Rigorous Simulation-based Testing for Autonomous Driving Systems -- Targeting the Achilles' Heel of Four Open Autopilots

TL;DR

This paper introduces a rigorous simulation-based testing method for autonomous driving systems that identifies critical failure scenarios by analyzing the decision-making process and traffic rule application, revealing significant safety issues in leading autopilots.

Contribution

The paper presents a novel, systematic approach to generate realistic, critical test cases for autonomous driving systems by analyzing their decision-making as a dynamic system, improving defect detection.

Findings

Identified major defects in Apollo, Autoware, Carla, and LGSVL autopilots.

Test cases reveal accidents, failures, and rule violations difficult to detect by random testing.

Results align with real-world safety concerns, indicating systems need further improvement.

Abstract

Simulation-based testing remains the main approach for validating Autonomous Driving Systems. We propose a rigorous test method based on breaking down scenarios into simple ones, taking into account the fact that autopilots make decisions according to traffic rules whose application depends on local knowledge and context. This leads us to consider the autopilot as a dynamic system receiving three different types of vistas as input, each characterizing a specific driving operation and a corresponding control policy. The test method for the considered vista types generates test cases for critical configurations that place the vehicle under test in critical situations characterized by the transition from cautious behavior to progression in order to clear an obstacle. The test cases thus generated are realistic, i.e., they determine the initial conditions from which safe control policies are possible, based on knowledge of the vehicle's dynamic characteristics. Constraint analysis identifies the most critical test cases, whose success implies the validity of less critical ones. Test coverage can therefore be greatly simplified. Critical test cases reveal major defects in Apollo, Autoware, and the Carla and LGSVL autopilots. Defects include accidents, software failures, and traffic rule violations that would be difficult to detect by random simulation, as the test cases lead to situations characterized by finely-tuned parameters of the vehicles involved, such as their relative position and speed. Our results corroborate real-life observations and confirm that autonomous driving systems still have a long way to go before offering acceptable safety guarantees.