • DOI: 10.1007/978-3-030-14984-0_17
  • Odkaz: https://doi.org/10.1007/978-3-030-14984-0_17
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Proprioceptive sensing encompasses the state of the robot given by its overall posture, forces, and torques acting on its body. It is an important source of information, especially for multi-legged walking robots because it enables efficient locomotion control that adapts to morphological and environmental changes. In this work, we focus on enhancing a simplified model of the multi-legged robot employed in a realistic robotic simulator to provide high-fidelity proprioceptive sensor signals. The proposed model enhancements are based on parameter identification and static and dynamic modeling of the robot. The enhanced model enables the V-REP robotic simulator to be used in real-world deployments of multi-legged robots. The performance of the developed simulation has been verified in the parameter search of dynamic locomotion gait to optimize the locomotion speed according to the limited maximal torques and self-collision free execution.

Multi-Vehicle Close Enough Orienteering Problem with Bézier Curves for Multi-Rotor Aerial Vehicles

  • DOI: 10.1109/ICRA.2019.8794339
  • Odkaz: https://doi.org/10.1109/ICRA.2019.8794339
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    This paper introduces the Close Enough Orienteering Problem (CEOP) for planning missions with multi-rotor aerial vehicles considering their maximal velocity and acceleration limits. The addressed problem stands to select the most rewarding target locations and sequence to visit them in the given limited travel budget. The reward is collected within a non-zero range from a particular target location that allows saving the travel cost, and thus collect more rewards. Hence, we are searching for the fastest trajectories to collect the most valuable rewards such that the motion constraints are not violated, and the travel budget is satisfied. We leverage on existing trajectory parametrization based on Bézier curves recently deployed in surveillance planning using unsupervised learning, and we propose to employ the learning in a solution of the introduced multi-vehicle CEOP. Feasibility of the proposed approach is supported by empirical evaluation and experimental deployment using multi-rotor vehicles.

On autonomous spatial exploration with small hexapod walking robot using tracking camera intel RealSense T265

  • DOI: 10.1109/ECMR.2019.8870968
  • Odkaz: https://doi.org/10.1109/ECMR.2019.8870968
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we report on the deployment of the combination of commercially available off-the-shelf embedded visual localization system and RGB-D camera in an autonomous robotic exploration performed by small hexapod walking robot. Since the multi-legged walking robot is capable of traversing rough terrains, the addressed exploration problem is to create a map of an unknown environment while simultaneously performing the traversability assessment of the explored environment to efficiently and safely reach next navigational waypoints. The proposed system is targeted to run onboard of small multi-legged robots, and therefore, the system design is focused on computationally efficient approaches using relatively lightweight components. Therefore, we take advantages of the recently introduced tracking camera Intel RealSense T265 and RGB-D camera Intel RealSense D435 that are deployed to our developed autonomous hexapod walking robot that is equipped with adaptive locomotion control. Together with the proposed computationally efficient data representation and traversability assessment, the developed system supports onboard mapping and online decision-making within the exploration strategy even on a platform with low computational capabilities. Based on the reported experimental evaluation of the tracking camera, the developed system provides sufficiently accurate localization, and the robot has been able to explore indoor and outdoor environments fully autonomously.

On Unsupervised Learning of Traversal Cost and Terrain Types Identification Using Self-organizing Maps

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Ing. Miloš Prágr,
  • Publikace: Artificial Neural Networks and Machine Learning – ICANN 2019: Theoretical Neural Computation. ICANN 2019. Basel: Springer, 2019. p. 654-668. Lecture Notes in Computer Science. vol. 11727. ISSN 0302-9743. ISBN 978-3-030-30486-7.
  • Rok: 2019
  • DOI: 10.1007/978-3-030-30487-4_50
  • Odkaz: https://doi.org/10.1007/978-3-030-30487-4_50
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    This paper reports on the deployment of self-organizing maps in unsupervised learning of the traversal cost for a hexapod walking robot. The problem is motivated by traversability assessment of terrains not yet visited by the robot, but for which shape and appearance features are available. The perception system of the robot is used to extract terrain features that are accompanied by traversal cost characterization captured from the real experience of the robot with the terrain, which is characterized by proprioceptive features. The learned model is employed to predict the traversal cost of new terrains based only on the shape and appearance features. Based on the experimental deployment of the robot in various terrains, a dataset of the traversal cost has been collected that is utilized in the presented evaluation of the traversal cost modeling using self-organizing map approach. In comparison with the Gaussian process, the self-organizing map provides competitive results and the found paths using the predicted traversal costs are close to the optimal path based on reference traversal cost of the particular terrain types. Besides, the self-organizing map can also be utilized for unsupervised identification of the terrain types, and it further supports incremental learning, which is more suitable for practical deployments of the robot in a priory unknown environments where reference traversal costs are not available.

Online Incremental Learning of the Terrain Traversal Cost in Autonomous Exploration

  • DOI: 10.15607/RSS.2019.XV.040
  • Odkaz: https://doi.org/10.15607/RSS.2019.XV.040
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we address motion efficiency in autonomous robot exploration with multi-legged walking robots that can traverse rough terrains at the cost of lower efficiency and greater body vibration. We propose a robotic system for online and incremental learning of the terrain traversal cost that is immediately utilized to reason about next navigational goals in building spatial model of the robot surrounding. The traversal cost experienced by the robot is characterized by incrementally constructed Gaussian Processes using Bayesian Committee Machine. During the exploration, the robot builds the spatial terrain model, marks untraversable areas, and leverages the Gaussian Process predictive variance to decide whether to improve the spatial model or decrease the uncertainty of the terrain traversal cost. The feasibility of the proposed approach has been experimentally verified in a fully autonomous deployment with the hexapod walking robot

Physical Orienteering Problem for Unmanned Aerial Vehicle Data Collection Planning in Environments with Obstacles

  • DOI: 10.1109/LRA.2019.2923949
  • Odkaz: https://doi.org/10.1109/LRA.2019.2923949
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    This paper concerns a variant of the Orienteering Problem (OP) that arises from multi-goal data collection scenarios where a robot with a limited travel budget is requested to visit given target locations in an environment with obstacles. We call the introduced OP variant the Physical Orienteering Problem (POP). The POP sets out to determine a feasible, collision-free, path that maximizes collected reward from a subset of the target locations and does not exceed the given travel budget. The problem combines motion planning and combinatorial optimization to visit multiple target locations. The proposed solution to the POP is based on the Variable Neighborhood Search (VNS) method combined with the asymptotically optimal sampling-based Probabilistic Roadmap (PRM*) method. The VNS-PRM* uses initial low-dense roadmap that is continuously expanded during the VNS-based POP optimization to shorten paths of the promising solutions, and thus allows maximizing the sum of the collected rewards. The computational results support the feasibility of the proposed approach by a fast determination of high-quality solutions. Moreover, an experimental verification demonstrates the applicability of the proposed VNS-PRM* approach for data collection planning for an unmanned aerial vehicle in an urban-like environment with obstacles.

Self-supervised learning of the biologically-inspired obstacle avoidance of hexapod walking robot

  • DOI: 10.1088/1748-3190/ab1a9c
  • Odkaz: https://doi.org/10.1088/1748-3190/ab1a9c
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we propose an integrated biologically inspired visual collision avoidance approach that is deployed on a real hexapod walking robot. The proposed approach is based on the Lobula giant movement detector (LGMD), a neural network for looming stimuli detection that can be found in visual pathways of insects, such as locusts. Although a superior performance of the LGMD in the detection of intercepting objects has been shown in many collision avoiding scenarios, its direct integration with motion control is an unexplored topic. In our work, we propose to utilize the LGMD neural network for visual interception detection with a central pattern generator (CPG) for locomotion control of a hexapod walking robot that are combined in the controller based on the long short-term memory (LSTM) recurrent neural network. Moreover, we propose self-supervised learning of the integrated controller to autonomously find a suitable setting of the system using a realistic robotic simulator. Thus, individual neural networks are trained in a simulation to enhance the performance of the controller that is then experimentally verified with a real hexapod walking robot in both collision and interception avoidance scenario and navigation in a cluttered environment.

Trajectory Planning for Aerial Vehicles in the Area Coverage Problem with Nearby Obstacles

  • Autoři: Marek, J., Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Modelling and Simulation for Autonomous Systems. Basel: Springer, 2019. p. 226-236. LNCS. vol. 11472. ISSN 0302-9743. ISBN 978-3-030-14983-3.
  • Rok: 2019
  • DOI: 10.1007/978-3-030-14984-0_18
  • Odkaz: https://doi.org/10.1007/978-3-030-14984-0_18
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we address the coverage path planning with curvature-constrained paths for a fixed-wing aerial vehicle. The studied problem is to provide a cost-efficient solution to cover a given area by the vehicle sensor from the specified altitude to provide a sufficient level of details in the captured snapshots of the area. In particular, we focus on scenarios where the area to be covered is surrounded by nearby obstacles such as trees or buildings, and the vehicle has to avoid collisions with the obstacles but maximizes the area coverage. We propose an extension of the existing coverage planning algorithm to determine a shortest collision-free path that is accompanied by Dubins Airplane model to satisfy the motion constraints of the vehicle. The reported results support the feasibility of the proposed approach to avoid nearby obstacles by optimal adjustments of the vehicle altitude while the requested complete coverage is satisfied. If such a solution is not found because of too close obstacles, a feasible solution maximizing the coverage is provided.

Traversal cost modeling based on motion characterization for multi-legged walking robots

  • DOI: 10.1109/ECMR.2019.8870912
  • Odkaz: https://doi.org/10.1109/ECMR.2019.8870912
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this paper, we concern a traversal cost estimation considering motion control of a hexapod walking robot. The proposed idea is motivated by the observation that the traversal cost depends not only on the traversed terrain but also on the robot motion. Based on the experimental deployments, the forward motion is preferable over some terrains; however, uphill and downhill locomotion over the particular terrain might differ significantly. Therefore, we propose to enhance the traversal cost model by a motion characterization. The model is learned using feature descriptor composed of terrain shape and appearance that is combined with the expected motion performance determined from the slope change and possible rotation of the robot. The traversal model enables to reason about the robot stability regarding placement of the robot legs and performed motion action. The proposed idea of motion characterization is demonstrated and experimentally verified on a simplified motion control using grid-based planning with the robot control decomposed into straight and turn movements.

Unsupervised learning‐based flexible framework for surveillance planning with aerial vehicles

  • DOI: 10.1002/rob.21823
  • Odkaz: https://doi.org/10.1002/rob.21823
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    The herein studied problem is motivated by practical needs of our participation in the Mohamed Bin Zayed International Robotics Challenge (MBZIRC) 2017 in which a team of unmanned aerial vehicles (UAVs) is requested to collect objects in the given area as quickly as possible and score according to the rewards associated with the objects. The mission time is limited, and the most time‐consuming operation is the collection of the objects themselves. Therefore, we address the problem to quickly identify the most valuable objects as surveillance planning with curvature‐constrained trajectories. The problem is formulated as a multivehicle variant of the Dubins traveling salesman problem with neighborhoods (DTSPN). Based on the evaluation of existing approaches to the DTSPN, we propose to use unsupervised learning to find satisfiable solutions with low computational requirements. Moreover, the flexibility of unsupervised learning allows considering trajectory parametrization that better fits the motion constraints of the utilized hexacopters that are not limited by the minimal turning radius as the Dubins vehicle. We propose to use Bézier curves to exploit the maximal vehicle velocity and acceleration limits. Besides, we further generalize the proposed approach to 3D surveillance planning. We report on evaluation results of the developed algorithms and experimental verification of the planned trajectories using the real UAVs utilized in our participation in MBZIRC 2017.

Variable Neighborhood Search for the Set Orienteering Problem and its application to other Orienteering Problem variants

  • DOI: 10.1016/j.ejor.2019.01.047
  • Odkaz: https://doi.org/10.1016/j.ejor.2019.01.047
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    This paper addresses the recently proposed generalization of the Orienteering Problem (OP), referred to as the Set Orienteering Problem (SOP). The OP stands to find a tour over a subset of customers, each with an associated profit, such that the profit collected from the visited customers is maximized and the tour length is within the given limited budget. In the SOP, the customers are grouped in clusters, and the profit associated with each cluster is collected by visiting at least one of the customers in the respective cluster. Similarly to the OP, the SOP limits the tour cost by a given budget constraint, and therefore, only a subset of clusters can usually be served. We propose to employ the Variable Neighborhood Search (VNS) metaheuristic for solving the SOP. In addition, a novel Integer Linear Programming (ILP) formulation of the SOP is proposed to find the optimal solution for small and medium-sized problems. Furthermore, we show other OP variants that can be addressed as the SOP, i.e., the Orienteering Problem with Neighborhoods (OPN) and the Dubins Orienteering Problem (DOP). While the OPN extends the OP by collecting a profit within the neighborhood radius of each customer, the DOP uses airplane-like smooth trajectories to connect individual customers. The presented computational results indicate the feasibility of the proposed VNS algorithm and ILP formulation, by improving the solutions of several existing SOP benchmark instances and requiring significantly lower computational time than the existing approaches.

Any-Time Trajectory Planning for Safe Emergency Landing

  • DOI: 10.1109/IROS.2018.8594225
  • Odkaz: https://doi.org/10.1109/IROS.2018.8594225
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    Loss of thrust is a critical situation for human pilots of fixed-wing aircraft which force them to select a landing site in the nearby range and perform an emergency landing. The time for the landing site selection is limited by the actual altitude of the aircraft, and it may be fatal if the correct decision is not chosen fast enough. Therefore, we propose a novel RRT* -based planning algorithm for finding the safest emergency landing trajectory towards a given set of possible landing sites. Multiple landing sites are evaluated simultaneously during the flight even before any mechanical issue occurs, and the roadmap of possible landing trajectories is updated permanently. Thus, the proposed algorithm has the any-time property and provides the best emergency landing trajectory almost instantly.

Autonomous Data Collection Using a Self-Organizing Map

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Hollinger, G.
  • Publikace: IEEE Transactions on Neural Networks and Learning Systems. 2018, 29(5), 1703-1715. ISSN 2162-237X.
  • Rok: 2018
  • DOI: 10.1109/TNNLS.2017.2678482
  • Odkaz: https://doi.org/10.1109/TNNLS.2017.2678482
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    The self-organizing map (SOM) is an unsupervised learning technique providing a transformation of a high-dimensional input space into a lower dimensional output space. In this paper, we utilize the SOM for the traveling salesman problem (TSP) to develop a solution to autonomous data collection. Autonomous data collection requires gathering data from predeployed sensors by moving within a limited communication radius. We propose a new growing SOM that adapts the number of neurons during learning, which also allows our approach to apply in cases where some sensors can be ignored due to a lower priority. Based on a comparison with available combinatorial heuristic algorithms for relevant variants of the TSP, the proposed approach demonstrates improved results, while also being less computationally demanding. Moreover, the proposed learning procedure can be extended to cases where particular sensors have varying communication radii, and it can also be extended to multivehicle planning.

Communication Architecture in Mixed-Reality Simulations of Unmanned Systems

  • DOI: 10.3390/s18030853
  • Odkaz: https://doi.org/10.3390/s18030853
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Verification of the correct functionality of multi-vehicle systems in high-fidelity scenarios is required before any deployment of such a complex system, e.g., in missions of remote sensing or in mobile sensor networks. Mixed-reality simulations where both virtual and physical entities can coexist and interact have been shown to be beneficial for development, testing, and verification of such systems. This paper deals with the problems of designing a certain communication subsystem for such highly desirable realistic simulations. Requirements of this communication subsystem, including proper addressing, transparent routing, visibility modeling, or message management, are specified prior to designing an appropriate solution. Then, a suitable architecture of this communication subsystem is proposed together with solutions to the challenges that arise when simultaneous virtual and physical message transmissions occur. The proposed architecture can be utilized as a high-fidelity network simulator for vehicular systems with implicit mobility models that are given by real trajectories of the vehicles. The architecture has been utilized within multiple projects dealing with the development and practical deployment of multi-UAV systems, which support the architecture’s viability and advantages. The provided experimental results show the achieved similarity of the communication characteristics of the fully deployed hardware setup to the setup utilizing the proposed mixed-reality architecture.

Cost of Transport Estimation for Legged Robot Based on Terrain Features Inference from Aerial Scan

  • Autoři: Ing. Miloš Prágr, Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). New York: IEEE Press, 2018. p. 1745-1750. ISSN 2153-0866. ISBN 978-1-5386-8094-0.
  • Rok: 2018
  • DOI: 10.1109/IROS.2018.8593374
  • Odkaz: https://doi.org/10.1109/IROS.2018.8593374
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    The effectiveness of the robot locomotion can be measured using the cost of transport (CoT) which represents the amount of energy that is needed for traversing from one place to another. Terrains excerpt different mechanical properties when crawled by a multi-legged robot, and thus different values of the CoT. It is therefore desirable to estimate the CoT in advance and plan the robot motion accordingly. However, the CoT might not be known prior the robot deployment, e.g., in extraterrestrial missions; hence, a robot has to learn different terrains as it crawls through the environment incrementally. In this work, we focus on estimating the CoT from visual and geometrical data of the crawled terrain. A thorough analysis of different terrain descriptors within the context of incremental learning is presented to select the best performing approach. We report on the achieved results and experimental verification of the selected approaches with a real hexapod robot crawling over six different terrains.

