Lidé

doc. Ing. David Šišlák, Ph.D.

Všechny publikace

Diverse Planning for UAV Trajectories

  • DOI: 10.1007/978-3-662-44440-5_17
  • Odkaz: https://doi.org/10.1007/978-3-662-44440-5_17
  • Pracoviště: Katedra počítačů
  • Anotace:
    Nowadays, unmanned aerial vehicles (UAVs) are more and more often used to solve various tasks in both the private and the public sector. Some of these tasks can often be performed completely autonomously while others are still dependent on remote pilots. They control an UAV using a command display where they can control it manually using joysticks or give it a simple task. The command displays allow to plan the UAV trajectory through waypoints while avoiding no- fly zones. Nevertheless, the operator can be aware of other preferences or soft restrictions for which it’s not feasible to be inserted into the system especially during time critical tasks. We propose to provide the opera- tor with several different alternative trajectories, so he can choose the best one for the current situation. In this contribution we propose sev- eral metrics to measure the diversity of the trajectories. Then we explore several algorithms for the alternative trajectories creation. Finally, we experimentally evaluate them in a benchmark 8-grid domain and we also present the evaluation by human operators.

Validation of an Air-Traffic Controller behavioral model for fast time simulation

  • DOI: 10.1109/ICNSurv.2014.6820020
  • Odkaz: https://doi.org/10.1109/ICNSurv.2014.6820020
  • Pracoviště: Katedra počítačů
  • Anotace:
    The US National Airspace System (NAS) is incredibly complex, and consists of many specific functions. Given predicted increases in air traffic, enhancement of the current system and development of the NextGEN system are critical to maintain safe and efficient operation. Each feature of the system needs to be carefully designed, developed, tested and validated. To this end, human-in-the-loop (HITL) simulations represent one of the most powerful and realistic testing tools. HITL simulations can provide valuable feedback on how new features influence the behavior of human operators. The drawbacks of HITL simulations include limited flexibility and scalability, and high cost. Computer simulations, which involve no direct human activity, can avoid some of these potential problems, and often represent an attractive alternative (or adjunct) to HITL simulation. A crucial question underlying the use of computer models is how well the model captures the human operator (in this case, the air traffic controller). This paper presents a validation of the AgentFly system, specifically a human en-route Air Traffic Controller (ATC) behavioral model. The ATC workload model is based on Multiple Resource Theory including visual scanning, radio emulation and different kinds of uncertainty. Validation of the AgentFly system was performed by comparing model output to HITL simulation data. The AgentFly system used simulated behavior of air traffic controllers and pilots to collect similar data. Both types of output data were processed and compared based on selected metrics.

Planning of diverse trajectories for UAV control displays

  • Pracoviště: Katedra počítačů
  • Anotace:
    Unmanned aerial vehicles (UAVs) are more and more often used to solve different tasks in both the private and the public sector. Some of these tasks, can be often performed completely autonomously, while the others are still dependent on the remote pilots. They control an UAV using a command display where they can control it manually using joysticks, or give it a simple task. The command display allow for the planning of the UAV trajectory through the waypoints while avoiding the no-fly zones. Nevertheless, the operator can be aware of other preferences, or soft restrictions, for which it's not feasible to be inserted into the system especially during the time critical tasks. We propose to provide the operator with several alternative trajectories which are different from the operator's point of view. So he can choose the best one for the current situation. In this contribution we evaluate previously presented techniques for diverse planning. We focus on an evaluation made by a group of human operators and show how it can be deployed in an UAV control display.

AGENTFLY: Multi-Agent Simulation of Air-Traffic Management

  • Autoři: doc. Ing. David Šišlák, Ph.D., Volf, P., Pavlíček, D., prof. Dr. Michal Pěchouček, MSc.,
  • Publikace: ECAI 2012 - 20th European Conference on Artificial Intelligence. Amsterdam: IOS Press, 2012, pp. 1019-1020. ISSN 0922-6389. ISBN 978-1-61499-097-0. Available from: http://www.booksonline.iospress.nl/Content/View.aspx?piid=31572
  • Rok: 2012
  • Pracoviště: Katedra počítačů
  • Anotace:
    The current air-traffic management (ATM) system involves thousands of people, a majority of them being human controllers [5]. Controllers organize the flow of air-traffic to safely maintain airplane distance and plans for assigned airspace sectors. The capacity of ATM depends on many factors, such as availability of air traffic control (i.e., each controller can handle only limited number of airplanes), current or forecasted weather condition, availability of airspace and capacity of airport facilities. An issue occurs at peak hours when the current ATM system reaches its limits.

