In this thesis an approach to verify existing collision avoidance algorithms is proposed. At first the mathematical definition of a car traffic scenario where a conflict situation may occur is given, referencing the structure of an optimal control problem. By solving the optimal control problem using OCPID-DAE1 software and ROC-HJ software, reference vehicle trajectories and sets of points reached by the reference vehicle, are computed. Moreover the implementation of a sensitivity analysis predicts effects of measurements errors on the trajectory or on the reachable set. Furthermore, the maximum possible sensor error for a given reference vehicle trajectory is estimated. The second step applies such techniques to construct a tool for verification of collision avoidance algorithms. Examples of verification procedures are given, based on the computation of the closest distance to an object for a rear-end collision scenario, or based on the computation of the safe set of initial states from which is possible for the reference vehicle to avoid a collision.
Verification of Collision Avoidance Systems using Optimal Control and Sensitivity Analysis
2015-01-01
Xausa, Ilaria: Verification of Collision Avoidance Systems using Optimal Control and Sensitivity Analysis. 2015
Theses
Electronic Resource
English
Logic Verification of Collision Avoidance System in Train Control Systems
British Library Conference Proceedings | 2009
|European Patent Office | 2020
|