GSOA: Growing Self-Organizing Array - Unsupervised learning for the Close-Enough Traveling Salesman Problem and other routing problems

  • DOI: 10.1016/j.neucom.2018.05.079
  • Odkaz: https://doi.org/10.1016/j.neucom.2018.05.079
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper presents a novel unsupervised learning procedure called the Growing Self-Organizing Array (GSOA) that is inspired by principles of the self-organizing maps for the Traveling Salesman Problem (TSP). The proposed procedure is a consolidation of principles deployed in solution of several variants of the generalized TSP with Neighborhoods (TSPN) for which the main ideas of the proposed unsupervised learning already demonstrates a wide range of applicability. The herein presented learning procedure is a conceptually simple algorithm which outperforms previous self-organizing map based approaches for the TSP in terms of the solution quality and required computational time. The main benefit of the proposed learning procedure is in solving routing problems that combine a combinatorial solution of some variant of the TSP with the continuous optimization, i.e., problems where it is needed to determine a sequence of visits to the given sets with determination of the particular waypoint location from each (possibly infinite) set. Low computational requirements of the proposed method are demonstrated in a solution of the Close-Enough Traveling Salesman Problem (CETSP), which is a special case of the TSPN with the disk-shaped neighborhoods. The results indicate the proposed procedure provides competitive solutions to the existing heuristics while it is about one order of magnitude faster and at least about two orders of magnitude faster than a heuristic solution of the discretized variant of the CETSP considered as the Generalized TSP. (C) 2018 Elsevier B.V. All rights reserved.

Learning Central Pattern Generator Network with Back-Propagation Algorithm

  • Autoři: Ing. Rudolf Jakub Szadkowski, Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the 18th Conference Information Technologies - Applications and Theory (ITAT 2018). Aachen: CEUR Workshop Proceedings, 2018. p. 116-123. vol. 2203. ISSN 1613-0073. ISBN 9781727267198.
  • Rok: 2018
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    An adaptable central pattern generator (CPG) that di-rectly controls the rhythmic motion of multilegged robotmust combine plasticity and sustainable periodicity. Thiscombination requires an algorithm that searches the para-metric space of the CPG and yields a non-stationary andnon-divergent solution. We model the CPG with the pi-oneering Matsuoka’s neural oscillator which is (mostly)non-divergent and provides constraints ensuring non-stationarity. We embed these constraints into the CPGformulation which we further implemented as a layer ofan artificial neural network. This enables the CPG to belearnable by back-propagation algorithm while sustainingthe desirable properties. Moreover, the proposed CPG canbe integrated into more complex networks and trained un-der different optimization objectives. In addition to thetheoretical properties of the developed system, its flexibil-ity is demonstrated in successful learning of the tripod mo-tion gait with its practical deployment on the real hexapodwalking robot.

On Locomotion Control Using Position Feedback Only in Traversing Rough Terrains with Hexapod Crawling Robot

  • Autoři: Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: IOP Conference Series: Materials Science and Engineering. Bristol: Institute of Physics Publishing, 2018. p. 1-10. ISSN 1757-899X.
  • Rok: 2018
  • DOI: 10.1088/1757-899X/428/1/012065
  • Odkaz: https://doi.org/10.1088/1757-899X/428/1/012065
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we report on our results on improved locomotion control of small and affordable hexapod crawling robot using only position feedback from the utilized servo motors. Multi-legged robots represent complex mechanical systems with many degrees of freedom from which they can benefit in traversing rough terrains. However, the crucial ability of multi-legged robots is maintaining stable locomotion over irregularities of the terrain which makes the locomotion control complex and requires reliable and timely detection of the leg contact point with the ground. Such detection may require additional sensory equipment which can increase the cost of the multi-legged platform. Therefore, we focus on exploiting capabilities of nowadays intelligent servo motors with position feedback to develop a minimalistic set up in which the robot uses solely the position feedback of the servo motors to sense the ground reaction force. The first achievements enable a small hexapod crawling robot to navigate rough terrains using stable pentapod gait, where only one leg moves at a time, and five legs support the robot. Later on, we improved the locomotion control to enable faster locomotion using three simultaneously moved legs in the so-called tripod motion gait. This paper reports on further advancements with a faster control loop enabled by hardware based acceleration of the communication latency with the utilized Dynamixel AX12 servo motors that improve the locomotion capabilities of the robot. The reported results indicate the robot locomotion with the used adaptive motion gait is speeded up by a factor of 1.4 with the same stability in traversing the rough terrain of the experimental laboratory mock-up.

On Unsupervised Learning based Multi-Goal Path Planning for Visiting 3D Regions

  • DOI: 10.1145/3297097.3297099
  • Odkaz: https://doi.org/10.1145/3297097.3297099
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we report on our early results on deploying unsupervised learning technique for solving a multi-goal path planning problem to determine a shortest path to visit a given set of 3D regions. The addressed problem is motivated by data collection missions in which a robotic vehicle is requested to visit a set of locations to perform particular measurements. Instead of precise visitation of the specified locations, it is allowed to take the measurements at the respective distance from the locations, and thus save the travel cost by exploiting non-zero sensing radius of the vehicle. In particular, the problem is formulated as a 3D variant of the Close-Enough Traveling Salesman Problem (CETSP), and the proposed approach is based on the recently introduced technique called the Growing Self-Organizing Array (GSOA). The GSOA is a neural network for routing problems that is accompanied with unsupervised learning procedure to determine a solution of the TSP-like problems in a finite number of learning epochs. Based on the reported results, the proposed GSOA-based approach provides competitive or better results than existing combinatorial heuristics based on the so-called Steiner zones, while the computational requirements are significantly lower.

Online Foot-Strike Detection Using Inertial Measurements for Multi-Legged Walking Robots

  • Autoři: Čížek, P., Ing. Jiří Kubík, prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). New York: IEEE Press, 2018. p. 7622-7627. ISSN 2153-0866. ISBN 978-1-5386-8094-0.
  • Rok: 2018
  • DOI: 10.1109/IROS.2018.8594010
  • Odkaz: https://doi.org/10.1109/IROS.2018.8594010
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Proprioceptive terrain sensing is essential for rough terrain traversal because it helps legged robots to negotiate individual steps by reacting to terrain irregularities. In this work, we propose to utilize inertial data in the detection of the contact between the leg and the terrain during the stride phase of the leg. We show that relatively cheap accelerometers can be utilized to reliably detect a foot-strike, and thus allow the robot to crawl irregular terrains. The continuous data processing is compared with the interrupt mode in which data are provided only around the foot-strike event. The interrupt mode exhibits significantly better performance, and it also supports generalization of the foot-strike event detector learned from data collected in slow locomotion to faster locomotion where the signals slightly change. The proposed solution is experimentally validated using a real hexapod walking robot for which the walking speed has been improved in comparison to the previous adaptive motion gait based on a force threshold-based position controller for the foot-strike detection.

Online planning for multi-robot active perception with self-organising maps

  • Autoři: Best, G., prof. Ing. Jan Faigl, Ph.D., Fitch, R.
  • Publikace: Autonomous Robots. 2018, 2018(42), 715-738. ISSN 0929-5593.
  • Rok: 2018
  • DOI: 10.1007/s10514-017-9691-4
  • Odkaz: https://doi.org/10.1007/s10514-017-9691-4
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    We propose a self-organising map (SOM) algorithm as a solution to a new multi-goal path planning problem for active perception and data collection tasks. We optimise paths for a multi-robot team that aims to maximally observe a set of nodes in the environment. The selected nodes are observed by visiting associated viewpoint regions defined by a sensor model. The key problem characteristics are that the viewpoint regions are overlapping polygonal continuous regions, each node has an observation reward, and the robots are constrained by travel budgets. The SOM algorithm jointly selects and allocates nodes to the robots and finds favourable sequences of sensing locations. The algorithm has a runtime complexity that is polynomial in the number of nodes to be observed and the magnitude of the relative weighting of rewards. We show empirically the runtime is sublinear in the number of robots. We demonstrate feasibility for the active perception task of observing a set of 3D objects. The viewpoint regions consider sensing ranges and self-occlusions, and the rewards are measured as discriminability in the ensemble of shape functions feature space. Exploration objectives for online tasks where the environment is only partially known in advance are modelled by introducing goal regions in unexplored space. Online replanning is performed efficiently by adapting previous solutions as new information becomes available. Simulations were performed using a 3D point-cloud dataset from a real robot in a large outdoor environment. Our results show the proposed methods enable multi-robot planning for online active perception tasks with continuous sets of candidate viewpoints and long planning horizons.

Optimal Solution of the Generalized Dubins Interval Problem

  • Autoři: Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the Robotics and Systems. Pensylvánie: Carnegie Melon University, Robotic Institute, 2018. ISBN 978-0-9923747-4-7.
  • Rok: 2018
  • DOI: 10.15607/RSS.2018.XIV.035
  • Odkaz: https://doi.org/10.15607/RSS.2018.XIV.035
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    The problem addressed in this paper is motivated by surveillance mission planning with curvature-constrained trajectories for Dubins vehicles that can be formulated as the Dubins Traveling Salesman Problem with Neighborhoods (DTSPN). We aim to provide a tight lower bound of the DTSPN, especially for the cases where the sequence of visits to the given regions is available. A problem to find the shortest Dubins path connecting two regions with prescribed intervals for possible departure and arrival heading angles of the vehicle is introduced. This new problem is called the Generalized Dubins Interval Problem (GDIP) and its optimal solution is addressed. Based on the solution of the GDIP, a tight lower bound of the above mentioned DTSPN is provided which is used to steer sampling-based algorithm to determine a feasible solution that is close to the optimum

Real-Time FPGA-Based Detection of Speeded-Up Robust Features Using Separable Convolution

  • Autoři: Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS. 2018, 14(3), 1155-1163. ISSN 1551-3203.
  • Rok: 2018
  • DOI: 10.1109/TII.2017.2764485
  • Odkaz: https://doi.org/10.1109/TII.2017.2764485
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we propose a novel architecture for efficient detection of Speeded Up Robust Features (SURF) for Field-programmable gate array (FPGA). The main benefits of the proposed architecture are in real-time low-latency performance and scalability. The proposed solution provides a significant acceleration of salient points extraction which is fundamental image processing technique for vision-based methods including the simultaneous localization and mapping (SLAM). Based on the presented practical results, the proposed architecture is capable of processing streaming image data at the rate of 140 Megapixels per second which roughly scales from the 640×480@420fps up to 1920×1080@60fps video streams on a low-end, low-cost FPGA solution (Cyclone V). Moreover, the proposed feature detection utilizes only about 20% of logic elements of the FPGA which supports further parallel processing of multiple inputs.

Surveillance Planning With Bezier Curves

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Váňa, P.
  • Publikace: IEEE Robotics and Automation Letters. 2018, 3(2), 750-757. ISSN 2377-3766.
  • Rok: 2018
  • DOI: 10.1109/LRA.2018.2789844
  • Odkaz: https://doi.org/10.1109/LRA.2018.2789844
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This letter concerns surveillance planning for an unmanned aerial vehicle (UAV) that is requested to periodically take snapshots of areas of interest by visiting a given set of waypoint locations in the shortest time possible. The studied problem can be considered as a variant of the combinatorial traveling salesman problem in which trajectories between the waypoints respect the kinematic constraints of the UAV. Contrary to the existing formulation for curvature-constrained vehicles known as the Dubins traveling salesman problem, the herein addressed problem is motivated by planning for multirotor UAVs which are not limited by the minimal required forward velocity and minimal turning radius as the Dubins vehicle, but rather by the maximal speed and acceleration. Moreover, the waypoints to be visited can be at different altitudes, and the addressed problem is to find a fast and smooth trajectory in three-dimensional (3-D) space from which all the areas of interest can be captured. The proposed solution is based on unsupervised learning in which the requested 3-D smooth trajectory is determined as a sequence of Bezier curves in a finite number of learning epochs. The reported results support feasibility of the proposed solution which has also been experimentally verified with a real UAV.

Terrain Classification with Crawling Robot Using Long Short-Term Memory Network

  • DOI: 10.1007/978-3-030-01424-7_75
  • Odkaz: https://doi.org/10.1007/978-3-030-01424-7_75
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    Terrain classification is a crucial feature for mobile robots operating across multiple terrains. One way to learn a terrain classifier is to use a stream of labeled proprioceptive data recorded during a terrain traversal. In this paper, we propose a new terrain classifier that combines a feature extraction from a data stream with the long short-term memory (LSTM) network. Features are extracted from the information-sparse data stream by applying a sliding window computing three central moments. The feature sequence is continuously classified by the LSTM network into multiple terrain classes. Furthermore, a modified bagging method is used to deal with a limited and unbalanced training set. In comparison to the previous work on terrain classifiers for a hexapod crawling robot using only servo-drive feedback, the proposed classifier provides continuous classification with the F1 score up to 0.88, and thus provide better results than SVM classifier learned on the same input data.

The Dubins Traveling Salesman Problem with Neighborhoods in the Three-Dimensional Space

  • Autoři: Váňa, P., Ing. Jakub Sláma, prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the 2018 IEEE International Conference on Robotics and Automation. Piscataway, NJ: IEEE, 2018. p. 374-379. ISSN 1050-4729. ISBN 978-1-5386-3081-5.
  • Rok: 2018
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    We introduce an extension of the Dubins Traveling Salesman Problem with Neighborhoods into the 3D space in which a fixed-wing aerial vehicle is requested to visit a set of target regions while the vehicle motion constraints are satisfied, i.e., the minimum turning radius and maximum climb and dive angles. The primary challenge is to address both the combinatorial optimization part of finding the sequence of target visits and the continuous optimization part of the final trajectory determination. Due to its high complexity, we propose to address both parts of the problem separately by a decoupled approach in which the sequence is determined by a new distance function designed explicitly for the utilized 3D Dubins Airplane model. The final trajectory is then frond by a local optimization which improves the solution quality. The proposed approach provides significantly better solutions than using Euclidean distance in the sequencing part of the problem. Moreover, the found solutions are of the competitive quality to the sampling-based algorithm while its computational requirements are about two orders of magnitude lower.

An experimental study on feature-based SLAM for multi-legged robots with RGB-D sensors

  • Autoři: Nowicki, M.R., Belter, D., Kostusiak, A., Čížek, P., prof. Ing. Jan Faigl, Ph.D., Skrzypczyński, P.
  • Publikace: Industrial Robot. 2017, 44(4), 428-441. ISSN 0143-991X.
  • Rok: 2017
  • DOI: 10.1108/IR-11-2016-0340
  • Odkaz: https://doi.org/10.1108/IR-11-2016-0340
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper aims to evaluate four different simultaneous localization and mapping (SLAM) systems in the context of localization of multi-legged walking robots equipped with compact RGB-D sensors. This paper identifies problems related to in-motion data acquisition in a legged robot and evaluates the particular building blocks and concepts applied in contemporary SLAM systems against these problems. The SLAM systems are evaluated on two independent experimental set-ups, applying a well-established methodology and performance metrics. Four feature-based SLAM architectures are evaluated with respect to their suitability for localization of multi-legged walking robots. The evaluation methodology is based on the computation of the absolute trajectory error (ATE) and relative pose error (RPE), which are performance metrics well-established in the robotics community. Four sequences of RGB-D frames acquired in two independent experiments using two different six-legged walking robots are used in the evaluation process. The evaluation was performed using indoor mockups of terrain. Experiments in more natural and challenging environments are envisioned as part of future research. The lack of accurate self-localization methods is considered as one of the most important limitations of walking robots. The main contribution lies in the integration of the state-of-the-art SLAM methods on walking robots and their thorough experimental evaluation using a well-established methodology. Moreover, a SLAM system designed especially for RGB-D sensors and real-world applications is presented in details.