AgentFly: Scalable, High-Fidelity Framework for Simulation, Planning and Collision Avoidance of Multiple UAVs

  • Pracoviště: Katedra počítačů
  • Anotace:
    AgentFly is a software prototype providing intelligent algorithms for autonomous unmanned aerial vehicles. AgentFly is implemented as a scalable multi-agent system in JAVA running on the top of the Aglobe platform which provides flexible middle-ware supporting seamless interaction among heterogenous software, hardware and human actors. Thanks to JAVA, AgentFly can be easily hosted on UAVs or computers with different operating systems. The multi-agent approach provides straightforward mapping - each airplane is controlled by one agent. Agents integrate intelligent algorithms providing a coordination-based control for autonomous UAVs. In the presented work, only algorithms which are fully distributed among airplanes are used. Such algorithms provide a real autonomous control for UAVs which do not require any central unit (a ground station or master airplane) controlling a group of UAVs. The main benefit is that the group of UAVs can operate also in the situations where the permanent communication link with the central unit or ground operating station is missing. Some of the algorithms presented in this chapter suppose that UAVs are equipped with communication modems which allow them to dynamically establish bi-directional communication channels based on their mutual position. Thus, airplanes utilize the mobile ad-hoc wireless network created by their communication modems. These algorithms provide the robust control in critical situations: loose of communication, destroyed airplane.

Autonomous UCAV Coordination in Dynamic Search and Destroy Missions

  • Pracoviště: Katedra počítačů
  • Anotace:
    Military operations of the last decades are no longer conflicts of only two participants. Coalitions of countries are participating on one or both sides of the conflict. These conflicts are composed of a number of missions of different nature. We focus on the search and destroy missions which form an irreplaceable part of most of the conflicts since the Vietnam War. With the introduction of Unmanned Combat Air Vehicles (UCAVs) the importance of these missions increases even more since it allows to further decrease the number of causalities among the allies. In this article we discuss the use of UCAVs in search and destroy missions and compare two approaches: a multi-agent negotiation approach, and Process Integrated Mechanism (PIM). Both approaches allow a high degree of autonomy of the UCAVs, promising to decrease the operator load and the base–UCAV communication. We propose several different quality metrics and use them to evaluate and compare both approaches. We also propose an interesting strategy that uses both approaches to create a coalition of autonomous UCAVs, taking advantage of the strengths of a multi-agent approach and PIM.

Multi-Agent Simulation of En-Route Human Air-Traffic Controller

  • Autoři: doc. Ing. David Šišlák, Ph.D., Volf, P., prof. Dr. Michal Pěchouček, MSc., Cannon, C.T., Nguyen, D.N., Regli, W.C.
  • Publikace: Proceedings of the 24th Innovative applications of artificial intelligence conference. Menlo Park, California: AAAI Press, 2012, pp. 2323-2328. ISBN 978-1-57735-568-7. Available from: http://www.aaai.org/ocs/index.php/IAAI/IAAI-12/paper/view/4873
  • Rok: 2012
  • Pracoviště: Katedra počítačů
  • Anotace:
    The Next-Generation Transportation program coordinates the evolution and transformation of the current air-traffic management (ATM) system for the National Airspace System (NAS). Currently the NAS has a limited capacity and cannot handle the increasing future air traffic demands. However, before newly proposed ATM concepts are deployed they must be rigorously evaluated under realistic conditions. This paper presents AGENTFLY, an emerging NAS-wide highfidelity multi-agent ATM simulator with precise emulation of the human controller operation workload model and human-system interaction. The simulator is validated using a flight scenario developed by the U.S. Federal Aviation Administration that is based on real data. We present preliminary results focusing on the accuracy of the simulated controllers within AGENTFLY.

Nas-wide en-route air-traffic controller modeling

  • DOI: 10.1109/ICNSurv.2012.6218417
  • Odkaz: https://doi.org/10.1109/ICNSurv.2012.6218417
  • Pracoviště: Katedra počítačů
  • Anotace:
    Increasing air-traffic demand implies that new air-traffic management (ATM) concepts lowering controller loads, maintaining safety and increasing efficiency need to be designed and implemented. Many of such ideas are prepared within NextGEN. Before they are deployed to real daily usage in National Airspace System (NAS), they must be rigorously evaluated under realistic conditions. The paper presents AGENTFLY, a NAS-wide high-fidelity distributed multi-agent simulator with precise emulation of the human controller operation workload model and human-system interaction. The current version of AGENTFLY provides precise modeling of the human radar controller (R-side) operating in en-route sector.

Sense and Avoid Concepts: Vehicle-Based SAA Systems (Vehicle-to-Vehicle)

  • Pracoviště: Katedra počítačů
  • Anotace:
    The various scenarios of unmanned aerial vehicles (UAVs) deployment require the ability to navigate UAVs in unknown terrain. UAV, while fulfilling the mission objectives, has to avoid static obstacles as well as moving obstacles like other UAVs, airplanes, balloons or areas with bad weather forecast or bad weather conditions. Furthermore, if UAV enters the commercial controlled airspace, it needs to be able to sense and avoid the potential conflicts considering the air-traffic control regulations. The concepts for development of automated systems providing the sense and avoid capability (also referred to as collision detection and resolution systems, CDR) came mainly from two domains. The first one is the air-traffic management domain, where automated tools like Traffic Collision Avoidance System (TCAS) and Precission Runway Monitor (PRM) are used to increase safety and fluency of the air-traffic. The second one is the artificial intelligence research and particularly robotics, where scientists investigated the trajectory planning and obstacle avoidance algorithms for aerial, ground and maritime systems.

