Verification of Collision Avoidance for CommonRoad Traffic Scenarios

11 pages•Published: December 6, 2021

Niklas Kochdumper, Philipp Gassert and Matthias Althoff

Abstract

We propose a benchmark for the verification of autonomous vehicles. By considering different traffic scenarios from the CommonRoad database, we obtain several thousands of different verification tasks, where the verification problem is to prove that the con- sidered tracking controller safely follows a given reference trajectory despite disturbances and measurement errors. The dynamic of the car is described by a nonlinear kinematic single-track model. Since the feedback matrix for the tracking controller is time-varying, the dynamic of the controlled system changes constantly. Because of this, the proposed benchmark is well-suited to evaluate how robustly reachability tools can handle changing system dynamics.

Keyphrases: autonomous vehicles, collision avoidance, control theory

In: Goran Frehse and Matthias Althoff (editors). 8th International Workshop on Applied Verification of Continuous and Hybrid Systems (ARCH21), vol 80, pages 184-194.

Links:	https://easychair.org/publications/paper/cqFP
	https://doi.org/10.29007/1973

BibTeX entry

@inproceedings{ARCH21:Verification_Collision_Avoidance_CommonRoad,
  author    = {Niklas Kochdumper and Philipp Gassert and Matthias Althoff},
  title     = {Verification of Collision Avoidance for CommonRoad Traffic Scenarios},
  booktitle = {8th International Workshop on Applied Verification of Continuous and Hybrid Systems (ARCH21)},
  editor    = {Goran Frehse and Matthias Althoff},
  series    = {EPiC Series in Computing},
  volume    = {80},
  publisher = {EasyChair},
  bibsource = {EasyChair, https://easychair.org},
  issn      = {2398-7340},
  url       = {/publications/paper/cqFP},
  doi       = {10.29007/1973},
  pages     = {184-194},
  year      = {2021}}

Download PDF Open PDF in browser