Data Collection Planning with Dubins Airplane Model and Limited Travel Budget

  • DOI: 10.1109/ECMR.2017.8098715
  • Odkaz: https://doi.org/10.1109/ECMR.2017.8098715
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    In this paper, we address the data collection planning problem for fixed-wing unmanned aircraft vehicle (UAV) with a limited travel budget. We formulate the problem as a variant of the Orienteering Problem (OP) in which the Dubins airplane model is utilized to extend the problem into the three-dimensional space and curvature-constrained vehicles. The proposed Dubins Airplane Orienteering Problem (DA-OP) stands to find the most rewarding data collection trajectory visiting a subset of the given target locations while the trajectory does not exceed the limited travel budget. Contrary to the original OP formulation, the proposed DA-OP combines not only the combinatorial part of determining a subset of the targets to be visited together with determining the sequence to visited them, but it also includes challenges related to continuous optimization in finding the shortest trajectory for Dubins airplane vehicle. The problem is addressed by sampling possible approaching angles to the targets, and a solution is found by the Randomized Variable Neighborhood Search (RVNS) method. A feasibility of the proposed solution is demonstrated by an empirical evaluation on modified benchmarks for the OP instances to the scenarios with varying altitude of the targets.

Data Collection Planning with Limited Budget for Dubins Airplane

  • Pracoviště: Katedra počítačů, Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    In this work, we address data collection planning with an Unmanned Aerial Vehicle (UAV) motivated by surveillance missions in which the UAV is requested to take snapshots at the given set of target locations. In particular, we focus on scenarios where UAV can be modeled by the Dubins airplane model in 3D and the travel budget is limited. In these problems, each target location has associated reward value representing an importance of the target, and thus the studied planning problem is to determine the most valuable targets together with the sequence to their visits such that the length of the data collection trajectory fits the travel budget.

Dubins Orienteering Problem

  • DOI: 10.1109/LRA.2017.2666261
  • Odkaz: https://doi.org/10.1109/LRA.2017.2666261
  • Pracoviště: Katedra kybernetiky, Centrum umělé inteligence
  • Anotace:
    In this paper, we address the Orienteering Problem (OP) for curvature constrained vehicle. For a given set of target locations, each with associated reward, the OP stands to find a tour from a prescribed starting location to a given ending location such that it maximizes collected rewards while the tour length is within a given travel budget constraint. The addressed generalization of the Euclidean OP is called the Dubins Orienteering Problem (DOP) in which the reward collecting tour has to satisfy the limited turning radius of the Dubins vehicle. The DOP consists not only of selecting the most valuable targets and determination of the optimal sequence to visit them, but it also involves the determination of the vehicle’s heading angle at each target location. The proposed solution is based on the Variable neighborhood search technique, and its feasibility is supported by an empirical evaluation in existing OP benchmarks. Moreover, an experimental verification in a real practical scenario further demonstrates the necessity of the proposed direct solution of the Dubins Orienteering Problem.

Dubins Orienteering Problem with Neighborhoods

  • DOI: 10.1109/ICUAS.2017.7991350
  • Odkaz: https://doi.org/10.1109/ICUAS.2017.7991350
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    In this paper, we address the Dubins Orienteering Problem with Neighborhoods (DOPN) a novel problem derived from the regular Orienteering Problem (OP). In the OP, one tries to find a maximal reward collecting path through a subset of given target locations, each with associated reward, such that the resulting path length does not exceed the specified travel budget. The Dubins Orienteering Problem (DOP) requires the reward collecting path to satisfy the curvature-constrained model of the Dubins vehicle while reaching precise positions of the target locations. In the newly introduced DOPN, the resulting path also respects the curvature constrained Dubins vehicle as in the DOP; however, the reward can be collected within a close distant neighborhood of the target locations. The studied problem is inspired by data collection scenarios for an Unmanned Aerial Vehicle (UAV), that can be modeled as the Dubins vehicle. Furthermore, the DOPN is a useful problem formulation of data collection scenarios for a UAV with the limited travel budget due to battery discharge and in scenarios where the sensoric data can be collected from a proximity of each target location. The proposed solution of the DOPN is based on the Variable Neighborhood Search method, and the presented computational results in the OP benchmarks support feasibility of the proposed approach.

Enhancing Neural Based Obstacle Avoidance with CPG Controlled Hexapod Walking Robot

  • Autoři: Čížek, P., prof. Ing. Jan Faigl, Ph.D., Ing. Jan Bayer,
  • Publikace: Proceedings of the 17th Conference on Information Technologies - Applications and Theory (ITAT 2017). Aachen: CEUR Workshop Proceedings, 2017. p. 65-70. 2017. vol. 1885. ISSN 1613-0073.
  • Rok: 2017
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    Avoiding collisions with obstacles and intercepting objects based on the visual perception is a vital survival ability of any animal. In this work, we propose an extension of the biologically based collision avoidance approach to the detection of intercepting objects using the Lobula Giant Movement Detector (LGMD) connected directly to the locomotion control unit based on the Central Pattern Generator (CPG) of a hexapod walking robot. The proposed extension uses Recurrent Neural Network (RNN) to map the output of the LGMD on the input of the CPG to enhance collision avoiding behavior of the robot in cluttered environments. The presented results of the experimental verification of the proposed system with a real mobile hexapod crawling robot support the feasibility of the presented approach in collision avoidance scenarios.

Foothold Placement Planning with a Hexapod Crawling Robot

  • Autoři: Čížek, P., Masri, D., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Piscataway: IEEE, 2017. p. 4096-4101. ISSN 2153-0858. ISBN 978-1-5386-2682-5.
  • Rok: 2017
  • DOI: 10.1109/IROS.2017.8206267
  • Odkaz: https://doi.org/10.1109/IROS.2017.8206267
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this work, we concern the problem of motion planning for a hexapod walking robot crawling in a semi-structured environment where a precise foot-tip positioning is necessary. We propose pipelined approach utilizing an RGB-D camera to perceive and map the forthcoming terrain in 2.5 D which is then processed for available foot-tip positions. The robot motion control is based on sampling-based planning to determine the most suitable leg supporting configurations for the individual body positions in the created terrain map. The individual body positions are connected into a roadmap with taking into account a feasibility of the robot transition between the individual configurations. The resulting trajectory is then planned in the created roadmap using a standard A* planner. The proposed method has been experimentally evaluated in the on-line and onboard setup with a real hexapod crawling robot. The herein reported results support feasibility of the proposed approach for a precise motion planning of small hexapod crawling robot in a semi-structured environment.

Mixed Reality Simulation for Incremental Development of Multi-UAV Systems

  • DOI: 10.1109/ICUAS.2017.7991351
  • Odkaz: https://doi.org/10.1109/ICUAS.2017.7991351
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Development of complex multi-robot systems requires time-consuming and expensive testing and, especially in a case of unmanned aerial systems, it aggregates risk of hardware failures and legal issues when operating more than one unmanned aircraft simultaneously. It is highly favorable to deal with most of the eventual design flaws and system bugs before the final field tests in a simulation where the risks are significantly lower. On the other hand, the fidelity of the simulation needs to rise as the system development approaches the final stages and since some phenomena are difficult to be modeled precisely, a partial embodiment of the simulation in the physical world is necessary. In this paper, we present our results in the utilization of mixed reality simulation for incremental development of multi-UAV systems. We present three use cases where this method was used for development of various systems to show its versatility: (i) an unmanned system consisting of heterogeneous team of autonomous unmanned aircraft; (ii) a system for verification of collision avoidance methods among fixed wing UAVs; and (iii) a system for planning collision-free paths for light-sport aircraft.

Neural based obstacle avoidance with CPG controlled hexapod walking robot

  • Autoři: Čížek, P., Milička, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the International Joint Conference on Neural Networks. IEEE Xplore, 2017. p. 650-656. ISSN 2161-4393. ISBN 978-1-5090-6181-5.
  • Rok: 2017
  • DOI: 10.1109/IJCNN.2017.7965914
  • Odkaz: https://doi.org/10.1109/IJCNN.2017.7965914
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this work, we are proposing a collision avoidance system for a hexapod crawling robot based on the detection of intercepting objects using the Lobula giant movement detector (LGMD) connected directly to the locomotion control unit based on the Central pattern generator (CPG). We have designed and experimentally verified the proposed approach that maps the output of the LGMD directly on the locomotion control parameters of the CPG. The results of the experimental verification of the system with real mobile hexapod crawling robot support the feasibility of the proposed approach in collision avoidance scenarios.

On Close Enough Orienteering Problem with Dubins Vehicle

  • DOI: 10.1109/IROS.2017.8206453
  • Odkaz: https://doi.org/10.1109/IROS.2017.8206453
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    In this paper, we address a generalization of the Orienteering Problem (OP) for curvature-constrained vehicles and to problems where it is allowed to collect a reward associated to each target location within a specified distance from the target. The addressed problem combines challenges of the combinatorial optimization of the OP (to select the most rewarding targets and find the optimal sequence to visit them) with the continuous optimization related to the determination of the waypoint locations and suitable headings at the waypoints for the considered Dubins vehicle such that the curvature-constrained path does not exceed the given travel budget and the sum of the collected rewards is maximized. The proposed generalization is called the Close Enough Dubins Orienteering Problem (CEDOP) and novel unsupervised learning approach is proposed to address computational requirements of this challenging planning problem. Based on the presented results, the proposed approach is feasible and provides a bit worse solution of CEDOP than the existing combinatorial approach but with significantly lower computational requirements.

On Self-Organizing Maps for Orienteering Problems

  • Autoři: prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the International Joint Conference on Neural Networks. IEEE Xplore, 2017. p. 2611-2620. ISSN 2161-4393. ISBN 978-1-5090-6181-5.
  • Rok: 2017
  • DOI: 10.1109/IJCNN.2017.7966175
  • Odkaz: https://doi.org/10.1109/IJCNN.2017.7966175
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper concerns principles of unsupervised learning of self-organizing maps (SOMs) to address optimization routing problems called the Orienteering Problem (OP) and its multi-vehicle variant called the Team Orienteering Problem (TOP). The problems are similar to the traveling salesman problem in finding an optimal tour to visit all the given locations, but here, each location has specified reward that can be collected by the tour and the problem is to select the most valuable subset of the locations that can be visited within the travel budget. In existing SOM for the OP, the locations to be visited are duplicated to adapt the network to locations with higher rewards more frequently. The proposed novel SOM-based solution overcomes this necessity and based on the presented results it significantly reduces the computational burden of the adaptation procedure. Besides, the proposed approach improves the quality of solutions and makes SOM competitive to existing heuristics for the OP, but still behind computationally expensive metaheuristics for the TOP. On the other hand, the main benefit of the SOM-based approaches over the existing heuristics is in solving the generalized variant of the OP and TOP with neighborhoods. These variants of the problem formulation allow to better utilize the travel budget for instances where the reward associated with the location can be collected by visiting a particular neighborhood of the location and not exactly the location itself. This generalized problem formulation better models situations of the robotic data collection, e.g., using wireless communication or range sensors.

On solution of the Dubins touring problem

  • DOI: 10.1109/ECMR.2017.8098685
  • Odkaz: https://doi.org/10.1109/ECMR.2017.8098685
  • Pracoviště: Centrum umělé inteligence, Multirobotické systémy
  • Anotace:
    The Dubins traveling salesman problem (DTSP) combines the combinatorial optimization of the optimal sequence of waypoints to visit the required target locations with the continuous optimization to determine the optimal headings at the waypoints. Existing decoupled approaches to the DTSP are based on an independent solution of the sequencing part as the Euclidean TSP and finding the optimal headings of the waypoints in the sequence. In this work, we focus on the determination of the optimal headings in a given sequence of waypoints and formulate the problem as the Dubins touring problem (DTP). The DTP can be solved by a uniform sampling of possible headings; however, we propose a new informed sampling strategy to find approximate solution of the DTP. Based on the presented results, the proposed algorithm quickly converges to a high-quality solution, which is less than 0.1% from the optimum. Besides, the proposed approach also improves the solution of the DTSP, and its feasibility has been experimentally verified in a real practical deployment.

RNN-based Visual Obstacle Avoidance with a CPG Controlled Hexapod Walking Robot

  • Autoři: Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS). Piscataway: IEEE, 2017. pp. 3146. ISSN 2153-0858. ISBN 978-1-5386-2682-5.
  • Rok: 2017
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    Avoiding collisions with obstacles and intercepting objects based on the visual perception is a vital survival ability of many animals. For a mobile robot moving from one place to another, the contact with a fixed or moving object may have fatal consequences. Therefore, collision avoidance skills of animals may be very useful also for mobile robots.

Self-organizing map for orienteering problem with dubins vehicle

  • Autoři: prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: 2017 12th International Workshop on Self-Organizing Maps and Learning Vector Quantization, Clustering and Data Visualization (WSOM). Marseille: IEEE, 2017. p. 125-132. ISBN 978-1-5090-6638-4.
  • Rok: 2017
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper reports on the application of the self-organizing map (SOM) to solve a novel generalization of the Orienteering Problem (OP) for curvature-constrained vehicles that is called the Dubins Orienteering Problem (DOP). Having a set of target locations, each with associated reward, and a given travel budget, the problem is to find the most valuable curvature-constrained path connecting the target locations such that the path does not exceed the travel budget. The proposed approach is based on two existing SOM-based approaches to solving the OP and Dubins Traveling Salesman Problem (Dubins TSP) that are further generalized to provide a solution of the more computational challenging DOP. DOP combines challenges of the combinatorial optimization of the OP and TSP to determine a subset of the most valuable targets and the optimal sequence of the waypoints to collect rewards of the targets together with the continuous optimization of determining headings of Dubins vehicle at the waypoints such that the total length of the curvature-constrained path is shorter than the given travel budget and the total sum of the collected rewards is maximized.

System for deployment of groups of unmanned micro aerial vehicles in GPS-denied environments using onboard visual relative localization

  • DOI: 10.1007/s10514-016-9567-z
  • Odkaz: https://doi.org/10.1007/s10514-016-9567-z
  • Pracoviště: Katedra kybernetiky, Centrum umělé inteligence
  • Anotace:
    A complex system for control of swarms of micro aerial vehicles (MAV), in literature also called as unmanned aerial vehicles (UAV) or unmanned aerial systems (UAS), stabilized via an onboard visual relative localization is described in this paper. The main purpose of this work is to verify the possibility of self-stabilization of multi-MAV groups without an external global positioning system. This approach enables the deployment of MAV swarms outside laboratory conditions, and it may be considered an enabling technique for utilizing fleets of MAVs in real-world scenar- ios. The proposed visual-based stabilization approach has been designed for numerous different multi-UAV robotic applications (leader-follower UAV formation stabilization, UAVswarmstabilizationanddeploymentinsurveillancesce- narios, cooperative UAV sensory measurement) in this paper. Deployment of the system in real-world scenarios truthfully verifies its operational constraints, given by limited onboard sensing suites and processing capabilities. The performance of the presented approach (MAV control, motion planning, MAV stabilization, and trajectory planning) in multi-MAV applications has been validated by experimental results in indoor as well as in challenging outdoor environments (e.g., in windy conditions and in a former pit mine).

Unsupervised learning for surveillance planning with team of aerial vehicles

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Váňa, P.
  • Publikace: Proceedings of the International Joint Conference on Neural Networks. IEEE Xplore, 2017. p. 4340-4347. ISSN 2161-4393. ISBN 978-1-5090-6181-5.
  • Rok: 2017
  • DOI: 10.1109/IJCNN.2017.7966405
  • Odkaz: https://doi.org/10.1109/IJCNN.2017.7966405
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we extent an existing self-organizing map (SOM)-based approach for the Dubins traveling salesman problem (DTSP) to solve its multi-vehicle variant generalized for visiting target regions called k-DTSP with Neighborhoods (k-DTSPN). The Dubins TSP is a variant of the combinatorial TSP for curvature-constrained vehicles. The problem is to determine a cost efficient path to visit a given set of continuous regions while the path allows to satisfy kinematic constraints of non-holonomic vehicles. The k-DTSPN is a generalization to determine k such paths, one for each vehicle. Although the k-DTSPN has been addressed by evolutionary methods, the proposed approach is able to provide solutions very quickly in units of seconds on conventional computationally resources which makes the proposed SOM-based approach suitable for on-line planning. The studied problem is motivated by surveillance task in which it is required to quickly provide information about the given set of target locations. Therefore, real computational requirements are crucial properties of the desired k-DTSPN solver. The proposed method meets this requirement and feasibility of the found solutions are demonstrated not only in computer simulations but also with a practical deployment on real aerial vehicles.

An Application of Self-Organizing Map for Multirobot Multigoal Path Planning with Minmax Objective

  • DOI: 10.1155/2016/2720630
  • Odkaz: https://doi.org/10.1155/2016/2720630
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, Self-Organizing Map (SOM) for the Multiple Traveling Salesman Problem (MTSP) with minmax objective is applied to the robotic problem of multigoal path planning in the polygonal domain. The main difficulty of such SOM deployment is determination of collision-free paths among obstacles that is required to evaluate the neuron-city distances in the winner selection phase of unsupervised learning. Moreover, a collision-free path is also needed in the adaptation phase, where neurons are adapted towards the presented input signal (city) to the network. Simple approximations of the shortest path are utilized to address this issue and solve the roboticMTSP by SOM. Suitability of the proposed approximations is verified in the context of cooperative inspection, where cities represent sensing locations that guarantee to "see" the whole robots' workspace. The inspection task formulated as the MTSP-Minmax is solved by the proposed SOM approach and compared with the combinatorial heuristic GENIUS. The results indicate that the proposed approach provides competitive results to GENIUS and support applicability of SOM for robotic multigoal path planning with a group of cooperating mobile robots. The proposed combination of approximate shortest paths with unsupervised learning opens further applications of SOM in the field of robotic planning.