Towards Parallel Real-Time Trajectory Planning

  • DOI: 10.1007/978-3-642-28786-2_11
  • Odkaz: https://doi.org/10.1007/978-3-642-28786-2_11
  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper exploits the computing power of widely available multi-core machines to accelerate the trajectory planning by parallelisation of the search algorithm. In particular we investigate the approach that schedules the workload on the cores using the hashing function based on the geographical partitioning of the search space. We use this approach to parallelize the AA* algorithm. In our solution, each partition of the geographical space is represented as an agent. The concept is evaluated on the simulation of real-time trajectory planning of aircraft respecting the environment and real aircraft performance models. We show that the approach decreases the planning time significantly on common multi-core machines preserving the quality of the trajectory provided by AA* algorithm.

Agent-Based Cooperative Decentralized Airplane-Collision Avoidance

  • DOI: 10.1109/TITS.2010.2057246
  • Odkaz: https://doi.org/10.1109/TITS.2010.2057246
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The efficiency of the current centralized air-traffic management is limited. A next-generation air transportation system should allow airplanes (manned and unmanned) to change their flight paths during the flight without approval from a centralized en route control. Such a scheme requires decentralized peer-to-peer conflict detection and collision-avoidance processes. In this paper, two cooperative (negotiation-based) conflictresolution algorithms are presented: iterative peer-to-peer and multiparty algorithms. They are based on high-level flight-plan variations using evasion maneuvers. The algorithms work with a different level of coordination autonomy, respect realistic assumptions of imprecise flight execution (integrating required navigation performance), and work in real time, where the planning and plan-execution phases interleave. Both algorithms provide a resolution in a 4-D domain (3-D space and time).

Agentfly: NAS-wide simulation framework integrating algorithms for automated collision avoidance

  • Autoři: doc. Ing. David Šišlák, Ph.D., Volf, P., Kopřiva, Š., prof. Dr. Michal Pěchouček, MSc.,
  • Publikace: Integrated Communications Navigation and Surveillance Conference. Piscataway: IEEE, 2011, pp. 1-11. ISSN 2155-4943. ISBN 978-1-4577-0592-2. Available from: http://ieeexplore.ieee.org/search/srchabstract.jsp?arnumber=5935278
  • Rok: 2011
  • DOI: 10.1109/ICNSURV.2011.5935278
  • Odkaz: https://doi.org/10.1109/ICNSURV.2011.5935278
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    AgentFly is a software prototype providing a distributed architecture for large-scale NAS-wide simulation implemented as a multi-agent system. AgentFly is implemented on top of the Aglobe platform which is both an implementation framework and a runtime engine for custom agents. It was selected over possible alternatives (e.g. JADE) for its outstanding performance and scalability supporting seamless interaction among heterogeneous software, hardware and human actors. AgentFly system has been developed for over five years. It was initially built f or simulation-based validation and comparison of various approaches for autonomous collision avoidance algorithms adopting the free-flight concept. Later, AgentFly has been extended with high-level control algorithms providing tactical control - i.e. coordination of several autonomous unmanned aerial vehicles (UAV).

Automated Conflict Resolution Utilizing Probability Collectives Optimizer

  • DOI: 10.1109/TSMCC.2010.2089448
  • Odkaz: https://doi.org/10.1109/TSMCC.2010.2089448
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Rising manned air traffic and deployment of unmanned aerial vehicles in complex operations requires integration of innovative and autonomous conflict detection and resolution methods. In this paper, the task of conflict detection and resolution is defined as an optimization problem searching for a heading control for cooperating airplanes using communication. For the optimization task, an objective function integrates both collision penalties and efficiency criteria considering airplanes' objectives (waypoints). The probability collectives optimizer is used as a solver for the specified optimization task. This paper provides two different implementation approaches to the presented optimization-based collision avoidance: 1) a parallel computation using multiagent deployment among participating airplanes and 2) semicentralized computation using the process-integratedmechanism architecture.

Large-Scale High-Fidelity Agent-Based Simulation in Air Traffic Domain

  • DOI: 10.1080/01969722.2011.610270
  • Odkaz: https://doi.org/10.1080/01969722.2011.610270
  • Pracoviště: Katedra počítačů
  • Anotace:
    We present the concept of dynamic partitioning of scalable, high-fidelity multi-agent simulation complemented with intelligent load-balancing processes. The simulation framework is designed to simulate entities to high details that require extended computation resources. To be able to simulate a huge amount of entities, distributed simulation is introduced using spatial partitioning and dynamic load balancing. A novel and important feature is the combination of the synchronous and asynchronous parts in the simulation. We use the domain of the air traffic simulation to verify the simulation framework. We present a method to perform spatial and temporal planning within 3D space and multilayer architecture using several collision avoidance algorithms to illustrate the high computational demands of each airplane. The platform has been used to support simulation of an entire civilian air traffic touching the national airspace of United States.

