Regulatory approval and safety guarantees for autonomous vehicles facing frequent functional updates and complex software stacks, including artificial intelligence, are a challenging topic. This paper proposes a concept and guideline for the development of an online verification module -- the Supervisor -- capable of handling the aforementioned challenges. The concept presented for the establishment of a Supervisor is designed in a way to identify and monitor an extensive list of features contributing to safe operation. As a result, a safe overall (sub)system is attained. Safeguarding a motion planner of an autonomous race vehicle is used to illustrate the procedure and practicability of the framework at hand. The capabilities of the proposed method are evaluated in a scenario-based test environment and on full-scale vehicle data.
@article{arxiv.2005.07740,
title = {Online Verification Concept for Autonomous Vehicles -- Illustrative Study for a Trajectory Planning Module},
author = {Tim Stahl and Matthis Eicher and Johannes Betz and Frank Diermeyer},
journal= {arXiv preprint arXiv:2005.07740},
year = {2020}
}
Comments
Accepted at The 23rd IEEE International Conference on Intelligent Transportation Systems, September 20 - 23, 2020