Evolution of multiple gaits for modular robots

  • DOI: 10.1109/SSCI.2016.7850182
  • Odkaz: https://doi.org/10.1109/SSCI.2016.7850182
  • Pracoviště: Katedra kybernetiky, Centrum umělé inteligence
  • Anotace:
    Modular robots are composed of many elementary mechatronic modules that can be connected to form a robot body of various shapes. This feature allows such a robot to adapt for a given task and particular environment. A motion of the modular robot is based on control of individual angles between the modules, and the robot locomotion can be realized using Central Pattern Generators (CPG). A robot motion in the environment with obstacles can be achieved using several locomotion controllers that are switched by a strategy based on motion planning techniques. Preparation of CPG-based gaits leads to a high-dimensional optimization that requires to design proper cost functions. Existing approaches optimize the gaits separately according to human-designed cost functions. In this paper, we investigate how to automatically derive a set of gaits suitable for modular robots without specifying low-level details about the gaits. We propose to optimize multiple gaits simultaneously using a single cost function. This cost function is based on the ability of motion planning to solve the task using the gaits being optimized. The proposed system is verified on several modular robots with unusual shapes including robots with failed modules.

Low-Latency Image Processing for Vision-Based Navigation Systems

  • Autoři: Čížek, P., prof. Ing. Jan Faigl, Ph.D., Masri, D.
  • Publikace: IEEE International Conference on Robotics and Automation. Budapešť: Institute of Electrical and Electronics Engineers Inc., 2016. p. 781-786. ISSN 1050-4729. ISBN 978-1-4673-8026-3.
  • Rok: 2016
  • DOI: 10.1109/ICRA.2016.7487207
  • Odkaz: https://doi.org/10.1109/ICRA.2016.7487207
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper concerns a problem of the latency reduction in the vision-based mobile robot navigation, which is considered as the crucial system property to determine a control command based on visual data in practical deployments of mobile robots. The problem is addressed by a processor centric FPGA-based System-on-Chip design allowing power and computationally efficient on-line image processing. The proposed architecture is considered in an autonomous vision-based navigation with a teach-and-repeat algorithm based on detection and tracking of image salient points. The architecture has been evaluated and compared with a CPU-based solution on different platforms and the results indicate that the proposed FPGA-based implementation outperforms pure CPU solutions in the overall latency, speed, and power consumption.

Multi-Robot Path Planning for Budgeted Active Perception with Self-Organising Maps

  • Autoři: Best, G., prof. Ing. Jan Faigl, Ph.D., Fitch, R.
  • Publikace: Intelligent Robots and Systems (IROS), 2016 IEEE/RSJ International Conference on. Piscataway: IEEE, 2016. p. 3164-3171. ISSN 2153-0866. ISBN 978-1-5090-3762-9.
  • Rok: 2016
  • DOI: 10.1109/IROS.2016.7759489
  • Odkaz: https://doi.org/10.1109/IROS.2016.7759489
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    We propose a self-organising map (SOM) algorithm as a solution to a new multi-goal path planning problem for active perception and data collection tasks. We optimise paths for a multi-robot team that aims to maximally observe a set of nodes in the environment. The selected nodes are observed by visiting associated viewpoint regions defined by a sensor model. The key problem characteristics are that the viewpoint regions are overlapping polygonal continuous regions, each node has an observation reward, and the robots are constrained by travel budgets. The SOM algorithm jointly selects and allocates nodes to the robots and finds favourable sequences of sensing locations. The algorithm has polynomial-bounded runtime independent of the number of robots. We demonstrate feasibility for the active perception task of observing a set of 3D objects. The viewpoint regions consider sensing ranges and self-occlusions, and the rewards are measured as discriminability in the ensemble of shape functions feature space. Simulations were performed using a 3D point cloud dataset from a real robot in a large outdoor environment. Our results show the proposed methods enable multi-robot planning for budgeted active perception tasks with continuous sets of candidate viewpoints and long planning horizons.

On Chaotic Oscillator-based Central Pattern Generator for Motion Control of Hexapod Walking Robot

  • Autoři: Milička, P., Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: ITAT 2016: Information Technologies - Applications and Theory: Conference on Theory and Practice of Information Technologies. Luxemburg: CreateSpace Independent Publishing Platform, 2016. p. 131-137. ISBN 978-1-5370-1674-0.
  • Rok: 2016
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we address a problem of motion control of a real hexapod walking robot along a trajectory of the prescribed curvature and desired motion gait. The proposed approach is based on a chaotic neural oscilátor that is employed as the central pattern generator (CPG). The CPG allows to generate various motion gaits according to the specified period of the chaotic oscillator. The output signal of the oscillator is processed by the proposed trajectory generator that allows to specify a curvature of the trajectory the robot is requested to traverse. Such a signal is then considered as an input for the inverse kinematic task which provides particular trajectories of individual legs that are directly send to the robot actuators. Thus, the main benefit of the proposed approach is that only two natural parameters are necessary to control the gait type and the robot motion. The proposed approach has been verified in real experiments. The experimental results support feasibility of the proposed concept and the robot is able to crawl desired trajectories with the tripod, ripple, low gear, and wave motion gaits.

On Construction of a Reliable Ground Truth for Evaluation of Visual SLAM Algorithms

  • Autoři: Ing. Jan Bayer, Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Acta Polytechnica CTU Proceedings. Praha: Česká technika - nakladatelství ČVUT, 2016. pp. 1-5. ISSN 2336-5382. ISBN 978-80-01-06022-3.
  • Rok: 2016
  • DOI: 10.14311/APP.2016.6.0001
  • Odkaz: https://doi.org/10.14311/APP.2016.6.0001
  • Pracoviště: Katedra počítačů, Centrum umělé inteligence
  • Anotace:
    In this work we are concerning the problem of localization accuracy evaluation of visual-based Simultaneous Localization and Mapping (SLAM) techniques. Quantitative evaluation of the SLAM algorithm performance is usually done using the established metrics of Relative pose error and Absolute trajectory error which require a precise and reliable ground truth. Such a ground truth is usually hard to obtain, while it requires an expensive external localization system. In this work we are proposing to use the SLAM algorithm itself to construct a reliable ground-truth by offline frame-by-frame processing. The generated ground-truth is suitable for evaluation of different SLAM systems, as well as for tuning the parametrization of the on-line SLAM. The presented practical experimental results indicate the feasibility of the proposed approach.

On Evaluation of Motion Gaits Energy Efficiency with a Hexapod Crawling Robot

  • Autoři: Černý, L., Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Acta Polytechnica CTU Proceedings. Praha: Česká technika - nakladatelství ČVUT, 2016. pp. 6-10. ISSN 2336-5382. ISBN 978-80-01-06022-3.
  • Rok: 2016
  • DOI: 10.14311/APP.2016.6.0006
  • Odkaz: https://doi.org/10.14311/APP.2016.6.0006
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this work, we are concerning the problem of energy efficient locomotion of a hexapod crawling robot. We are emphasizing a practical verification and deployment on a real walking robot to evaluate relations between the energy consumption, motion speed, and terrain type with a particular motion gait. The tripod, tetrapod, and pentapod motion gaits are considered in the presented evaluation report.

On Localization and Mapping with RGB-D Sensor and Hexapod Walking Robot in Rough Terrains

  • Autoři: Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of 2016 IEEE International Conference on Systems, Man, and Cybernetics. IEEE, 2016. ISBN 978-1-5090-1897-0.
  • Rok: 2016
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we address a problem of precise online localization of a hexapod walking robot operating in rough terrains. We consider an existing Simultaneous Localization and Mapping approach with a low cost structured light (RGB-D) sensor. We propose to combine this sensor and localization method with the developed adaptive motion gait that allows the robot to crawl various types of terrain, such as stairs, ramps, or small wooden blocks. Such an environment requires a full 6- DOF pose estimation to create a map of the robot surroundings and allows us to asses impact of the individual terrain types and influence of the SLAM method parametrization on the localization accuracy. The reported evaluation results indicate the relations between the terrain type, parametrization of the method and the localization accuracy.

On Self-Organizing Map and Rapidly-Exploring Random Graph in Multi-Goal Planning

  • Autoři: prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of the 11th Workshop on Self-Organizicng Maps - Advances in Intelligent Systems and Computing. Heidelberg: Springer, 2016. pp. 143-153. ISSN 2194-5357. ISBN 978-3-319-28517-7.
  • Rok: 2016
  • DOI: 10.1007/978-3-319-28518-4_12
  • Odkaz: https://doi.org/10.1007/978-3-319-28518-4_12
  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper reports on ongoing work towards an extension of the self-organizing maps for the traveling salesman problem to more challenging problems of multi-goal trajectory planning for complex robots with a high-dimensional configuration space. The main challenge of this problem is that the distance function needed to find a sequence of the visits to the goals is not known a priori and it is not easy to compute. To address this challenge, we propose to utilize the unsupervised learning in a trade-off between the exploration of the distance function and exploitation of its current model. The proposed approach is based on steering the sampling process in a randomized sampling-based motion planning technique to create a suitable motion planning roadmap, which represents the required distance function. The presented results shows the proposed approach quickly provides an admissible solution, which may be further improved by additional samples of the configuration space

Random Inspection Tree Algorithm in Visual Inspection with a Realistic Sensing Model and Differential Constraints

  • Autoři: Kafka, P., prof. Ing. Jan Faigl, Ph.D., Váňa, P.
  • Publikace: IEEE International Conference on Robotics and Automation. Budapešť: Institute of Electrical and Electronics Engineers Inc., 2016. p. 2782-2787. ISSN 1050-4729. ISBN 978-1-4673-8026-3.
  • Rok: 2016
  • DOI: 10.1109/ICRA.2016.7487440
  • Odkaz: https://doi.org/10.1109/ICRA.2016.7487440
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we consider existing asymptotically optimal inspection planning algorithm in coverage path planning with realistic visibility constraints of standard cameras. Although the existing approach is able to provide an optimal solution with omnidirectional sensing and limited sensing range, it is prohibitively computationally expensive for problems with only few objects to be covered and limited field of view. Based on the analysis of the utilized sampling-based strategy, we propose a heuristic approach to decrease computational requirements in problems with restricted viewing frustum, which is a more realistic model of a digital camera. In addition, we also consider a minimal distance and angle under which the object to be covered is captured by the forward looking camera to make a snapshot of the object with the required details.

Road Following with Blind Crawling Robot

  • Autoři: Stejskal, M., Mrva, J., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: IEEE International Conference on Robotics and Automation. Budapešť: Institute of Electrical and Electronics Engineers Inc., 2016. p. 3612-3617. ISSN 1050-4729. ISBN 978-1-4673-8026-3.
  • Rok: 2016
  • DOI: 10.1109/ICRA.2016.7487544
  • Odkaz: https://doi.org/10.1109/ICRA.2016.7487544
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we consider road following to autonomously navigate a mobile robot through an environment while keeping the robot on the specified road. Contrary to existing approaches based on a forward looking camera, we consider the problem for a technically blind walking robot without any exteroceptive sensors. The only feedback considered is an estimation of tactile information that is determined from the robot servo drives. The proposed control law is based on an on-line classification of the previously learned terrains which is utilized to identify a situation when a robot starts to crawl off the desired road terrain. The controller steers the robot to keep its body and all its legs on the road while crawling forward with a constant velocity. The experimental results support feasibility of the proposed minimalistic approach and allows the robot to autonomously navigate in an outdoor environment and follow urban park pathways and avoid off-road parts.

Self-Organizing Map for Data Collection Planning in Persistent Monitoring with Spatial Correlations

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Váňa, P.
  • Publikace: Proceedings of 2016 IEEE International Conference on Systems, Man, and Cybernetics. IEEE, 2016. ISBN 978-1-5090-1897-0.
  • Rok: 2016
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper introduces an extension of the unsupervised learning method to solve data collection planning problems where particular sensor measurements can be spatially correlated. The problem is motivated by monitoring tasks formulated as the Prize-Collecting Traveling Salesman Problem with Neighborhoods (PC-TSPN). A solution of the PC-TSPN consists of a selection of important sensors, determination of the locations to read data from these sensors, and finding the shortest path to visit the locations. The solution cost is defined as a sum of the travel cost and penalty characterizing additional cost associated to sensors from which data are not retrieved. The penalty represents importance of particular sensor measurements to the quality of the model and existing solutions assume the penalties are constant values. However, for spatially close sensor locations, data from one sensor may contain also information about nearby locations and thus, its penalty depends on locations selected for data collection. The proposed generalization of the PC-TSPN solver allows to consider spatial correlations of sensor measurements and the proposed approach provides better solutions than the previous algorithm with fixed penalties.

Self-Organizing Map for the Curvature-Constrained Traveling Salesman Problem

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Váňa, P.
  • Publikace: ARTIFICIAL NEURAL NETWORKS AND MACHINE LEARNING - ICANN 2016, PT II. Düsseldorf: Springer VDI Verlag, 2016. pp. 497-505. ISSN 0302-9743. ISBN 978-3-319-44780-3.
  • Rok: 2016
  • DOI: 10.1007/978-3-319-44781-0_59
  • Odkaz: https://doi.org/10.1007/978-3-319-44781-0_59
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we consider a challenging variant of the traveling salesman problem (TSP) where it is requested to determine the shortest closed curvature-constrained path to visit a set of given locations. The problem is called the Dubins traveling salesman problem in literature and its main difficulty arises from the fact that it is necessary to determine the sequence of visits to the locations together with particular headings of the vehicle at the locations. We propose to apply principles of unsupervised learning of the self-organizing map to simultaneously determine the sequence of the visits together with the headings. A feasibility of the proposed approach is supported by an extensive evaluation and comparison to existing solutions. The presented results indicate that the proposed approach provides competitive solutions to existing heuristics, especially in dense problems, where the optimal sequence of the visits cannot be determined as a solution of the Euclidean TSP.

Self-Organizing Map-based Solution for the Orienteering Problem with Neighborhoods

  • Pracoviště: Katedra kybernetiky, Centrum umělé inteligence
  • Anotace:
    In this paper, we address the Orienteering problém (OP) by the unsupervised learning of the self-organizing map (SOM). We propose to solve the OP with a new algorithm based on SOM for the Traveling salesman problem (TSP). Both problems are similar in finding a tour visiting the given locations; however, the OP stands to determine the most valuable tour that maximizes the rewards collected by visiting a subset of the locations while keeping the tour length under the specified travel budget. The proposed stochastic search algorithm is based on unsupervised learning of SOM and it constructs a feasible solution during each learning epoch. The reported results support feasibility of the proposed idea and show the performance is competitive with existing heuristics. Moreover, the key advantage of the proposed SOM-based approach is the ability to address the generalized OP with Neighborhoods, where rewards can be collected by traveling anywhere within the neighborhood of the locations. This problem generalization better fits data collection missions with wireless data transmission and it allows to save unnecessary travel costs to visit the given locations.

Stereo Vision-Based Localization for Hexapod Walking Robots Operating in Rough Terrains

  • Autoři: Fischer, T., Pire, T., Čížek, P., Cristóforis, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Intelligent Robots and Systems (IROS), 2016 IEEE/RSJ International Conference on. Piscataway: IEEE, 2016. p. 2492-2497. ISSN 2153-0866. ISBN 978-1-5090-3762-9.
  • Rok: 2016
  • DOI: 10.1109/IROS.2016.7759388
  • Odkaz: https://doi.org/10.1109/IROS.2016.7759388
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    This paper concerns the self-localization problem of a hexapod walking robot operating in rough terrains. Given that legged robots exhibit higher terrain passability than wheeled or tracked platforms when operating in harsh environments, they constitute a challenge for the localization techniques because the camera motion between consecutive frames can be arbitrary due to the motion gait and terrain irregularities. In this paper, we present and evaluate an inertially assisted Stereo Parallel Tracking and Mapping (S-PTAM) method deployed on a hexapod crawling robot in a rough terrain. The considered deployment scenario is motivated by autonomous navigation in an unknown environment in an open loop fashion. The reported results and comparison with an existing RGB-D SLAM technique show the feasibility of the proposed approach and its suitability for navigation of crawlers in harsh environments.