Parallel Real-time Trajectory Planning in U.S. NAS

  • Pracoviště: Katedra počítačů
  • Anotace:
    This paper exploits the computing power of widely available multi-core machines to accelerate the trajectory planning by parallelisation of the search algorithm. In particular we investigate the approach that schedules the workload on the cores using the hashing function based on the geographical partitioning of the search space. We use this approach to parallelize the AA* algorithm. The concept is evaluated on the simulation of real-time trajectory planning of aircraft respecting the environment and real aircraft performance models. We show that this approach decreases the planning time significantly on common multi-core machines preserving the quality of the trajectory provided by AA* algorithm.

Surveillance of Unmanned Aerial Vehicles Using Probability Collectives

  • DOI: 10.1007/978-3-642-23181-0_23
  • Odkaz: https://doi.org/10.1007/978-3-642-23181-0_23
  • Pracoviště: Katedra počítačů
  • Anotace:
    A rising deployment of unmanned aerial vehicles in complex environment operations requires advanced coordination and planning methods. We address the problem of multi-UAV-based area surveillance and collision avoidance. The surveillance problem contains non-linear components and non-linear constraints which makes the optimization problem a hard one. We propose discretization of the problem based on the definition of the points of interest and time steps to reduce its complexity. The objective function integrates both the area surveillance and collision avoidance sub-problems. The optimization task is solved using a probability collection solver that allows to distribute computation of the optimization. We have implemented the probability collective solver as a multi-agent simulation. The results show the approach can be used for this problem.

Distributed Scheduling Using Constraint Optimization and Multiagent Path Planning

  • Autoři: Cannon, Ch., Lass, R., Sultanik, E., Regli, W., doc. Ing. David Šišlák, Ph.D., prof. Dr. Michal Pěchouček, MSc.,
  • Publikace: Proceedings of AAMAS 2010 The 12th International Workshop on Distributed Constraint Reasoning. Philadelphia: Drexel university, 2010, pp. 22-34. ISBN 978-0-9826571-1-9. Available from: https://www.cs.drexel.edu/files/eas28/dcr2010papers/dcr2010_8.pdf
  • Rok: 2010
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The goal of the distributed scheduling problem is to minimize the global cost of assigning n decentralized workers to m tasks at time points. This problem is further complicated in continuous environments because the entire state space cannot be searched. This paper presents a decentralized approach of dividing the distributed scheduling in continuous environments problem into two subproblems: distributed set covering and distributed multiagent path planning. First, we represent the rpblem of assigning workers (i.e., covers) to tasks (i.e., sets) as a Distributed Constraint Optimization Problem (DCOP). Then, the DCOP solver passes its solution to the distributed multiagent path planner who creates a conflict-free path for each worker to its assigned tasks. By first representing the problem as a DCOP, it restricts the plan space to only a set of feasible plans. We apply this approach to the scenario of distributed scheduling for unmanned aerial vehicle surveillance.

Iterative Accelerated A* Path Planning

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The paper provides a description of an iterative version of the Accelerated A* algorithm for path planning and its application in the air traffic domain for airplanes with defined motion dynamics operationg in the Earth-centered, Earth-fixed coordinate system (GPS) on a sperical model of the Earth constrained by the landscape and special use airspaces (SUA). The motion dynamics of the airplanes is modeled using the Base of Aircraft Data (BADA) airplane performance models. The presented algorithm provides an extension of the A* algorithm that significantly reduces the serch space and makes planning of the flight trajectories computationally tractable.

Large Scale Agent Based Simulation of Air traffic

  • Autoři: doc. Ing. David Šišlák, Ph.D., Volf, P., prof. Dr. Michal Pěchouček, MSc.,
  • Publikace: Proceedings of the Twentieth European Meeting on Cybernetics and Systems Research. Vienna: Austrian Society for Cybernetics Studies, 2010, pp. 527-532. ISBN 978-3-85206-178-8. Available from: http://www.springerlink.com/content/8kx6507332735361/
  • Rok: 2010
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    We describe a multi-agent platform \textsc{AgentFly} used for large-scale high-detail simulation of air traffic. The platform introduces spatial and temporal planning within 3D space, multi-layer architecture using several collision avoidance algorithms. To be able to simulate a huge amount of airplanes, distributed simulation is introduced using spatial partitioning and dynamic load-balancing. The platform has been used to support simulation of an entire civilian air traffic touching National Air-Space of United States. Thorough evaluation of the system has been performed, confirming that it can scale up to a very high number of complex agents operating simultaneously (thousands of aircraft) with full detailed models.

