Ing. David Vošahlík, Ph.D.

Advanced driver assistant systems in vehicles, such as anti-lock brake system, electronic stability program, and traction control, heavily rely on the traction force of the wheels, which is determined by the wheels’ longitudinal slip ratio. A common method used to describe this dependency is the well-known Pacejka (magic) formula. The Pacejka (magic) formula suggests the existence of a unique optimal slip ratio, denoted as lambda_opt, for which the maximum traction force is achieved. Although the lambda_opt value slightly varies with changes in the tire-to-road interface, most studies tend to neglect the variation of lambda_opt due to its relatively low impact on the introduced error. Instead, most studies focus on estimating the maximum friction coefficient mu_max . In this paper, we address this oversimplification and propose an estimation method for lambda_opt. This paper introduces two novel approaches for real-time estimation of the slip ratio lambda_opt based on wheel dynamics. Both approaches were derived using a nonlinear twin-track model implemented in MATLAB & Simulink. The first approach uses recursive least squares, whereas the second approach employs the Unscented Kalman filter algorithm. In addition, a traction force estimator is presented because both estimators rely on either a measurement or an estimation of the traction force. To validate the performance of the proposed estimators, a high-fidelity twin-track vehicle model was initially employed. Finally, real-world experiments are presented in which the proposed algorithms are validated using an RC subscale platform.

This paper uses the brake control allocation method for Electric Vehicles (EVs) based on system-level vehicle Reference Point (RP) motion feedback. The RP motion control is an alternative to the standard brake torque allocation methods and results in improved vehicle stability in both longitudinal and lateral directions without requiring additional measurements beyond what is available in EVs with ABS and ESP. The proposed control law simplifies the brake torque allocation algorithm, reduces overall development time and effort, and merges most of the braking systems into one. Additionally, the measured or estimated signals required are reduced compared to the standard approach. The system-level RP measurements and references are transformed into individual wheel coordinate systems, where tracking is ensured by actuating both friction torques and electric motor regenerative torques using a proposed brake torque blending mechanism. The whole control system is validated in simulations using the Simulink vehicle dynamic simulator and IPG Carmaker high-fidelity simulator, utilizing the CTU FEE EFORCE formula model. Finally, a brake HiL stand validation is presented.

A novel vehicle traction algorithm solving the traction force allocation problem based on vehicle center point motion feedback controller is proposed in this paper. The center point motion feedback control system proposed utilizes individual wheel torque actuation assuming all wheels are individually driven. The approach presented is an alternative to the various direct optimization-based traction force/torque allocation schemes. The proposed system has many benefits, such as significant reduction of the algorithm complexity by merging most traction system functionalities into one. Such a system enables significant simplification, unification, and standardization of powertrain control design. Moreover, many signals needed by conventional traction force allocation methods are not required to be measured or estimated with the proposed approach, which are among others vehicle mass, wheel loading (normal force), and vehicle center of gravity location. Vehicle center point trajectory setpoints and measurements are transformed to each wheel, where the tracking is ensured using the wheel torque actuation. The proposed control architecture performance and analysis are shown using the nonlinear twin-track vehicle model implemented in Matlab & Simulink environment. The performance is then validated using high fidelity FEE CTU in Prague EFORCE formula model implemented in IPG CarMaker environment with selected test scenarios. Finally, the results of the proposed control allocation are compared to the state-of-the-art approach.

State trajectory planning is recently one of the main challenges for self driving/autonomous vehicles technology. The Minimum Violation Planing (MVP) approach provides a reliable and user intuitive framework for states trajectory planning. It provides possibility of various logic constraints, boundary conditions, dynamic constraints, and many others. The MVP is revisited in this paper with a special focus on the time efficiency of the algorithm. Two modifications of the MVP reducing the calculation time while not significantly compromising the trajectory quality are proposed in this paper. The vehicle yaw, yaw rate, north and east position, velocity, and battery state of charge variables planning is selected to compare the proposed planning framework modifications. The presented modifications and the original MVP algorithm are compared in selected test scenario, where significant calculation time reduction is shown while the plan optimality is not affected remarkably

State trajectory planning is one of the primary self-driving cars technology enablers. However, state trajectory planning is a more complex and computationally demanding task compared to path planning. The vehicle’s east and north position, yaw, yaw rate, velocity, and battery state of charge variables trajectory planning with a particular focus on the safety and economy of the vehicle operation is concerned in this paper. Comparison of Model Predictive Control (MPC) and Minimum Violation Planning (MVP) used for trajectory planning is brought in this paper. The latter is a sampling-based algorithm based on the RRT* algorithm compared to the other optimization-based algorithm. A heuristic is introduced to convert the complex non-convex optimization planning task to a convex optimization problem. Next, MVP algorithm enhancement is proposed to reduce the calculation time. Both algorithms are tested on a selected testing scenario using a high fidelity nonlinear single-track model implemented in Matlab & Simulink environment.

The visual predictor of a drivable surface friction ahead of the vehicle is presented. The image recognition neural network is trained in self-supervised fashion, as an alternative to tedious, error-prone, and subjective human annotation. The training images are labelled automatically by surface friction estimates from vehicle response during ordinary driving. The Unscented Kalman Filter algorithm is used to estimate tire-to-road interface friction parameters, taking into account the highly nonlinear nature of tire dynamics. Finally, the overall toolchain was validated using an experimental subscale platform and real-world driving scenarios. The resulting visual predictor was trained using about 3 000 images and validated on an unseen set of 800 test images, achieving 0.98 crosscorrelation between the visually predicted and the estimated value of surface friction.

Trend of autonomous vehicles and e-mobility is in favor of an advanced control system development and deployment. Vehicle dynamics level control systems providing safety limits and high performance response, especially during high dynamics maneuvers, are necessary. This work provides solution for vehicle longitudinal dynamics (vehicle acceleration) considering physical limits given by road, tire and vehicle dynamics respectively. The goal is to maximize vehicle longitudinal acceleration by controlling each wheel longitudinal slip ratio. Considered mathematical model is non-linear single track model incorporating non-linear Pacejka magic formula as a tire model. Design model for control system is derived as a linearized state-space model at constant acceleration operation point. Therefore, the common linearization approach, at system equilibrium, is not possible and the linearization along system trajectory is used. Such solution results in involvement of LPV techniques, as vehicle velocity is state variable. Finally, the LQ optimal control framework is employed to deliver control algorithms providing constant vehicle acceleration trajectory tracing. This is accomplished by longitudinal slip ratio control for each wheel.