The Dubins Traveling Salesman Problem with Constrained Collecting Maneuvers

  • Autoři: Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Acta Polytechnica CTU Proceedings. Praha: Česká technika - nakladatelství ČVUT, 2016. pp. 34-39. ISSN 2336-5382. ISBN 978-80-01-06022-3.
  • Rok: 2016
  • DOI: 10.14311/APP.2016.6.0034
  • Odkaz: https://doi.org/10.14311/APP.2016.6.0034
  • Pracoviště: Centrum umělé inteligence
  • Anotace:
    In this paper, we introduce a variant of the Dubins traveling salesman problem (DTSP) that is called the Dubins traveling salesman problem with constrained collecting maneuvers (DTSP-CM). In contrast to the ordinary formulation of the DTSP, in the proposed DTSP-CM, the vehicle is requested to visit each target by specified collecting maneuver to accomplish the mission. The proposed problem formulation is motivated by scenarios with unmanned aerial vehicles where particular maneuvers are necessary for accomplishing the mission, such as object dropping or data collection with sensor sensitive to changes in vehicle heading. We consider existing methods for the DTSP and propose its modifications to use these methods to address a variant of the introduced DTSP-CM, where the collecting maneuvers are constrained to straight line segments.

Agent-based Approach to Illegal Maritime Behavior Modeling

  • DOI: 10.17402/026
  • Odkaz: https://doi.org/10.17402/026
  • Pracoviště: Katedra počítačů
  • Anotace:
    Maritime shipping is a set of complex activities with a large number of actors involved. We focus on a subset of illegal maritime activities, such as armed robberies, maritime piracy or contraband smuggling. To fight against them and minimize their negative impact naval authorities typically introduce a number of countermeasures, such as deployed patrols or surveillance agents. Due to very high costs of countermeasures it is often beneficial to evaluate their impact using a simulation, allowing what-if analysis and evaluation of a range of scenarios before actually deploying the countermeasures. We introduce BANDIT, an agent-based computational platform, which is designed to evaluate scenarios with an accent on the modeling of different types of illegal behavior and on the interaction between agents. The platform consists of an agent behavior modeling system and a multi-agent maritime simulator. The platform allows the definition of a number of scenarios through a simple configuration and it offers the means to run these scenarios in a single or a batch mode and evaluate the results as single or aggregate data sets respectively. We demonstrate the usefulness of the platform on the scenarios of the drug smuggling problem in the seas surrounding Central America. Senario outcomes (e.g., heatmaps of activities, set of trajectories etc.) are subsequently used to help with the design of effective countermeasures, i.e., allocating naval patrols and planning their patrol routes.

Bi-objective maritime route planning in pirate-infested waters

  • DOI: 10.17402/046
  • Odkaz: https://doi.org/10.17402/046
  • Pracoviště: Katedra počítačů
  • Anotace:
    Contemporary maritime shipping is subject to a large number of constraints given by tight shipping schedules and very low margins. Additionally, problematic areas with increased security needs dynamically changing in time, combined with seasonal oceanographic and meteorological conditions pose a challenging voyage planning problem. In this work we present a risk-aware voyage planner taking into account spatio-temporal environmental conditions. The planner is based on a graph-based search algorithm A*. We discretize the required area into a graph, we store various layers of information into the edges of the graph (such as risk and weather conditions) in a form of numeric weights and we define a bi-objective planning problem with a tradeoff between security and duration of the voyage. The nature of the algorithm guarantees a complete and optimal solution in a form of an optimized voyage with respect to the criterion function composed of the two weighted components, i.e, duration and security of the voyage. We demonstrate the approach on our area of interest: Indian Ocean. We use NATO piracy activity risk surface as the risk layer and we compute all transit voyages between relevant routing points in the area. Finally, thanks to the discretization of the problem, we are able to integrate corridors imposed by the shipping authorities and evaluate additional what-if scenarios with extended corridor systems. The resulting planner is exposed to the public using a web service with an easy interface requiring start time of the voyage and the origin and the destination point of voyage. Combined with an expressive visualization, this tool demonstrates the capabilities of the proposed solution.

Comparison of Task-Allocation Algorithms in Frontier-Based Multi-robot Exploration

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Simonin, O., Charpillet, F.
  • Publikace: Multi-Agent Systems: 12th European Conference on Multi-Agent Systems. Heidelberg: Springer, 2015. p. 101-110. ISSN 0302-9743. ISBN 978-3-319-17129-6.
  • Rok: 2015
  • DOI: 10.1007/978-3-319-17130-2_7
  • Odkaz: https://doi.org/10.1007/978-3-319-17130-2_7
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we address the problem of efficient allocation of the navigational goals in the multi-robot exploration of unknown environment. Goal candidate locations are repeatedly determined during the exploration. Then, the assignment of the candidates to the robots is solved as the task-allocation problem. A more frequent decision-making may improve performance of the exploration, but in a practical deployment of the exploration strategies, the frequency depends on the computational complexity of the task-allocation algorithm and available computational resources. Therefore, we propose an evaluation framework to study exploration strategies independently on the available computational resources and we report a comparison of the selected task-allocation algorithms deployed in multi-robot exploration.

Feature Extraction for Terrain Classification with Crawling Robots

  • Autoři: Mrva, J., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: ITAT 2015 conference proceedings. Aachen: CEUR Workshop Proceedings, 2015, pp. 179-185. ISSN 1613-0073. ISBN 978-1-5151-2065-0. Available from: http://ceur-ws.org/Vol-1422/179.pdf
  • Rok: 2015
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we address the problem of terrain classification using a technically blind hexapod walking robot. The proposed approach is built on top of the existing method based on analysis of the feedback from the robot’s actuators and the desired trajectory. The formed method uses features for the Support Vector Machine classification method that assumes a regular time-invariant gait to control the robot. However, such a gait does not allow the robot to traverse rough terrains, and therefore, it is necessary to consider adaptive motion gait to deal with small obstacles, which is, unfortunately, not a regular gait with some fixed predefined period. Therefore, we propose to alter the features extraction process to utilize the terrain classification method also for an adaptive motion gait, which enables the robot to traverse rough terrains. The proposed method has been experimentally verified on several terrains that are not traversable by a default regular gait. The achieved results not only confirmed the high accuracy of the terrain classification as the existing approach, but also expanded the area of operation of a hexapod walking robot into more challenging terrains.

Impact Assessment of Image Feature Extractors on the Performance of SLAM Systems

  • Autoři: Pire, T., Fischer, T., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: ACTA POLYTECHNICA CTU PROCEEDINGS: PAIR'15 Student Conferences on Planning in Artificial Intelligence and Robotics. Praha: České vysoké učení technické v Praze, 2015, pp. 45-50. ISSN 2336-5382. Available from: https://ojs.cvut.cz/ojs/index.php/APP/article/view/APP.2015.1.0045/3188
  • Rok: 2015
  • DOI: 10.14311/APP.2015.1.0045
  • Odkaz: https://doi.org/10.14311/APP.2015.1.0045
  • Pracoviště: Katedra počítačů
  • Anotace:
    This work evaluates an impact of image feature extractors on the performance of a visual SLAM method in terms of pose accuracy and computational requirements. In particular, the S-PTAM (Stereo Parallel Tracking and Mapping) method is considered as the visual SLAM framework for which both the feature detector and feature descriptor are parametrized. The evaluation was performed with a standard dataset with ground-truth information and six feature detectors and four descriptors. The presented results indicate that the combination of the GFTT detector and the BRIEF descriptor provides the best trade-off between the localization precision and computational requirements among the evaluated combinations of the detectors and descriptors.

On FPGA Based Acceleration of Image Processing in Mobile Robotics

  • Autoři: Čížek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: ACTA POLYTECHNICA CTU PROCEEDINGS: PAIR'15 Student Conferences on Planning in Artificial Intelligence and Robotics. Praha: České vysoké učení technické v Praze, 2015, pp. 8-14. ISSN 2336-5382. Available from: https://ojs.cvut.cz/ojs/index.php/APP/article/view/APP.2015.1.0008
  • Rok: 2015
  • DOI: 10.14311/APP.2015.1.0008
  • Odkaz: https://doi.org/10.14311/APP.2015.1.0008
  • Pracoviště: Katedra počítačů
  • Anotace:
    In visual navigation tasks, a lack of the computational resources is one of the main limitations of micro robotic platforms to be deployed in autonomous missions. It is because the most of nowadays techniques of visual navigation relies on a detection of salient points that is computationally very demanding. In this paper, an FPGA assisted acceleration of image processing is considered to overcome limitations of computational resources available on-board and to enable high processing speeds while it may lower the power consumption of the system. The paper reports on performance evaluation of the CPU-based and FPGA-based implementations of a visual teach-and-repeat navigation system based on detection and tracking of the FAST image salient points. The results indicate that even a computationally efficient FAST algorithm can benefit from a parallel (low-cost) FPGA-based implementation that has a competitive processing time but more importantly it is a more power efficient.

On sampling based methods for the Dubins Traveling Salesman Problem with Neighborhoods

  • Autoři: Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: ACTA POLYTECHNICA CTU PROCEEDINGS: PAIR'15 Student Conferences on Planning in Artificial Intelligence and Robotics. Praha: České vysoké učení technické v Praze, 2015, pp. 57-61. ISSN 2336-5382. Available from: https://ojs.cvut.cz/ojs/index.php/APP/article/view/3407
  • Rok: 2015
  • DOI: 10.14311/APP.2015.1.0057
  • Odkaz: https://doi.org/10.14311/APP.2015.1.0057
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we address the problem of path planning to visit a set of goal regions by Dubins vehicle, which is also known as the Dubins Traveling Salesman Problem with Neighborhoods (DTSPN). We propose a modification of the existing sampling-based approach to use an increasing number of samples per goal region and thus improve the solution quality if a more computational time is available. The performance of the proposed modified sampling-based algorithm has been compared with existing approaches for the DTSPN and results of the quality of found solutions and the required computational time are presented.

On the Dubins Traveling Salesman Problem with Neighborhoods

  • Autoři: Váňa, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: IROS 2015: Proceedings IEEE/RSJ International Conference on Inteligent Robots and Systems. Los Alamitos: IEEE Computer Society, 2015. pp. 4029-4034. ISSN 2153-0858. ISBN 978-1-4799-9994-1.
  • Rok: 2015
  • DOI: 10.1109/IROS.2015.7353945
  • Odkaz: https://doi.org/10.1109/IROS.2015.7353945
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we address the problem of optimal path planning to visit a set of regions by Dubins vehicle, which is also known as the Dubins Traveling Salesman Problem with Neighborhoods (DTSPN). This problem can be tackled by a transformation to other variants of the TSP or evolutionary algorithms. We address the DTSPN as a problem to find Dubins path to visit a given sequence of regions and propose a simple iterative optimization procedure to find Dubins path visiting the regions. The proposed approach allows to efficiently solve the DTSPN and based on the presented comparison with existing approaches, the proposed algorithm provides solutions of competitive quality to the evolutionary techniques while it is significantly less computationally demanding.

On Traversability Cost Evaluation from Proprioceptive Sensing for a Crawling Robot

  • Autoři: Mrva, J., Stejskal, M., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: ACTA POLYTECHNICA CTU PROCEEDINGS: PAIR'15 Student Conferences on Planning in Artificial Intelligence and Robotics. Praha: České vysoké učení technické v Praze, 2015, pp. 1-6. ISSN 2336-5382.
  • Rok: 2015
  • Pracoviště: Katedra počítačů
  • Anotace:
    Traversability characteristics of the robot working environment are crucial in planning an efficient path for a robot operating in rough unstructured areas. In the literature, approaches to wheeled or tracked robots can be found, but a relatively little attention is given to walking multi-legged robots. Moreover, the existing approaches for terrain traversability assessment seem to be focused on gathering key features from a terrain model acquired from range data or camera image and only occasionally supplemented with proprioceptive sensing that expresses the interaction of the robot with the terrain. This paper addresses the problem of traversability cost evaluation based on proprioceptive sensing for a hexapod walking robot while optimizing different criteria. We present several methods of evaluating the robot-terrain interaction that can be used as a cost function for an assessment of the robot motion that can be utilized in high-level path-planning algorithms.

Reaction Diffusion Voronoi Diagrams: From Sensors Data to Computing

  • Autoři: Vázquez-Otero, A., prof. Ing. Jan Faigl, Ph.D., Dormido, R., Duro, N.
  • Publikace: Sensors - Open Access Journal. 2015, 15(6), 12736-12764. ISSN 1424-8220.
  • Rok: 2015
  • Pracoviště: Katedra počítačů
  • Anotace:
    n this paper, a new method to solve computational problems using reaction diffusion (RD) systems is presented. The novelty relies on the use of a model configuration that tailors its spatiotemporal dynamics to develop Voronoi diagrams (VD) as a part of the system's natural evolution. The proposed framework is deployed in a solution of related robotic problems, where the generalized VD are used to identify topological places in a grid map of the environment that is created from sensor measurements. The ability of the RD-based computation to integrate external information, like a grid map representing the environment in the model computational grid, permits a direct integration of sensor data into the model dynamics. The experimental results indicate that this method exhibits significantly less sensitivity to noisy data than the standard algorithms for determining VD in a grid. In addition, previous drawbacks of the computational algorithms based on RD models, like the generation of volatile solutions by means of excitable waves, are now overcome by final stable states.

Tactile Sensing with Servo Drives Feedback only for Blind Hexapod Walking Robot

  • Autoři: Mrva, J., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Proceedings of 10th International Workshop on Robot Motion and Control. Piscataway: IEEE, 2015. p. 240-245. ISBN 978-1-4799-7043-8.
  • Rok: 2015
  • DOI: 10.1109/RoMoCo.2015.7219742
  • Odkaz: https://doi.org/10.1109/RoMoCo.2015.7219742
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we address the problem of traversing rough terrains with hexapod walking robots. Although there can be found several approaches to deal with the terrain complexity, the proposed approach is strictly focused on a minimalist and cheap sensor equipment without additional inertial and exteroceptive sensors. The main idea of the proposed approach is to consider only the feedback from the intelligent servo drives to detect the contact point of a leg with the surface. During the leg motion, a relation of the joint torque and difference of the current and required joint positions is utilized to emulate a dedicated tactile sensor and thus the only equipment needed are the robot actuators. The proposed approach has been experimentally verified in a series of scenarios where a regular motion gait does not allow the robot to traverse the terrain while the proposed detection method enables a smooth motion of the technically blind robot in rough terrains of various difficulty.

A Practical Multirobot Localization System

  • DOI: 10.1007/s10846-014-0041-x
  • Odkaz: https://doi.org/10.1007/s10846-014-0041-x
  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    We present a fast and precise vision-based software intended for multiple robot localization. The core component of the software is a novel and efficient algorithm for black and white pattern detection. The method is robust to variable lighting conditions, achieves sub-pixel precision and its computational complexity is independent of the processed image size. With off-the-shelf computational equipment and low-cost cameras, the core algorithm is able to process hundreds of images per second while tracking hundreds of objects with millimeter precision. In addition, we present the method’s mathematical model, which allows to estimate the expected localization precision, area of coverage, and processing speed from the camera’s intrinsic parameters and hardware’s processing capacity. The correctness of the presented model and performance of the algorithm in real-world conditions is verified in several experiments. Apart from the method description, we also make its source code public at http://purl.org/robotics/whycon; so, it can be used as an enabling technology for various mobile robotic problems.

Multi-Goal Trajectory Planning with Motion Primitives for Hexapod Walking Robot

  • Autoři: Vaněk, P., prof. Ing. Jan Faigl, Ph.D., Masri, D.
  • Publikace: Proceedings of the 11th International Conference on Informatics in Control, Automation and Robotics. Porto: SciTePress - Science and Technology Publications, 2014, pp. 599-604. ISBN 978-989-758-040-6.
  • Rok: 2014
  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper presents our early results on multi-goal trajectory planning with motion primitives for a hexapod walking robot. We propose to use an on-line unsupervised learning method to simultaneously find a solution of the underlying traveling salesman problem together with particular trajectories between the goals. Using this technique, we avoid pre-computation of all possible trajectories between the goals for a graph based heuristic solvers for the traveling salesman problem. The proposed approach utilizes principles of self-organizing map to steer the randomized sampling of configuration space in promising areas regarding the multi-goal trajectory. The presented results indicate the proposed steering mechanism provides a feasible multi-goal trajectory in a less number of samples than an approach based on a priori known sequence of the goals visits.

Reaction-Diffusion based Computational Model for Autonomous Mobile Robot Exploration of Unknown Environments

  • Autoři: Vazquez-Otero, A., prof. Ing. Jan Faigl, Ph.D., Duro, N., Dormido, R.
  • Publikace: International Journal of Unconventional Computing. 2014, 10(4), 295-316. ISSN 1548-7199.
  • Rok: 2014
  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper introduces a computational model in which the main decision logic is based on principles arising from the dynamics of reaction-diffusion systems. The approach is an extension of our previous work where similar principles were used to develop a path planning algorithm. In this work, we select a mobile robot exploration task as a platform for exhibiting the core properties of the proposed computational framework. The functionalities represent particular building blocks that provide decision-logic capability of the exploration strategy. Beside a single mobile robot exploration, the proposed principles can also be generalized for multi-robot exploration, which is supported by the presented results.

