Saved in:
Bibliographic Details
Main Authors: Zhang, Weifan, Zhao, Xiaofeng, Bazzi, Adel, Li, Mingrui, Wei, Yifan, Sun, Dengfeng
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2605.09153
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913107437158400
author Zhang, Weifan
Zhao, Xiaofeng
Bazzi, Adel
Li, Mingrui
Wei, Yifan
Sun, Dengfeng
author_facet Zhang, Weifan
Zhao, Xiaofeng
Bazzi, Adel
Li, Mingrui
Wei, Yifan
Sun, Dengfeng
contents Closed-loop traffic simulation requires agents that are both scalable and behaviorally realistic. Recent self-play reinforcement learning approaches demonstrate strong scalability, but their equilibrium strategies fail to capture the socially aware behaviors of real human drivers. We propose a hierarchical architecture that goes beyond self-play by combining high-level multi-agent interaction reasoning with low-level continuous trajectory realization. Specifically, a Stackelberg-style Multi-Agent Reinforcement Learning (MARL) module generates interaction-aware intention commands. These commands condition a low-level continuous motion module, translating the strategic intent into physically consistent, scene-responsive control sequences. To mitigate distribution shift in closed-loop deployment, we introduce a hybrid co-training scheme combining MARL with auxiliary recovery supervision. Experiments on a SUMO-based urban network demonstrate that the proposed framework achieves superior control smoothness and safety compared to self-play and passive imitation baselines, while maintaining competitive traffic efficiency.
format Preprint
id arxiv_https___arxiv_org_abs_2605_09153
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Beyond Self-Play: Hierarchical Reasoning for Continuous Motion in Closed-Loop Traffic Simulation
Zhang, Weifan
Zhao, Xiaofeng
Bazzi, Adel
Li, Mingrui
Wei, Yifan
Sun, Dengfeng
Robotics
Artificial Intelligence
Closed-loop traffic simulation requires agents that are both scalable and behaviorally realistic. Recent self-play reinforcement learning approaches demonstrate strong scalability, but their equilibrium strategies fail to capture the socially aware behaviors of real human drivers. We propose a hierarchical architecture that goes beyond self-play by combining high-level multi-agent interaction reasoning with low-level continuous trajectory realization. Specifically, a Stackelberg-style Multi-Agent Reinforcement Learning (MARL) module generates interaction-aware intention commands. These commands condition a low-level continuous motion module, translating the strategic intent into physically consistent, scene-responsive control sequences. To mitigate distribution shift in closed-loop deployment, we introduce a hybrid co-training scheme combining MARL with auxiliary recovery supervision. Experiments on a SUMO-based urban network demonstrate that the proposed framework achieves superior control smoothness and safety compared to self-play and passive imitation baselines, while maintaining competitive traffic efficiency.
title Beyond Self-Play: Hierarchical Reasoning for Continuous Motion in Closed-Loop Traffic Simulation
topic Robotics
Artificial Intelligence
url https://arxiv.org/abs/2605.09153