Adversarial Testing with Reinforcement Learning: A Case Study on Autonomous Driving
Authors:
- 27 May 2024
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 293-304
Abstract
Testing autonomous driving systems (ADSs) is essential to ensure their safety. Existing testing techniques manipulate the objects of the driving environment in order to trigger a misbehavior of the ADS under test. Reinforcement learning (RL) approaches have been applied to effectively modify the dynamic objects of the environment (e.g., pedestrians and other vehicles), also known as Non-Playable Characters (NPCs). However, existing RL approaches implement centralized controllers of the environment, resulting in possibly unrealistic and even invalid behaviors of the NPCs. In this paper, we propose to model NPCs as independent and fully autonomous agents to challenge the ADS under test (i.e., the ego ADS). In the first step of our approach, we train an adversarial ADS by designing a reward function as a linear combination of two components: (1) a component that encourages it to drive well in the given driving scenario, and (2) an adversarial component, that smoothly guides the agent towards a collision with the ego ADS. In our second step, we resume training of the ego ADS, to increase its robustness towards the behaviors of the adversarial ADS. Our experiments on a highway driving scenario show that the adversarial ADS is significantly more effective at generating collisions of the ego ADS than a random baseline. Moreover, adversarial retraining induces safe behaviors of the ego ADS, preventing the adversarial ADS from colliding.Keywords
This publication has 33 references indexed in Scilit:
- Many-Objective Reinforcement Learning for Online Testing of DNN-Enabled Systems2023
- Adversarial Deep Reinforcement Learning for Improving the Robustness of Multi-agent Autonomous Driving Policies2022
- Quick Learner Automated Vehicle Adapting its Roadmanship to Varying Traffic Cultures with Meta Reinforcement Learning2021
- Gym-µRTS: Toward Affordable Full Game Real-time Strategy Games Research with Deep Reinforcement Learning2021
- Intelligent driving intelligence test for autonomous vehicles with naturalistic and adversarial environmentNature Communications, 2021
- AV-FUZZER: Finding Safety Violations in Autonomous Driving Systems2020
- Generating Avoidable Collision Scenarios for Testing Autonomous Driving Systems2020
- Human-level control through deep reinforcement learningNature, 2015
- A Hitchhiker's guide to statistical tests for assessing randomized algorithms in software engineeringSoftware Testing, Verification and Reliability, 2012
- Congested traffic states in empirical observations and microscopic simulationsPhysical Review E, 2000