Self-organizing map for determination of goal candidates in mobile robot exploration

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Vaněk, P., Kulich, M.
  • Publikace: Proceedings of the 22nd European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning. Louvain la Neuve: Ciaco - i6doc.com, 2014. pp. 589-594. ISBN 978-2-87419-095-7.
  • Rok: 2014
  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper addresses a problem of determining goal candidates in the frontier-based mobile robot exploration. The proposed solution is based on self-organizing map for the traveling salesman problem with neighborhoods and it allows to study the exploration formulated as a problem of repeated coverage of the current frontiers where the minimal number of goal candidates is determined simultaneously together with the expected cost to visit the candidates. The early results enabled by the proposed self-organizing map-based solution indicate exploration improvement for the proposed problem formulation. The presented work demonstrates how neural network approach can provide interesting insights and ground for studying optimizations problems arising in robotics.

Self-Organizing Map for the Prize-Collecting Traveling Salesman Problem

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Hollinger, G.A.
  • Publikace: Advances in Intelligent Systems and Computing. Düsseldorf: Springer VDI Verlag, 2014. pp. 281-291. ISSN 2194-5357. ISBN 978-3-319-07694-2.
  • Rok: 2014
  • DOI: 10.1007/978-3-319-07695-9_27
  • Odkaz: https://doi.org/10.1007/978-3-319-07695-9_27
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we propose novel adaptation rules for the self-organizing map to solve the prize-collecting traveling salesman problem (PC-TSP). The goal of the PC-TSP is to find a cost-efficient tour to collect prizes by visiting a subset of a given set of locations. In contrast with the classical traveling salesman problem, where all given locations must be visited, locations in the PC-TSP may be skipped at the cost of some additional penalty. Using the self-organizing map, locations for the final solution may be selected during network adaptation, and locations where visitation would be more expensive than their penalty can be avoided. We have applied the proposed self-organizing map learning procedure to autonomous data collection problems, where the proposed approach provides results competitive with an existing combinatorial solver. ˆ Springer International Publishing Switzerland 2014.

Unifying multi-goal path planning for autonomous data collection

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Hollinger, G.A.
  • Publikace: Proceedings of IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2014. p. 2937-2942. ISSN 2153-0858. ISBN 978-1-4799-6934-0.
  • Rok: 2014
  • DOI: 10.1109/IROS.2014.6942967
  • Odkaz: https://doi.org/10.1109/IROS.2014.6942967
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we propose a framework for solving variants of the multi-goal path planning problem with applications to autonomous data collection. Autonomous data collection requires optimizing the trajectory of a mobile vehicle to collect data from a number of stationary sensors in a known configuration. The proposed approach utilizes the self-organizing map (SOM) architecture to provide a unified solution to multi-goal path planning problems. Our approach applies to cases where the vehicle must move within a radius of a sensor to collect data and also where some sensors can be ignored due to a lower priority. We compare our proposed approach to state-of-the-art approximate solutions to variants of the Traveling Salesman Problem (TSP) for random deployments and in an underwater monitoring application domain. Our results demonstrate that the SOM approach outperforms combinatorial heuristic algorithms and also provides a unified approach for solving variants of the multi-goal path planning problem.

Unsupervised Learning of Growing Roadmap in Multi-Goal Motion Planning Problem

  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we address the multi-goal motion planning problem in which it is required to determine an order of visits of a pre-specified set of goals together with the shortest trajectories connecting the goals. The considered problem is inspired by inspection planning missions, where multiple goals must be visited with a required precision. The problem combines challenges of the combinatorial traveling salesman problem with difficulties of the motion planning. The presented approach is based on unsupervised learning of the self-organizing map technique for the traveling salesman problem applied in the configuration space. This learning technique takes an advantage of acquiring information about exploring the configuration space into a topology of the map that is simultaneously exploited in determination of the multi-goal trajectory and further directions of motion planning roadmap expansion. Presented results indicate that the proposed approach is feasible and it is able to provide a solution of the multi-goal motion planning problem without a priori known sequence of the goals visits.

A cooperative driver model for traffic simulations

  • DOI: 10.1109/INDIN.2013.6622979
  • Odkaz: https://doi.org/10.1109/INDIN.2013.6622979
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, a cooperative driver model for a multi-agent traffic simulation is proposed. The model combines maneuver-based trajectory planning of the vehicles with a cooperative conflict resolving. The proposed model is able to provide a safe drive in complex traffic situations at the highest possible speed. The idea of the model and its feasibility have been verified in complex scenarios such as line change under heavy traffic, highway entering or highway crossing. Moreover, the developed cooperative driver model is being integrated with a human operated driving simulator that enables verification of the proposed model in mixed scenarios enriching the simulation for a human driver with highly cooperative background traffic; thus, providing a platform for further studies on benefits of assistive technologies. The paper provides description of the proposed model and its early evaluation on the selected scenarios in a multi-agent traffic simulation.

Asynchronous decentralized prioritized planning for coordination in multi-robot system

  • DOI: 10.1109/IROS.2013.6696903
  • Odkaz: https://doi.org/10.1109/IROS.2013.6696903
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, the multi-robot motion coordination planning problem is addressed. Although a centralized prioritized planning strategy can be used to solve the problem, we rather consider a decentralized variant, which is a more suitable for a robotic system of cooperating unmanned aerial vehicles (UAVs) due to communication limitations, privacy concerns, and a better exploitation of computational resources distributed among the individual robots. However, the existing decentralized prioritized planning algorithm contains synchronization points that all the agents must be able to pass synchronously, which is impractical and inefficient for a real-world deployment of the robotic systems. Therefore, we introduce a new asynchronous decentralized prioritized planning algorithm and show that the method can converge faster than both the synchronous decentralized and centralized algorithms. Further, we demonstrate the applicability of the proposed method as a coordination mechanism within a complex mission planning for a real robotic system consisting of several autonomous UAVs.

External Localization System for Mobile Robotics

  • DOI: 10.1109/ICAR.2013.6766520
  • Odkaz: https://doi.org/10.1109/ICAR.2013.6766520
  • Pracoviště: Katedra počítačů
  • Anotace:
    We present a fast and precise vision-based software intended for multiple robot localization. The core component of the proposed localization system is an efficient method for black and white circular pattern detection. The method is robust to variable lighting conditions, achieves sub-pixel precision, and its computational complexity is independent of the processed image size. With off-the-shelf computational equipment and low-cost camera, its core algorithm is able to process hundreds of images per second while tracking hundreds of objects with millimeter precision. We propose a mathematical model of the method that allows to calculate its precision, area of coverage, and processing speed from the camera’s intrinsic parameters and hardware’s processing capacity. The correctness of the presented model and performance of the algorithm in real-world conditions are verified in several experiments. Apart from the method description, we also publish its source code; so, it can be used as an enabling technology for various mobile robotics problems.

External Localization System for Mobile Robotics

  • Autoři: Krajník, T., Nitsche, M., prof. Ing. Jan Faigl, Ph.D., Mejail, M., Duckett, T., Přeučil, L.
  • Publikace: Proceedings of the International Conference on Advanced Robotics 2013. Piscataway: IEEE, 2013, pp. 1-6. ISBN 978-1-4799-2722-7. Available from: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=6766520&isnumber=6766447
  • Rok: 2013
  • DOI: 10.1109/ICAR.2013.6766520
  • Odkaz: https://doi.org/10.1109/ICAR.2013.6766520
  • Pracoviště: Katedra počítačů
  • Anotace:
    We present a fast and precise vision-based software intended for multiple robot localization. The core component of the proposed localization system is an efficient method for black and white circular pattern detection. The method is robust to variable lighting conditions, achieves sub-pixel precision, and its computational complexity is independent of the processed image size. With off-the-shelf computational equipment and low-cost camera, its core algorithm is able to process hundreds of images per second while tracking hundreds of objects with millimeter precision. We propose a mathematical model of the method that allows to calculate its precision, area of coverage, and processing speed from the camera's intrinsic parameters and hardware's processing capacity. The correctness of the presented model and performance of the algorithm in real-world conditions are verified in several experiments. Apart from the method description, we also publish its source code; so, it can be used as an enabling technology for various mobile robotics problems.

Growing neural gas efficiently

  • DOI: 10.1016/j.neucom.2012.10.004
  • Odkaz: https://doi.org/10.1016/j.neucom.2012.10.004
  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper presents optimization techniques that substantially speed up the Growing Neural Gas (GNG) algorithm. The GNG is an example of the Self-Organizing Map algorithm that is a subject of an intensive research interest in recent years as it is used in various practical applications. However, a poor time performance on large scale problems requiring neural networks with a high amount of nodes can be a limiting factor for further applications (e.g., cluster analysis, classification, 3-D reconstruction) or a wider usage. We propose two optimization techniques that are aimed exclusively on an efficient implementation of the GNG algorithm internal structure rather than on a modification of the original algorithm. The proposed optimizations preserve all properties of the GNG algorithm and enable to use it on large scale problems with reduced computational requirements in several orders of magnitude.

Low-Cost Embedded System for Relative Localization in Robotic Swarms

  • DOI: 10.1109/ICRA.2013.6630694
  • Odkaz: https://doi.org/10.1109/ICRA.2013.6630694
  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    In this paper, we present a small, light-weight, low-cost, fast and reliable system designed to satisfy requirements of relative localization within a swarm of micro aerial vehicles. The core of the proposed solution is based on off-the-shelf components consisting of the Caspa camera module and Gumstix Overo board accompanied by a developed efficient image processing method for detecting black and white circular patterns. Although the idea of the roundel recognition is simple, the developed system exhibits reliable and fast estimation of the relative position of the pattern up to 30 fps using the full resolution of the Caspa camera. Thus, the system is suited to meet requirements for a vision based stabilization of the robotic swarm. The intent of this paper is to present the developed system as an enabling technology for various robotic tasks.

On Determination of Goal Candidates in Frontier-Based Multi-Robot Exploration

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Kulich, M.
  • Publikace: Proceedings of 6th European Conference on Mobile Robots. Barcelona: Institut de Robotica i Informatica Industrial, 2013. p. 210-215. ISBN 978-1-4799-0263-7.
  • Rok: 2013
  • DOI: 10.1109/ECMR.2013.6698844
  • Odkaz: https://doi.org/10.1109/ECMR.2013.6698844
  • Pracoviště: Katedra počítačů
  • Anotace:
    Frontier-based approach can be considered as a de facto standard method for a mobile robot exploration task. Many variants have been proposed; however, relatively little attention has been made to study the influence of goal candidates generation to the performance of the exploration. In regular approaches, frontiers are considered as eventual goals for the next-best-view selection using a utility function combining a distance cost and expected information gain. The aim of this paper is to show that using goal candidates that are independent of the distance cost can improve the performance of exploration strategies. The found insights are supported by a statistical evaluation of thousands of trials performed for various environments.

Reaction-Diffusion Process Based Computational Model for Mobile Robot Exploration Task

  • Autoři: Vazquez-Otero, A., prof. Ing. Jan Faigl, Ph.D., Duro, N., Dormido, R.
  • Publikace: Workshop Proceedings on Unconventional Approaches to Robotics. Piscataway: IEEE, 2013, pp. 16-18. ISBN 978-1-4673-5642-8.
  • Rok: 2013
  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper presents an exploration algorithm based on properties of reaction-diffusion models. The approach is based on our previous work on this topic, in which a novel path planning algorithm was developed providing competitive paths to standard approaches like smoothness of the found solution. In this paper, it is shown how the developed principles can be applied in exploration of unknown environment with a mobile robot. The presented approach represents a novel exploration strategy where the main decision logic is based on principles arising from the underlying dynamics of the reaction-diffusion systems.

Speeding Up Coverage Queries in 3D Multi-Goal Path Planning

  • Autoři: Janoušek, P., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: ICRA2013: Proceedings of 2013 IEEE International Conference on Robotics and Automation. Piscataway: IEEE, 2013. p. 5067-5072. ISSN 1050-4729. ISBN 978-1-4673-5641-1.
  • Rok: 2013
  • DOI: 10.1109/ICRA.2013.6631303
  • Odkaz: https://doi.org/10.1109/ICRA.2013.6631303
  • Pracoviště: Katedra počítačů
  • Anotace:
    In this paper, we present a supporting structure for speeding up visibility queries needed for a 3D multi-goal path planning arising from a robotic coverage problem where goals are sensing locations from which an object of interest can be covered. Although such coverage problems can be addressed by a decomposed approach where sensing locations are determined prior finding the sequence of their visits, the proposed approach is motivated by a solution of the problem in which sensing locations are simultaneously determined together with evaluation of the path connecting them in order to provide a cost effective inspection path. The proposed structure divides the space into elements that support determination of suitable sensing locations to cover the objects during solution of the multi-goal path planning.

SyRoTek - Distance Teaching of Mobile Robotics

  • DOI: 10.1109/TE.2012.2224867
  • Odkaz: https://doi.org/10.1109/TE.2012.2224867
  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    E-learning is a modern and effective approach for training in various areas and at different levels of education. This paper gives an overview of SyRoTek - an e-learning platform for mobile robotics, artificial intelligence, control engineering and related domains. SyRoTek provides remote access to a set of fully autonomous mobile robots placed in a restricted area with dynamically reconfigurable obstacles, which enables solving a huge variety of problems. A user is able to control the robots in real-time by their own developed algorithms as well as being able to analyze gathered data and observe activity of the robots by provided interfaces. The system is currently used for education at the Czech Technical University in Prague and at the University of Buenos Aires, and it is freely accessible to other institutions. In addition to the system overview, the paper presents the experience gained from the actual deployment of the system in teaching activities.

Visiting Convex Regions in a Polygonal Map

  • DOI: 10.1016/j.robot.2012.08.013
  • Odkaz: https://doi.org/10.1016/j.robot.2012.08.013
  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    This paper is concerned with a variant of the multi-goal path planning in which goals are represented as convex polygons. The problem is to find a closed shortest path in a polygonal map such that all goals are visited. The proposed solution is based on a self-organizing map (SOM) algorithm for the traveling salesman problem. Neurons’ weights are considered as nodes inside the polygonal domain and connected nodes represent a path that evolves according to the proposed adaptation rules. In addition, a reference algorithm based on the solution of the traveling salesman problem and the consecutive touring polygons problem is provided to find high quality solutions of the created set of problems. The problems are designed to represent various inspection and patrolling tasks and can form a kind of benchmark set for multi-goal path planning algorithms. The performance of the algorithms is examined in this problem set, which includes an instance of the watchman route problem with restricted visibility range. The proposed SOM based algorithms provide a unified approach to solve various visibility based routing problems in polygonal maps while they provide a competitive quality of solutions to the reference algorithm with significantly lower computational requirements.

Advanced Methods for UAV Autonomy

  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    This paper presents advanced technologies for Micro Unmanned Aerial Vehicles (µ-UAVs) developed by the Intelligent and Mobile Robotics Group of the Czech Technical University in Prague. These methods, based on artificial intelligence, allow the µ-UAVs to fly fully autonomously with a minimal required interaction with an operator. The paper aims to show an applicability of the developed methods in tasks of autonomous navigation, periodical surveillance, inspection, ground unit localization, and mapping. The main intention of this contribution is to demonstrate technologies that can extend the operational deployment of today's u-UAV systems.

Goal Assignment using Distance Cost in Multi-Robot Exploration

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Kulich, M., Přeučil, L.
  • Publikace: Proceedings of 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2012. p. 3741-3746. ISBN 978-1-4673-1735-1.
  • Rok: 2012
  • DOI: 10.1109/IROS.2012.6385660
  • Odkaz: https://doi.org/10.1109/IROS.2012.6385660
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we discuss the problem of goal assignment in the multi-robot exploration task. The presented work is focused on the underlying optimal assignment problem of the multi-robot task allocation that is addressed by three state-of-the art approaches. In addition, we propose a novel exploration strategy considering allocation of all current goals (not only immediate goal) for each robot, which leads to the multiple traveling salesman problem formulation. Although the problem is strongly NP-hard, we show its approximate solution is computationally feasible and its overall requirements are competitive to the previous approaches. The proposed approach and three well-known approaches are compared in series of problems considering various numbers of robots and sensor ranges. Based on the evaluation of the results the proposed exploration strategy provides shorter exploration times than the former approaches.

Low Cost MAV Platform AR-Drone in Experimental Verifications of Methods for Vision Based Autonomous Navigation

  • DOI: 10.1109/IROS.2012.6386277
  • Odkaz: https://doi.org/10.1109/IROS.2012.6386277
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Several navigation tasks utilizing a low-cost Micro Aerial Vehicle (MAV) platform AR-drone are presented in this paper to show how it can be used in an experimental verification of scientific theories and developed methodologies. An important part of this paper is an attached video showing a set of such experiments. The presented methods rely on visual navigation and localization using on-board cameras of the AR-drone employed in the control feedback. The aim of this paper is to demonstrate flight performance of this platform in real world scenarios of mobile robotics.

