Long-term Microscopic Traffic Simulation with History-Masked Multi-agent Imitation Learning

TL;DR

This paper introduces a history-masked multi-agent imitation learning approach for long-term microscopic traffic simulation, effectively addressing covariate shift and improving realism in urban traffic modeling.

Contribution

The paper proposes a novel history-masked imitation learning method that enhances long-term stability and accuracy in traffic simulation by removing historical trajectory dependence.

Findings

Achieves better microscopic similarity to real-world data.

Improves long-term macroscopic traffic pattern accuracy.

Outperforms state-of-the-art baselines on pNEUMA dataset.

Abstract

A realistic long-term microscopic traffic simulator is necessary for understanding how microscopic changes affect traffic patterns at a larger scale. Traditional simulators that model human driving behavior with heuristic rules often fail to achieve accurate simulations due to real-world traffic complexity. To overcome this challenge, researchers have turned to neural networks, which are trained through imitation learning from human driver demonstrations. However, existing learning-based microscopic simulators often fail to generate stable long-term simulations due to the \textit{covariate shift} issue. To address this, we propose a history-masked multi-agent imitation learning method that removes all vehicles' historical trajectory information and applies perturbation to their current positions during learning. We apply our approach specifically to the urban traffic simulation problem and evaluate it on the real-world large-scale pNEUMA dataset, achieving better short-term microscopic and long-term macroscopic similarity to real-world data than state-of-the-art baselines.