Imitating driver behavior with generative adversarial networks

TLDR

Accurately predicting and simulating human driving behavior is essential for intelligent transportation systems, yet traditional methods rely on simple parametric models and behavioral cloning. The paper aims to overcome cascading errors in prior approaches to produce realistic, trajectory‑perturbation‑robust driving behavior. The authors extend Generative Adversarial Imitation Learning to train recurrent policies, thereby mitigating cascading errors and producing realistic driving behavior. The model rivals rule‑based controllers and maximum likelihood models in realistic highway simulations, reproducing emergent human behaviors such as lane‑change rates while maintaining realistic control over long horizons.

Abstract

The ability to accurately predict and simulate human driving behavior is critical for the development of intelligent transportation systems. Traditional modeling methods have employed simple parametric models and behavioral cloning. This paper adopts a method for overcoming the problem of cascading errors inherent in prior approaches, resulting in realistic behavior that is robust to trajectory perturbations. We extend Generative Adversarial Imitation Learning to the training of recurrent policies, and we demonstrate that our model rivals rule-based controllers and maximum likelihood models in realistic highway simulations. Our model both reproduces emergent behavior of human drivers, such as lane change rate, while maintaining realistic control over long time horizons.