On Localization Uncertainty in an Autonomous Inspection

  • DOI: 10.1109/ICRA.2012.6224706
  • Odkaz: https://doi.org/10.1109/ICRA.2012.6224706
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents a multi-goal path planning framework based on a self-organizing map algorithm and a model of the navigation describing evolution of the localization error. The framework combines finding a sequence of goals' visits with a goal-to-goal path planning considering localization uncertainty. The approach is able to deal with local properties of the environment such as expected visible landmarks usable for the navigation. The local properties affect the performance of the navigation, and therefore, the framework can take the full advantage of the local information together with the global sequence of the goals' visits to find a path improving the autonomous navigation. Experimental results in real outdoor and indoor environments indicate that the framework provides paths that effectively decreases the localization uncertainty; thus, increases the reliability of the autonomous goals' visits.

Path Planning Based on Reaction-Diffusion Process

  • Autoři: Otero, A.V., prof. Ing. Jan Faigl, Ph.D., Muňuzuri, A.P.
  • Publikace: Proceedings of 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems. Piscataway: IEEE, 2012. pp. 896-901. ISBN 978-1-4673-1735-1.
  • Rok: 2012
  • DOI: 10.1109/IROS.2012.6385592
  • Odkaz: https://doi.org/10.1109/IROS.2012.6385592
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we present a novel path planning algorithm based on properties that reaction-diffusion (RD) models exhibit by the underlying non-linear dynamics of the considered system. In particular herein considered a two-variable RD model provides advantages of natural parallelism, noise resistance, and especially the non-annihilating feature that traveling fronts separating two stable states exhibit upon a collision. Based on this, we developed a path planning algorithm that provides paths with lengths competitive to standard path planning approaches. Moreover, the results presented indicates the paths are smoother and also within a safe distance from obstacles; thus, the found paths combine advantages of two fundamental approaches, namely the DT algorithm and Voronoi diagram.

Pokročilé metody řízení autonomních bezpilotních prostředků

  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    Tento článek prezentuje pokročilé metody řízení bezpilotních prostředků (UAV) vyvinuté skupinou inteligentní a mobilní robotiky Českého vysokého učení technického v Praze. Tyto metody, založené na technikách umělé inteligence, umožňují bezpilotním prostředkům plnit řadu úkolů zcela samostatně s minimální nezbytnou interakcí s operátorem. V článku jsou uvedeny příklady aplikací těchto metod v úlohách autonomní navigace, inspekce, dohledu, lokalizace a mapování. Hlavním záměrem příspěvku je demonstrace technologií umožňujících rozšířit nasazení malých bezpilotních prostředků v rozličných, zejména dohledových, průzkumných a záchranných misích.

SyRoTek - systém pro robotickou televýuku a experimentování

  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    Tento článek popisuje SyRoTek - platformu pro vzdálené vzdělávání v oblastech mobilní robotiky, umělé inteligence, řízení a příbuzných oborech a pro provádění experimentů s reálnými roboty a senzory. SyRoTek poskytuje přístup ke skupině plně autonomních robotů operujících v omezeném prostoru s automaticky regulovatelnými překážkami a tím umožňuje řešení rozličných úloh. Systém byl navržen se speciálním zaměřením na dlouhodobé a intenzivní používání tak, aby byl dostupný v režimu 24/7. Uživatelé mohou nejenom sledovat a zpracovávat reálná data získaná ze senzorů, ale zejména řidit roboty v reálném čase vlastními aplikacemi. Systém je aktivně využíván pro výzkum I výuku na ČVUT v Praze a na Universidad de Buenos Aires. Mimoto je volně přístupný ostatním zájemcům, ať již jednotlivcům či institucím.

A Multi-Goal Path Planning for Goal Regions in the Polygonal Domain

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper concerns a variant of the multi-goal path planning problem in which goals may be polygonal regions. The problem is to find a closed shortest path in a polygonal map such that all goals are visited. The proposed solution is based on a self-organizing map algorithm for the traveling salesman problem, which is extended to the polygonal goals. Neurons' weights are considered as nodes inside the polygonal domain and connected nodes represent a path that evolves according to the proposed adaptation rules. Performance of the rules is evaluated in a set of problems including an instance of the watchman route problem with restricted visibility range. Regarding the experimental results the proposed algorithm provides flexible approach to solve various NP-hard routing problems in polygonal maps.

A Sampling Schema for Rapidly Exploring Random Trees Using a Guiding Path

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, a novel sampling schema for Rapidly Exploring Random Trees (RRT) is proposed to address the narrow passage issue. The introduced method employs a guiding path to steer the tree growth towards a given goal. The main idea of the proposed approach stands in a preference of the sampling of the configuration space C along a given guiding path instead of sampling of the whole space. While for a low-dimensional C the guiding path can be found as a geometric path in the robot workspace, such a path does not provide useful information for efficient sampling of a high-dimensional C. We propose an iterative scaling approach to find a guiding path in such high-dimensional configuration spaces. The approach starts with a scaled geometric model of the robot to a fraction of its original size for which a guiding path is found using the RRT algorithm. Then, such a path is iteratively used in the proposed RRT-Path algorithm for a larger robot up to its original size

A Sensor Placement Algorithm for a Mobile Robot Inspection Planning

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Kulich, M., Přeučil, L.
  • Publikace: Journal of Intelligent and Robotic Systems. 2011, 62(3-4), 329-353. ISSN 0921-0296.
  • Rok: 2011
  • DOI: 10.1007/s10846-010-9449-0
  • Odkaz: https://doi.org/10.1007/s10846-010-9449-0
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we address the inspection planning problem to "see" the whole area of the given workspace by a mobile robot. The problem is decoupled into the sensor placement problem and the multi-goal path planning problem to visit found sensing locations. However the decoupled approach provides a feasible solution, its overall quality can be poor, because the sub-problems are solved independently. We propose a new randomized approach that considers the path planning problem during solution of the sensor placement problem. The proposed algorithm is based on a guiding of the randomization process according to prior knowledge about the environment. The algorithm is compared with two algorithms already used in the inspection planning. Performance of the algorithms is evaluated in several real environments and for a set of visibility ranges. The proposed algorithm provides better solutions in both evaluated criterions: a number of sensing locations and a length of the inspecti

A TECHNICAL SOLUTION OF A ROBOTIC E-LEARNING SYSTEM IN THE SYROTEK PROJECT

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    SyRoTek (a system for a robotic e-learning) is a robotic virtual laboratory being developed at Czech Technical University in Prague. SyRoTek provides access to real mobile robots placed in an arena with dynamically reconfigurable obstacles enabling variety of tasks in the field of mobile robotics and artificial intelligence. The robots are equipped with several sensors allowing students to realize how robots' perception works and how to deal with uncertainties of the real world. An insight to a technical solution of the SyRoTek project is presented in this paper.

An Application of the Self-Organizing Map in the non-Euclidean Traveling Salesman Problem

  • DOI: 10.1016/j.neucom.2010.08.026
  • Odkaz: https://doi.org/10.1016/j.neucom.2010.08.026
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    An application of the self-organizing map (SOM) to the Traveling Salesman Problem (TSP) has been reported by many researchers, however these approaches are mainly focused on the Euclidean TSP variant.We consider the TSP as a problem formulation for the multi-goal path planning problem in which paths among obstacles have to be found.Weapply a simple approximation of the shortest path that seems to be suitable for the SOM adaptation procedure. The approximation is based on a geometrical interpretation of SOM, where weights of neurons represent nodes that are placed in the polygonal domain. The approximation is verified in a set of real problems and experimental results show feasibility of the proposed approach for the SOM based solution of the non-Euclidean TSP.

AR Drone as a Platform for Robotic Research and Education

  • DOI: 10.1007/978-3-642-21975-7_16
  • Odkaz: https://doi.org/10.1007/978-3-642-21975-7_16
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents the AR-Drone quadrotor helicopter as a robotic platform usable for research and education. Apart from the description of hardware and software, we discuss several issues regarding drone equipment, abilities and performance. We show, how to perform basic tasks of position stabilization, object following and autonomous navigation. Moreover, we demonstrate the drone ability to act as an external navigation system for a formation of mobile robots. To further demonstrate the drone utility for robotic research, we describe experiments in which the drone has been used. We also introduce a freely available software package, which allows researches and students to quickly overcome the initial problems and focus on more advanced issues.

Estimation of Mobile Robot Pose from Optical Mouses

  • Autoři: Mudrová, L., prof. Ing. Jan Faigl, Ph.D., Halgašík, J., doc. Ing. Tomáš Krajník, Ph.D.,
  • Publikace: Eurobot Conference 2010, International Conference on Research and Education in Robotics. Bern: University of Applied Sciences, 2011, pp. 93-107. ISSN 1865-0929. ISBN 978-3-642-27271-4.
  • Rok: 2011
  • DOI: 10.1007/978-3-642-27272-1_9
  • Odkaz: https://doi.org/10.1007/978-3-642-27272-1_9
  • Pracoviště: Katedra kybernetiky, Katedra počítačů
  • Anotace:
    This paper describes a simple method of dead-reckoning based on off-the-shelf components: optical mouses and a laptop. The problem is formulated as finding a transformation of mouses positions to position of the robot. The formulation of the transformation is based on a method already used in range-based localization. Beside a solution of the transformation, the paper provides description of practical application of mouse based localization for a home made robot. The paper considers identification and mouse data reading procedures as well. The presented approach has been evaluated in several real experiments and the proposed localization provides competitive results to the odometry based on high-precision stepper motors.

Inspection Planning in the Polygonal Domain by Self-Organizing Map

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Přeučil, L.
  • Publikace: Applied Soft Computing. 2011, 11(8), 5028-5041. ISSN 1568-4946.
  • Rok: 2011
  • DOI: 10.1016/j.asoc.2011.05.055
  • Odkaz: https://doi.org/10.1016/j.asoc.2011.05.055
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Inspection planning is a problem of finding a (closed) shortest path from which a robot "sees" the whole workspace. The problem is closely related to the Traveling Salesman Problem (TSP) if the discrete sensing is performed only at the finite number of sensing locations. For the continuous sensing, the problem can be formulated as the Watchman Route Problem (WRP), which is known to be NP-hard for the polygonal representation of the robot workspace. Although several Self-Organizing Map (SOM) approaches have been proposed for the TSP, they are strictly focused to the Euclidean TSP, which is not the case of the inspection path planning in the polygonal domain. In this paper, a novel SOM adaptation schema is proposed to address both variants of the inspection planning with discrete and continuous sensing in the polygonal domain. The schema is compared with the state of the art SOM schema for the TSP in a set of multi-goal path planning problems and WRPs.

On Distance Utility in the Exploration Task

  • Autoři: Kulich, M., prof. Ing. Jan Faigl, Ph.D., Přeučil, L.
  • Publikace: ICRA2011: Proceedings of 2011 IEEE International Conference on Robotics and Automation. Madison: Omnipress, 2011. p. 4455-4460. ISSN 1050-4729. ISBN 978-1-61284-386-5.
  • Rok: 2011
  • DOI: 10.1109/ICRA.2011.5980221
  • Odkaz: https://doi.org/10.1109/ICRA.2011.5980221
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Performance of exploration strategies strongly depends on the process of determination of a~next robot goal. Current approaches define different utility functions how to evaluate and select possible next goal candidates. One of the mostly used evaluation criteria is the distance cost that prefers candidates close to the current robot position. If this is the only criterion, simply the nearest candidate is chosen as the next goal. Although this criterion is simple to implement and gives feasible results there are situations where the criterion leads to wrong decisions. This paper presents the distance cost that reflects traveling through all goal candidates. The cost is determined as a~solution of the Traveling Salesman Problem using the Chained Lin-Kernighan heuristic. % to find a near optimal solution. The cost can be used as a~stand-alone criterion as well as it can be integrated into complex decision systems. Experimental results for open-space and office-like experiments show that the proposed approach outperforms the standard one in the length of the traversed trajectory during the exploration while the computational burden is not significantly increased.

On the Performance of Self-organizing Maps for the Non-Euclidean Traveling Salesman Problem in the Polygonal Domain

  • DOI: 10.1016/j.ins.2011.05.019
  • Odkaz: https://doi.org/10.1016/j.ins.2011.05.019
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, two state-of-the-art algorithms for the Traveling Salesman Problem (TSP) are examined in the multi-goal path planning problem motivated by inspection planning in the polygonal domain W. Both algorithms are based on the self-organizing map (SOM) for which an application in W is not typical. The first is Somhom's algorithm, and the second is the Co-adaptive net. These algorithms are augmented by a simple approximation of the shortest path among obstacles in W. Moreover, the competitive and cooperative rules are modified by recent adaptation rules for the Euclidean TSP, and by proposed enhancements to improve the algorithms' performance in the non-Euclidean TSP. Based on the modifications, two new variants of the algorithms are proposed that reduce the required computational time of their predecessors by an order of magnitude, therefore making SOM more competitive with combinatorial heuristics.

Self-Organizing Map for the Multi-Goal Path Planning with Polygonal Goals

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Přeučil, L.
  • Publikace: Proceedings of Artificial Neural Networks and Machine Learning. Heidelberg: Springer, 2011. pp. 85-92. ISBN 978-3-642-21734-0.
  • Rok: 2011
  • DOI: 10.1007/978-3-642-21735-7_11
  • Odkaz: https://doi.org/10.1007/978-3-642-21735-7_11
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents a self-organizing map approach for the multi-goal path planning problem with polygonal goals. The problem is to find a shortest closed collision free path for a mobile robot operating in a planar environment represented by a polygonal map W. The requested path has to visit a given set of areas where the robot takes measurements in order to find an object of interest. Neurons' weights are considered as points in W and the solution is found as approximate shortest paths connecting the points (weights). The proposed self-organizing map has less number of parameters than a previous approach based on the self-organizing map for the traveling salesman problem. Moreover, the proposed algorithm provides better solutions within less computational time for problems with high number of polygonal goals.

Uncertainty of Mobile Robot Localization in Cooperative Inspection Tasks

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents an overview of the rst year of a two year project aiming at multi-goal path planning and navigation methods for a team of real mobile robots performing an inspection task in known environments. The real robots are considered and therefore their localization estimation provides position with an uncertainty. This uncertainty is modeled by an iterative equation, which describes the evolution of a position error for a robot navigating by a reliable and provably stable method. In this project, we consider the model in the path planning algorithm for nding a path visiting given set of locations. The proposed planning procedure considers not only the length of the planned path, but also keeps the robot position error low at the locations. The experiments show that although the planned path is longer, the reliability of visiting the inspected locations is increased.

A Monocular Navigation System for RoboTour Competition

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we present a mobile robot navigation system used in the RoboTour challenge. We describe the basic principles of the navigation methods and show how to combine monocular vision and odometry. We propose to use the monocular vision to determine only the robot's heading and the odometry to estimate only the traveled distance. We show that the heading estimation itself can suppress odometric cumulative errors and outline a mathemtical proof of this statement. The practical result of the proof is that even simple algorithms capable to estimate just the heading can be used as a base for "record and replay" techniques. Beside the navigational principles, practical implementation of our navigation system is described. It is based on image processing algorithms for path following and landmark-based crossing traversing. An overview of experimental results is presented as well.

A Simple Yet Reliable Visual Navigation System

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    We present a simple monocular camera-based navigation system for an autonomous vehicle. The method utilizes off-the-shelf components (camera, compass and odometry) and standard algorithms. It does not require any additional infrastructure like radio beacons or GPS. The basic idea of our system is a simple, yet novel method of position estimation based on monocular vision and odometry. Contrary to traditional localization methods, which use advanced mathematical methods to determine vehicle position, our method uses a more practical approach. A monocular vision technique determines heading of the vehicle and the odometry is used to estimate the traveled distance. Though the system is simple, it can deal with variable illumination, seasonal changes of the environment, dynamic objects and obstacles. We believe, that our navigation method is useful in areas, where GPS signal suffers from occlusions and reflections, e.g. forests or canyons.

A Visual Navigation System for RoboTour Competition

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we present our approach to the navigation system for the RoboTour challenge. We describe our intentions, ideas and main principles of our navigation methods which lead to the system that won in years 2008 and 2009. The main idea of our system is a simple yet novel method of position estimation based on monocular vision and odometry. Unlike in other systems, the monocular vision is used to determine only the robot's heading and the odometry is used to estimate only the traveled distance. We show that the heading estimation itself can suppress odometric cumulative errors and prove this statement mathematically and experimentally. The practical result of the proof is that even simple algorithms capable to estimate just the heading can be used as a base for ``record and replay'' techniques. Beside the navigational principles, practical implementation of our navigation system is described.