Accelerated A* Path Planning

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The paper addresses the area of path planning for nonholonomic vehicles operating in a large-scale dynamic continuous three-dimensional space where the vehicle has to avoid given obstacles and restricted areas.

Accelerated A* Trajectory Planning: Grid based Path Planning Comparison

  • Autoři: doc. Ing. David Šišlák, Ph.D., Volf, P., prof. Dr. Michal Pěchouček, MSc.,
  • Publikace: Proceedings of the 19th International Conference on Automated Planning & Scheduling/ 4th Workshop on Planning and Plan Execution for Real-World Systems. Menlo Park, California: AAAI Press, 2009, pp. 74-81. ISBN 978-1-57735-407-9.
  • Rok: 2009
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The contribution of the paper is a high performance pathplanning algorithm designed to be used within a multi-agent planning framework solving a UAV collision avoidance problem. Due to the lack of benchmark examples and available algorithms for 3D+time planning, the algorithm performance has been compared in the classical domain of path planning in grids with blocked and unblocked cells. The Accelerated A* algorithm has been compared against the Theta* path planner, Rapid-Exploring Random Trees-based planners and the original A* searching in graphs providing the shortest any-angle paths. Experiments have shown that Accelerated A* finds the shortest paths in all scenarios including many randomized configurations. Experiments document that Accelerated A* is slower than Theta* and RRT-based planers in many cases, but it is faster than the original A*.

Agent Based Approach to Free Flight Planning, Control, and Simulation

  • DOI: 10.1109/MIS.2009.1
  • Odkaz: https://doi.org/10.1109/MIS.2009.1
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Intelligent agents technology provides techniques and algorithms for distributed coordination and distributed decision making. The authors developed AgentFly, a multiagent prototype for air traffic control of free flight based operations of multiple aerial assets, based on intelligent agents. AgentFly provides mechanisms for distributed planning, negotiation based collision avoidance, and multiagent flight simulation. The US Air Force supports this project, but the Federal Aviation Administration is also studying AgentFly for planning mixed traffic of manned and unmanned air traffic.

Autonomous UAV Surveillance of Complex Urban Environments

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    We address the problem of multi-UAV surveillance in complex urban environments with occlusions. The problem consists of coordinating the flight of UAVs with on-board cameras so that the coverage and recency of the information about a designated area is maximized. In contrast to the existing work, sensing constraints due to occlusions and UAV flight constraints are modeled realistically and taken into account. We propose a novel occlusion-aware surveillance algorithm based on a decomposition of the surveillance problem into a variant of the three-dimensional art gallery problem and the multi-traveling salesmen problem for Dubins vehicles. The algorithm is thoroughly evaluated on the high-fidelity AGENTFLY UAV simulation testbed which accurately models all constraints and effects involved. The results confirm the importance of occlusion-aware flight path planning, in particular in the case of narrow street areas and low UAV flight altitudes.

Distributed Platform for Large-Scale Agent-Based Simulations

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    We describe a distributed architecture for situated largescale agent-based simulations with predominately local interactions. The approach, implemented in AglobeX Simulation platform, is based on a spatially partitioned simulated virtual environment and allocating a dedicated processing core to the environment simulation within each partition. In combination with dynamic load-balancing, such partitioning enables virtually unlimited scalability of the simulation platform. The approach has been used to extend the AgentFly air-traffic testbed to support simulation of a complete civilian air-traffic touching National Air-Space of United States. Thorough evaluation of the system has been performed, confirming that it can scale up to a very high number of complex agents operating simultaneously (thousands of aircraft) and determining the impact of different configurations of the simulation architecture on its overall performance.

Flight Trajectory Path Planning

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The paper describes an application of the A* algorithm for flight path planning for airplanes with defined motion dynamics operating in a continuous three-dimensional space constrained by existing physical obstacles. The presented A* algorithm modification provides significant acceleration (reduction of the state space) of the path planning process. The described algorithm is able to find a path through small gaps between obstacles using a pre-defined searching precision. The paper documents a set of path planning benchmarks where the solution quality and internal algorithm properties are compared against the A* algorithm for flight trajectory path planning.

Integration of Probability Collectives for Collision Avoidance in AGENTFLY

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Rising deployment of Unmanned Aerial Assets in komplex operations requires innovative automatic coordination algorithms, especially for decentralized collision avoidance. The paper investigates a stochastic Probability Collectives (PC) optimizer for the collision avoidance problem defined as an optimization task where efficiency criteria, collision penalties and airplanes' missions are integrated in an objektive function. The paper provides two different implementation approaches of the proposed method - complex multiagent deployment distributes the PC optimization process among airplanes and the Process Integrated Mechanism approach replaces communication with optimizer migration. Both implementations have been developed and tested on an airspace multi-agent framework AGENTFLY.

