Flow-ERD: Agent-type Aware Flow Matching with Entropy-Regularized Distillation for Diverse Traffic Simulation

2026-07-07

AI Summary

Existing traffic simulation methods struggle to simultaneously achieve realism and diversity, limiting thorough validation of autonomous driving systems. This work proposes Flow-ERD, a multi-agent simulation framework that jointly optimizes both objectives for the first time. It introduces Agent-Type Aware Flow Matching to model type-aware multimodal motion distributions and incorporates Entropy-Regularized Distillation for closed-loop trajectory refinement, effectively mitigating mode collapse and enhancing fine-grained behavioral variation. By integrating flow matching, inverse KL divergence optimization, and log-free diversity evaluation, the method attains Pareto optimality among reproducible approaches on the WOSAC benchmark, significantly outperforming current state-of-the-art solutions.

Abstract

Realistic and diverse traffic simulation is essential to autonomous driving development. Yet prevailing benchmarks predominantly reward realism, and recent methods have optimized accordingly, leaving diversity underexplored. We introduce \textbf{Flow-ERD}, a multi-agent simulator that pursues realism and diversity jointly. Its backbone, \textbf{Agent-Type Aware Flow Matching} (AFM), couples flow matching's multi-modal expressiveness with type-specific kinematic execution. It preserves fine-grained diversity while keeping motions consistent with each agent type. A second stage, \textbf{Entropy-Regularized Distillation} (ERD), fine-tunes the closed-loop rollout distribution with an entropy-regularized reverse-KL objective. This mitigates covariate shift while explicitly preventing collapse onto high-density modes. We evaluate Flow-ERD with a log-free diversity metric alongside standard realism scores. Flow-ERD ranks first on the WOSAC test benchmark and dominates the realism--diversity Pareto front among reproducible baselines. Our project page is available \href{https://seulbinhwang.github.io/flow-erd-project-page/}{here}.

Problem

• traffic simulation
• realism
• diversity
• multi-agent simulation

Innovation

• Flow Matching
• Agent-Type Awareness
• Entropy-Regularized Distillation
• Diverse Traffic Simulation
• Multi-Agent Simulation