A Visualization System for Teaching Intelligent Mobile

  • Autoři: Kulich, M., prof. Ing. Jan Faigl, Ph.D., Chudoba, J., Košnar, K.
  • Publikace: Proceedings of The 2nd International Conference on Simulation, Modeling, and Programming for Autonomus Robots. Berlin: Springer, 2010, ISSN 0302-9743. ISBN 978-3-642-17318-9.
  • Rok: 2010
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper is concerned with a visualization of sensory data in the SyRoTek project that aims to create an e­learning platform for supporting teaching mobile robotics and artificial intelligence. The visualization is based on the Stage simulator that is extended to support multi­view configuration and combination of real scenes and simulated environments scenes in a single visualization window. The modified Stage becomes a standalone visualization application that can be used for on­line and off­line video creation. Moreover, the application can be used in the development process of robot control application. It helps students to debug their applications and to understand the robot perception of the real surrounding environment in real­time or from recorded/logged data.

Approximate Solution of the Multiple Watchman Routes Problem With Restricted Visibility Range

  • Autoři: prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: IEEE Transactions on Neural Networks. 2010, 21(10), 1668-1679. ISSN 1045-9227.
  • Rok: 2010
  • DOI: 10.1109/TNN.2010.2070518
  • Odkaz: https://doi.org/10.1109/TNN.2010.2070518
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, a new self-organizing map (SOM) based adaptation procedure is proposed to address the Multiple Watchmen Route Problem with the restricted visibility range in the polygonal domain W. A watchman route is represented by a ring of connected neurons weights that evolves in W while obstacles are considered by approximation of the shortest path. The adaptation procedure considers a coverage of W by the ring in order to attract nodes towards uncovered parts of W. The proposed procedure is experimentally verified in a set of environments and several visibility ranges. Performance of the procedure is compared with the decoupled approach based on solutions of the Art Gallery Problem and the consecutive Traveling Salesman Problem. The experimental results show suitability of the proposed procedure based on relatively simple supporting geometrical structures enabling application of SOM principles to watchman route problems in W.

Mobile Robotics at FEE CTU

  • Autoři: prof. Ing. Jan Faigl, Ph.D., doc. Ing. Tomáš Krajník, Ph.D., Košnar, K., Szücsová, H., Chudoba, J., Grimmer, V., Přeučil, L.
  • Publikace: First International Conference on Robotics in Education, Bratislava. Bratislava: Slovak University of Technology in Bratislava, 2010. pp. 43-48. ISBN 978-80-227-3353-3.
  • Rok: 2010
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we describe concepts and main ideas of the labs of the Mobile Robotics course at FEE, CTU in Prague. We present our gained experience from three years of teaching of the course. We consider the students' contact with real hardware and real sensor data as the most important part of mobile robotics as the mobile robot can quickly lose information about its position in contrast to stationary robotic manipulators. To achieve our desired pedagogical goals we have decided to develop a new small platform that will be based mostly on off-the-shelf components and it will have sufficient computation power to use the Player robotic framework. The labs are organized into four consecutive assignments and a final assignment that combines particular students' results from the previous tasks. The final assignment is to create an algorithm that navigates the mobile robot in order to create a topological map of the environment and reuse this map for later navigation.

Mobile Robotics at FEE CTU

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we describe concepts and main ideas of the labs of the Mobile Robotics course at FEE, CTU in Prague. We present our gained experience from three years of teaching of the course. We consider the students' contact with real hardware and real sensor data as the most important part of mobile robotics as the mobile robot can quickly lose information about its position in contrast to stationary robotic manipulators. To achieve our desired pedagogical goals we have decided to develop a new small platform that will be based mostly on off-the-shelf components and it will have sufficient computation power to use the Player robotic framework. The labs are organized into four consecutive assignments and a final assignment that combines particular students' results from the previous tasks. The final assignment is to create an algorithm that navigates the mobile robot in order to create a topological map of the environment and reuse this map for later navigation.

On a Mobile Robotics E-learning System

  • Autoři: Kulich, M., Košnar, K., Chudoba, J., prof. Ing. Jan Faigl, Ph.D., Přeučil, L.
  • Publikace: Proceedings of the Twentieth European Meeting on Cybernetics and Systems Research. Vienna: Austrian Society for Cybernetics Studies, 2010, pp. 597-602. ISBN 978-3-85206-178-8.
  • Rok: 2010
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    With huge expansion of artificial intelligence and mobile robotics technologies into many industrial applications as well as day-to-day activities it is necessary to train students to understand and manage these technologies. The key of successful education is then to provide access to real data from nowadays sensors used in the area. The main goal of the SyRoTek system described in the paper is to afford students in basic and advanced courses in the field opportunity to control real robots with real sensors via web. Moreover, the system will be open to skilled public for test/verification purposes. The design SyRoTek system comprises a team of 12 tele-operated mobile robots acting in 24/7 maintance-free environment equipped with charging docks and reconfigurable system of obstacles, all being observable and accessible via Internet.

RoboTour 2009 - soutěž outdoorových robotů

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Koncem září minulého roku proběhl v brněnském parku Lužánky čtvrtý ročník soutěže autonomních robotů RoboTour 2009. Soutěž je zjednodušenou analogií orientačního běhu pro mobilní roboty.

Simple, Yet Stable Bearing-Only Navigation

  • DOI: 10.1002/rob.20354
  • Odkaz: https://doi.org/10.1002/rob.20354
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This article describes a simple monocular navigation system for a mobile robot based on the map-and-replay technique. The presented method is robust, easy to implement, does not require sensor calibration or structured environment and its computational complexity is independent of the environment size. The method can navigate a robot while sensing only one landmark at a time, making it more robust than other monocular approaches. The aforementioned properties of the method allow even low-cost robots to effectively act in large outdoor and indoor environments with natural landmarks only. The basic idea is to utilize a monocular vision to correct only the robot's heading and leaving distance measurements just to the odometry. The heading correction itself can suppress the odometric error and prevent the overall position error from diverging. The influence of a map-based heading estimation and odometric errors on the overall position uncertainty is examined.

Surveillance Planning with Localization Uncertainty for UAVs

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents a new multi-goal path planning method that incorporates the localization uncertainty in a visual inspection surveillance task. It is shown that the reliability of the executed found plan is increased if the localization uncertainty of the used navigation method is taken into account during the path planning. The navigation method follows the map&replay technique based on a combination of monocular vision and dead-reckoning. The mathematical description of the navigation method allows efficient computation of the evolution of the robot position uncertainty that is used in the proposed path planning algorithm. The algorithm minimizes the length of the inspection path while the robot position error at the goals is decreased. The presented experimental results indicate that probability of the goals visits can be increased by the proposed algorithm.

SyRoTek - A Robotic System for Education

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents insight to ideas and the current state of the project SyRoTek (System for a robotic e-learning) that aims to create a platform for students' practical verification of gained knowledge in the fields of Robotics and Artificial Intelligence. A set of real mobile robots is being developed in order to provide remote access to real hardware for enrolled students. The advantage of the real system over a pure virtual simulated environment is in realistic confrontation with noise and uncertainty that is an indivisible part of the real world. In such a system, students can acquire in deep understanding of main studied principles in an attractive form, as students (especially future engineers) like to control real things. Robots are designed with special attention to long-term and heavy duty usage. Moreover, support for semi-autonomous evaluation of students' solution of their assignments is a part of the system.

SyRoTek - A Robotic System for Education

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents insight to ideas and the current state of the project SyRoTek (System for a robotic e-learning) that aims to create a platform for students' practical verification of gained knowledge in the fields of Robotics and Artificial Intelligence. A set of real mobile robots is being developed in order to provide remote access to real hardware for enrolled students. The advantage of the real system over a pure virtual simulated environment is in realistic confrontation with noise and uncertainty that is an indivisible part of the real world. In such a system, students can acquire in deep understanding of main studied principles in an attractive form, as students (especially future engineers) like to control real things. Robots are designed with special attention to long-term and heavy duty usage. Moreover, support for semi-autonomous evaluation of students' solution of their assignments is a part of the system.

FPGA-based Speeded Up Robust Features

  • DOI: 10.1109/TEPRA.2009.5339646
  • Odkaz: https://doi.org/10.1109/TEPRA.2009.5339646
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    We present an implementation of the Speeded Up Robust Features (SURF) on a Field Programmable Gate Array (FPGA). The SURF algorithm extracts salient points from image and computes descriptors of their surroundings that are invariant to scale, rotation and illumination changes. The interest point detection and feature descriptor extraction algorithm is often used as the first stage in autonomous robot navigation, object recognition and tracking etc. However, detection and extraction are computationally demanding and therefore can't be used in systems with limited computational power. We took advantage of algorithm's natural parallelism and implemented it's most demanding parts in FPGA logic. Several modifications of the original algorithm have been made to increase it's suitability for FPGA implementation.

RRT-Path: a guided Rapidly exploring Random Tree

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    During last decade Rapidly Exploring random trees (RRT) became widely used for solving a motion planning problem in various areas. Poor performance of these algorithms has been noticed in environments with narrow passages. Several variants have been developed to address this issue. This paper presents a new variation of the RRT designed for sampling environments with narrow passages. Performance of the proposed method has been experimentally verified and results are compared with the original RRT, RRT-Bidirectional na RRT-Blossom algorithms.

SyRoTek - On an e-Learning System for Mobile Robotics and Artificial Intelligence

  • Autoři: Kulich, M., prof. Ing. Jan Faigl, Ph.D., Košnar, K., Přeučil, L., Chudoba, J.
  • Publikace: ICAART 2009. Setúbal: INSTICC Press, 2009. pp. 275-280. ISBN 978-989-8111-66-1.
  • Rok: 2009
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The paper deals with motivations and design leading to succeeding development of a system for remote learning of mobile robotics topics. Specifically, the designed SyRoTek system comprises a team of 12 tele-operated mobile robots acting in 24/7 maintenance-free environment equipped with charging docks and reconfigurable system of obstacles, all being observable and accessible via Internet. The SyRoTek system together with an attached e-learning environment it is aimed to provide the features real data gathering and real robot motion execution. The whole set-up is targeted on training purposes in basic and advanced courses in the field of Intelligent and Mobile Robotics and Collective Robotics as well as for test/verification purposes in a research domain.

SyRoTek - A System for Robotic E-learning

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The paper describes the SyRoTek ("System for robotic e-learning") project in general, its overall design, main hardware and software components and status of the project after one year of solution. The aim of the SyRoTek project is to research, design, and develop novel methods and approaches for building a multi-robot system for distance learning. The foreseen system will allow its remote users to get acquainted with algorithms from areas of modern mobile and collective robotics, artificial intelligence, control, and many other related domains. Advanced users will be able to develop own algorithms and monitor behavior of these algorithms on-line during real experiments. The proposed system reduces a development process and allows a wide spectrum of both individuals and institutions to work with real robotic equipment.

Cooperative Planning in Multiple Robots Systems

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Kulich, M., Přeučil, L.
  • Publikace: WORKSHOP 2006. Praha: České vysoké učení technické v Praze, 2006, pp. 144-145. ISBN 80-01-03439-9.
  • Rok: 2006
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Search and rescue mission is an application of cooperative path planning. Paths for robots are planned according to criterion to find victims as fast as possible. Solution is based on decomposition of inspection task and solution of Multiple Traveling Salesmen Problem with MinMax criterion.

Decision Support by Simulation in a Robotic Soccer Domain

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper describes a decision support system for a robotic soccer team. In a software system managing robots during a soccer match, this component acts as an information provider for other system components. The core of this decision support system is a physics simulator, which can predict game situation in the moment, when a particular decision is delivered to robotic players. This compensates the negative effects of transport delays originating in the system. Moreover, testing of planning and control algorithms on a virtual system is also made possible by this support module. The most important contribution of this component is a strategic decision support by in-game feasibility testing of proposed actions. This is done by evaluating the outcome of such action by physical simulation.

Iterative Prototype Optimisation with Evolved Improvement Steps

  • Autoři: Kubalík, J., prof. Ing. Jan Faigl, Ph.D.,
  • Publikace: Genetic Programming, Proceedings of 9th European Conference, EuroGP 2006. Heidelberg: Springer, 2006, pp. 154-165. ISSN 0302-9743. ISBN 3-540-33143-3.
  • Rok: 2006
  • DOI: 10.1007/11729976_14
  • Odkaz: https://doi.org/10.1007/11729976_14
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper, we introduce a new framework called Iterative Prototype Optimisation with Evolved Improvement Steps. This is a general optimisation framework, where an initial prototype solution is being improved iteration by iteration. In each iteration, a sequence of actions/operations, which improves the current prototype the most, is found by an evolutionary algorithm. The proposed algorithm has been tested on problems from two different optimisation problem domains - binary string optimisation and the traveling salesman problem.

Reasoning and planning for robotsoccer

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper presents architecture of system G-Bots for robotic soccer. Apart from our open architecture description, we focus on new approaches applied in reasoning and planning system components.

Sensing Locations Positioning for Multi-robot Inspection Planning

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Kulich, M.
  • Publikace: 2006 International IEEE Workshop on Distributed Intelligent Systems (DIS 2006). Piscataway: IEEE, 2006. p. 79-84. ISBN 0-7695-2589-X.
  • Rok: 2006
  • DOI: 10.1109/DIS.2006.66
  • Odkaz: https://doi.org/10.1109/DIS.2006.66
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Problems of cooperative multi-robot inspection and exploration play an important role in many practical applications. This paper presents an algorithm for inspection planning based on decomposition of the problem into two subproblems - Art Gallery Problem (AGP) that finds guards (sensing locations) and Multiple Traveling Salesmen Problem (MTSP) that connects the found guards by routes. While standard approaches for Art Gallery Problem try to minimize a number of guards, the proposed method is designed to optimise lengths found by a MTSP solver and therefore to minimise time needed by a team of robots to inspect the working environment. The proposed algorithm has been implemented and tested. Influence of the method to quality of the inspection planning solution and comparison with the Randomized Dual Sampling Schema are discussed.

Transformed Net - Collision Avoidance Algorithm for Robotic Soccer

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Fast and robust obstacle avoidance plays an important role for design of a successful robot soccer team, although not all teams use it nowadays. The standard algorithms assume that a working environment is static or changing slowly. Moreover, computation time and time needed for realization of the planned path is usually not crucial. Speed of robot soccer players (which act as obstacles) can be several meters per second, what requires low reaction time. One criterion for obstacle avoidance is therefore to plan a path far enough from opponent robots to guarantee that their trajectories will not collide with the planned one. This is in contradiction to a primary goal of robot soccer - to reach the desired position as fast as possible. An obstacle avoidance algorithm suited for robot soccer should find acceptable compromise between these two antagonistic requirements.

Cooperative Planning for Heterogeneous Teams in Rescue Operations

  • Autoři: Kulich, M., prof. Ing. Jan Faigl, Ph.D., Přeučil, L.
  • Publikace: IEEE International Workshop on Safety, Security and Rescue Robotics. Piscataway: IEEE, 2005, ISBN 0-7803-8946-8.
  • Rok: 2005

Multiple Traveling Salesmen Problem with Hierarchy of Cities in Inspection Task with Limited Visibility

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Kulich, M., Přeučil, L.
  • Publikace: 5th Workshop on Self-Organizing Maps. Orsay: Université Paris-Sud, 2005. p. 91-98.
  • Rok: 2005
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper introduces a hierarchical self-organizing neural network solving multiple traveling salesmen problem in an inspection tasks with limited visibility in an environment with obstacles. The working environment is represented as a polygon with holes. The inspection task is decomposed into two sub-problems: a) generation of sensing locations (Art Gallery Problem - AGP) and b) connecting of found locations by a set of paths (Multiple Traveling Salesmen Problem - MTSP). Both sub-problems are NP-hard and therefore algorithms finding approximate solutions are used. The AGP solver is based on randomized sampling of the environment, while a self-organizing neural network solves the MTSP. Moreover, the visibility range is limited in real applications. Low visibility leads to increase of number of sensing locations and therefore increases time costs of the MTSP planning.

Path Planning For Multi-robot Inspection Task Considering Acceleration Limits

  • Autoři: prof. Ing. Jan Faigl, Ph.D., Klačnar, G., Matko, D., Kulich, M.
  • Publikace: Proceedings of the fourteenth International Electrotechnical and Computer Science Conference ERK 2005. Ljubljana: IEEE Slovenia Section, 2005, pp. 138-141. ISSN 1581-4572.
  • Rok: 2005
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Paper presents path planning for multi robotic steam in an inspection task with cosideration acceleration of robots

Rescue Operation Planning by Soft Computing Techniques

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Aplication of three soft computing technigues to solve multiple travelling salesmen problem, self organized neural networks, genetic algorithm and ant colony optimization..

Za stránku zodpovídá: Ing. Mgr. Radovan Suk