Optimization based Collision Avoidance for Cooperating Airplanes

  • Autoři: doc. Ing. David Šišlák, Ph.D., Volf, P., prof. Dr. Michal Pěchouček, MSc., Suri, N., Nicholson, D., Woodhouse, D.
  • Publikace: 2009 IEEE/WIC/ACM INTERNATIONAL JOINT CONFERENCES ON WEB INTELLIGENCE (WI) AND INTELLIGENT AGENT TECHNOLOGIES (IAT), VOL 2. Los Alamitos: IEEE Computer Society, 2009, pp. 375-378. ISBN 978-1-4244-5331-3.
  • Rok: 2009
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Application of autonomous intelligent systems into airspace domain is very important nowadays. The paper presents decentralized collision avoidance algorithm utilizing a solution of the defined optimization problem where efficiency criteria, collision penalties and airplanes' missions are integrated in an objective function. Two different implementation approaches used for stochastic Probability Collectives optimizer are presented and evaluated - a complex distributed multi-agent deployment among participating airplanes and the Process Integrated Mechanism inspired architecture. Both approaches have been validated and evaluated on the multi-agent framework AGENTFLY providing precise simulation for airspace operations.

AGENTFLY: A Multi-Agent Airspace Test-bed

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The contribution presents a multi-agent technology in the domain of the air-traffic control of several autonomous seriál vehicles (manned as well as unmanned). The system has been validated mainly by the collision avoidance (CA) tasks. Several cooperative and non-cooperative CA methods have been integrated in the system to validate and compare thein properties in the scalable experiments. The AGENTFLY system is suitable also for the developing and testing of algorithms for the collective flight control. The operation of the underlying multi-agent system has been integrated with freely available geographical and tactical data sources. The system provides real time 2D/3D visualization and also the web-access component.

AGENTFLY: Towards Multi-Agent Technology in Free Flight Air Traffic Control

  • Autoři: doc. Ing. David Šišlák, Ph.D., prof. Dr. Michal Pěchouček, MSc., Volf, P., Pavlíček, D., Samek, J., Mařík, V., Losiewicz, P.
  • Publikace: Defense Industry Applications of Autonomous Agents and Multi-Agent Systems. Heidelberg: Birkhäuser Verlag AG, 2008. p. 73-96. Whitestein Series in Software Agent Technologies and Autonomic Computing. ISBN 978-3-7643-8570-5.
  • Rok: 2008
  • DOI: 10.1007/978-3-7643-8571-2_5
  • Odkaz: https://doi.org/10.1007/978-3-7643-8571-2_5
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Ever rising deployment of Unmanned Aerial Assets (UAAs) in complex military and rescue operations require novel and innovative methods for intelligent planning and collision avoidance among a high number of heterogeneous, semi-trusted flying assets in well specified and constrained areas [1]. We have studied the free flight concept as an alternative to the classical, centralized traffic control. In free flight the unmanned aerial assets are provided with flight trajectory that has been elaborated without consideration of other flying objects that may occupy the same air space. The collision threads are detected by each of the aircraft individually and the collisions are avoided by an asset-to-asset negotiation. Multi-agent technology is very well suited as a technological platform for supporting the free-flight concept among the heterogeneous UAAs. In this chapter we present AGENTFLY, multi-agent system for free-flight simulation and flexible collision avoidance.

Decentralized Algorithms for Collision Avoidance in Airspace

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The paper proposes decentralized deconfliction algorithms deployed on multiple autonomous aerial vehicles in freeflight operations. The paper provides two separate algorithms for collision avoidance - one based on the iterative peer-to-peer negotiation solving a singular collision and second based on multi-party negotiation about a cluster of collisions. The presented decentralized algorithms allow the vehicles operating in the same area to utilize the given airspace more efficiently. The algorithms have been developed and tested on a multi-agent prototype and the properties of both algorithms are discussed on a set of large scale experiments.

Agent-Based Multi-Layer Collision Avoidance to Unmanned Aerial Vehicles

  • DOI: 10.1109/KIMAS.2007.369837
  • Odkaz: https://doi.org/10.1109/KIMAS.2007.369837
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This contribution presents a distributed, multi-layer collision avoidance architecture supporting efficient utilization of air space shared by several autonomous aerial vehicles. Presented multi-layer architecture is based on deliberative deployment of several collision avoidance methods by the aircraft at the same time. Both cooperative and non-cooperative collision avoidance methods are presented in the paper. The robustness of the architecture is justified by means of experimental validation of multi-agent simulation.

