Learning to falsify automated driving vehicles with prior knowledge

TL;DR

This paper introduces a learning-based falsification framework for testing automated driving functions in simulation, leveraging prior knowledge to efficiently identify scenarios that violate safety specifications.

Contribution

It combines prior knowledge with learning to improve falsification efficiency and effectiveness in automated driving systems.

Findings

Higher reward falsifying scenarios achieved

Outperforms purely learning-based methods

Outperforms purely model-based methods

Abstract

While automated driving technology has achieved a tremendous progress, the scalable and rigorous testing and verification of safe automated and autonomous driving vehicles remain challenging. This paper proposes a learning-based falsification framework for testing the implementation of an automated or self-driving function in simulation. We assume that the function specification is associated with a violation metric on possible scenarios. Prior knowledge is incorporated to limit the scenario parameter variance and in a model-based falsifier to guide and improve the learning process. For an exemplary adaptive cruise controller, the presented framework yields non-trivial falsifying scenarios with higher reward, compared to scenarios obtained by purely learning-based or purely model-based falsification approaches.