Development of A Stochastic Traffic Environment with Generative Time-Series Models for Improving Generalization Capabilities of Autonomous Driving Agents

TL;DR

This paper introduces a GAN-based traffic simulator trained on real data to enhance the generalization of autonomous driving agents, addressing the limitations of simple simulators in realistic scenarios.

Contribution

The paper presents a novel GAN-based traffic environment that improves the realism of traffic simulations for training autonomous driving agents.

Findings

RL agents trained on GAN-based simulator generalize better to real-world scenarios.

GAN-generated traffic trajectories resemble real traffic interactions.

Enhanced simulation realism leads to improved autonomous driving policy robustness.

Abstract

Automated lane changing is a critical feature for advanced autonomous driving systems. In recent years, reinforcement learning (RL) algorithms trained on traffic simulators yielded successful results in computing lane changing policies that strike a balance between safety, agility and compensating for traffic uncertainty. However, many RL algorithms exhibit simulator bias and policies trained on simple simulators do not generalize well to realistic traffic scenarios. In this work, we develop a data driven traffic simulator by training a generative adverserial network (GAN) on real life trajectory data. The simulator generates randomized trajectories that resembles real life traffic interactions between vehicles, which enables training the RL agent on much richer and realistic scenarios. We demonstrate through simulations that RL agents that are trained on GAN-based traffic simulator has stronger generalization capabilities compared to RL agents trained on simple rule-driven simulators.