Collision Avoidance Algorithms: Multi-agent Approach

  • Autoři: Vrba, P., Mařík, V., Přeučil, L., Kulich, M., doc. Ing. David Šišlák, Ph.D.,
  • Publikace: Holonic and Multi-Agent Systems for Manufacturing - HoloMAS 2007. Munich: Springer, 2007. p. 348-360. ISSN 0302-9743. ISBN 978-3-540-74478-8.
  • Rok: 2007
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The paper deals with the methods for detection and avoidance of collisions of autonomously moving vehicles utilizing the principles and techniques of multi-agent systems. Three different scenarios are discussed: (i) movement of AGVs in 2D space with fixed trajectories, (ii) movement of autonomous robots in an open 2D space and (iii) collision-free flights of unmanned aerial vehicles. For each category, an agent-based solution is proposed. Presented experiments show that the cooperative approach to detecting and avoiding collisions based on negotiation and goal sharing of agents, representing vehicles, seems to be highly efficient. The multi-layer architecture combining cooperative approach with algorithms based on dynamic no-access zones for avoiding non-cooperative vehicles provides good results in generating collision-free flight corridors.

Convergence of Peer to Peer Collision Avoidance among Unmanned Aerial Vehicles

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this article we study the theoretical aspects of the collision avoidance among the collectives of unmanned aerial vehicles (UAVs) engaged in the free flight operation. The free flight based operations do not provide collision-free trajectories as a flight specification to the UAVs. Collision avoidance of the trajectories is implemented by means of peer-to-peer negotiations among the individual UAVs. In this paper we study the role of interaction in the collision avoidance process. We prove theoretically convergence of the specific negotiating protocol that has been deployed in a practical implementation of the software prototype of the distributed collision avoidance system.

Multi party Collision Avoidance among Unmanned Aerial Vehicles

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper addresses the problem of distributed cooperative collision avoidance that supports efficient utilization of air space shared by several autonomous unmanned aerial vehicles. The novel multi-party collision avoidance (MPCA) algorithm is described. It is compared to the iterative peer-to-peer collision avoidance (IPPCA) algorithm that iteratively optimizes social welfare. The paper provides a set of experiments and a comparison of different collision avoidance mechanisms in a multi-agent model of air traffic.

Od osamocených robotů ke kolaborativní robotice

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Od osamocených robotů ke kolaborativní robotice

Autonomous Agents for Air Traffic Deconfliction

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This contribution presents a deployment exercise of multi-agent technology in the domain of deconflicted air-traffic control among several autonomous aerial vehicles. Negotiation based deconfliction algorithm have been developed and integrated in the agent-based model of the individual flight. Operation of the underlying multi-agent system has been integrated with freely available, geographical and tactical data sources in order to demonstrate openness of the technology. An additional, web client visualization and access component has been developed in order to facilitate a multi-user, platform independent use of the system.

Deployment of A globe Multi Agent Platform

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    A-globe is a freeware, simulation-oriented multi-agent platform featuring agent migration, communication inaccessibility simulation and high scalability with moderate hardware requirements. The paper presents several deployment of A-globe platform.

Negotiation Based Approach to Unmanned Aerial Vehicles

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    We present a framework for agent based aircraft deconfliction mechanism to enable efficient airspace use by various UAVs during coalition operations. In our approach, each vehicle is autonomous, but cooperative: it actively shares its flight plan with near aircrafts so that potential collisions can be detected and resolved using norm-based system. Non-cooperative and utility-based deconfliction approaches are also discussed as they offer a possibility to achieve more efficient and robust mechanism in the future. System is validated on multi-agent simulation that uses the public online-accessible data from various information sources.

Stand in Agents in Large scale Multi agent Environment with Dynamic Communication Inaccessibility

  • Autoři: doc. Ing. David Šišlák, Ph.D.,
  • Publikace: Proceedings of Workshop 2006. Praha: České vysoké učení technické v Praze, 2006, pp. 132-133. ISBN 80-01-03439-9.
  • Rok: 2006
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Autonomous decentralized algorithm for message relaying in dynamic Ad-hoc network. Agents use the social dominance models and virtual palyments.

A-globe: Agent Development Platform with Inaccessibility and Mobility Support

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    At present several Java-based multi-agent platforms from different developers are available, but none of them fully supports agent mobility and communication inaccessibility simulation. They are thus unsuitable for experiments with large scale real-world simulation. In this chapter we describe architecture of A-globe, fast, scalable and lightweight agent development platform with environmental simulation and mobility support. Beside the functions common to most agent platforms it provides a position-based messaging service, so it can be used for experiments with extensive environment simulation and communication inaccessibility. Simple benchmarks that compare the A-globe performance against other available agent platforms are also included.

A-globe: Multi-agent Platform with Advanced Simulation and Visualization Support

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    A-Globe is a simulation oriented multi-agent platform featuring agent migration, communication inaccessibility simulation and high scalability with moderate hardware requirements. Using dedicated simulation messaging together with 2D and 3D visualization support, large agent systems can be engineered, tested and visualized on a single machine. A-globe agents are fully fledged JAVA agents, each with its own independent thread, that can autonomously migrate between platforms running on different hosts. Thanks to the separation of simulation and agent code, deployment of agents to embedded devices is straightforward. Platform is not natively FIPA-compliant, as the interoperability was sacrificed to support the scalability and efficiency.

