Towards Physically Adversarial Intelligent Networks (PAINs) for Safer Self-Driving
- 19 December 2022
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Control Systems Letters
- Vol. 7, 1063-1068
Abstract
Neural networks in autonomous vehicles suffer from overfitting, poor generalizability, and untrained edge cases due to limited data availability. Researchers often synthesize randomized edge-case scenarios to assist in the training process, though simulation introduces the potential for overfitting to latent rules and features. Automating worst-case scenario generation could yield informative data for improving self-driving. To this end, we present a “Physically Adversarial Intelligent Network", wherein self-driving vehicles interact aggressively in the CARLA simulation. We train two agents, a protagonist, and an adversary, using dueling double deep Q networks with prioritized experience replay. The coupled networks alternately seek to collide and avoid collisions such that the “defensive” avoidance algorithm increases the mean time to failure and distance traveled under non-hostile operating conditions. The trained protagonist becomes more resilient to environmental uncertainty and less prone to corner case failures resulting in collisions than the agent trained without an adversary.Keywords
This publication has 12 references indexed in Scilit:
- Learning to Collide: An Adaptive Safety-Critical Scenarios Generating Method2020
- Generation of Driving Scenario Trajectories with Generative Adversarial Networks2020
- SADA: Semantic Adversarial Diagnostic Attacks for Autonomous ApplicationsProceedings of the AAAI Conference on Artificial Intelligence, 2020
- Generating Adversarial Driving Scenarios in High-Fidelity Simulators2019
- Risk Averse Robust Adversarial Reinforcement Learning2019
- DeepRoad: GAN-based metamorphic testing and input validation framework for autonomous driving systems2018
- End-to-End Driving Via Conditional Imitation Learning2018
- GAN(Generative Adversarial Nets)Journal of Japan Society for Fuzzy Theory and Intelligent Informatics, 2017
- Mask R-CNN2017
- Deep Reinforcement Learning framework for Autonomous DrivingElectronic Imaging, 2017