Optimizing Agents Operation in Partially Inaccessible and Disruptive Environment

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The algorithm we present in this paper aims to optimally distribute and connect the community of loosely coupled middle agents ensuring communication accessibility in a dynamic, inaccessible environment. Complete decentralization, autonomous adaptation to local and global changes in the environment and domain independence are main features of the presented algorithm. The main motivation of this approach is to significantly decrease the number of messages required for communication relaying in inaccessible system operation. We use social dominance models and virtual payments to obtain such behavior.

Solving Inaccessibility in Multi-Agent Systems By Mobile Middle-Agents

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This paper analyzes the problem of communication inaccessibility in the multi-agent systems. Apart from listing the main reasons for existence of communication inaccessibility, it suggest a complete inaccessibility model and metrics. Various inaccessibility solutions are presented, together with their applicability in the environments with different degrees of inaccessibility, as verified by experiments. Major contribution of this paper is in suggesting a novel solution to solving inaccessibility based on community of autonomous, migrating middle-agents. A distributed algorithm for dynamic allocation of the middle-agents in a network of partially inaccessible agents is proposed. The suggested solution is supported by a set of experiments comparing the distributed and centralized algorithm for middle-agents allocation.

System Innovation Award at CIA2004, Erfurt

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    The A-Globe platform is primarily aimed at large scale, real world simulations with fully fledged agents. To support this goal, it includes a special infrastructure for envirnmental simulation.

Using Stand-In Agents in Partially Accessible Multi-Agent Environment

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This contribution defines a metrics and proposes a solution for the problem of agents inaccessibility in multi-agent systems. We define the stand-in pattern for knowledge maintenance and remote presence in distributed agent systems with communication inaccessibility. Our implementation has been designed and tested in the A-globe agent platform. We also present a set of measurements quantifying agents' inaccessibility in our domain and comparing the usefulness of different solution in the environments with different inaccessibility

A-Globe: Agent Platform with Inaccessibility and Mobility Support

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    In this paper we describe architecture of newly developed agent platform A-Globe. It is fast and lightweight platform with agent mobility support. Beside the functions common to most of agent platforms it provides the Geographical Information System service to user, so it can be used for experiments with environment simulation and communication inaccessibility. A-Globe performance benchmarks compared against other agent platforms are also stated in this paper.

Metrics of Inaccessibility in MAS

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Within our previous work the problem of agents inaccessibility has been properly studied, the concept of agent's acquaintance models and their use in the partially inaccessible environment has been investigated, the model of the stand-in agent has been designed and implemented and the A-Globe agent interaction platform has been developed. In this paper we will metrics in order to select appropriate method to overcome inaccessibility itself.

Simulating Agents´ Mobility and Inaccessibility with A-Globe Multi-agent System

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Simulating Agents´ Mobility and Inaccessibility with A-Globe Multi-agent System

Solutions for Cooperation with Limited Communication

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    We define a set of metrics for communication inaccessibility measurements, together with a comparison of existing solutions for inaccessibility in a MAS. We compare the solutions such as relaying, middle agents, stand-in agents and analyze the performance of these solutions for various values of inaccessibility properties. The measurements are obtained on a test domain implemented using the A-Globe agent platform with inaccessibility support.

Solving Communication Inaccessibility in Coalition Operations

  • Pracoviště: Katedra kybernetiky
  • Anotace:
    Inability to communicate the exact information in the right time has become one of the key bottlenecks in real-time control of military operations as well as in planning large scale coalition operations. In this paper we study situations in which communication and coordination inaccessibility can occur and discuss three different solutions of various level of sophistication -- relay agents, acquaintance models and stand-in agents. These methods were deployed in the aglobe multi-agent system. Deployment in two different scenarios is discussed here.

Control of Elevator

  • Autoři: doc. Ing. David Šišlák, Ph.D.,
  • Publikace: Poster 2003. Praha: České vysoké učení technické v Praze, Fakulta elektrotechnická, 2003. p. IC 33. ISSN 0277-786X. ISBN 0-8194-5368-4.
  • Rok: 2003
  • Pracoviště: Katedra kybernetiky
  • Anotace:
    This contribution presents research that was initiated by diploma thesisControl of Elevator. The physical model of elevator is used for training students of its control now. The model consist of a cabin, buttons for calling the cabin into the relevant floor and buttons for choosing the destination. The cabin can move up and down in the range of four floors. The physical model is connected to the Programmable Logic Controller (PLC). Students create control algorithms for elevator. This concept has some disadvantages: model parameters can't be changed, simulation of passengers is made by student himself. I propose another way for training control of elevator with the virtual simulator of elevator on the computer.

Za stránku zodpovídá: Ing. Mgr. Radovan Suk