An LLM-Powered Cooperative Framework for Large-Scale Multi-Vehicle Navigation

2025-10-09

AI Summary

Urban-scale multi-vehicle dynamic navigation suffers from poor scalability and difficulty in modeling nonlinear, coupled traffic dynamics. To address this, we propose CityNav—a hierarchical cooperative framework integrating global traffic coordination with local adaptive path planning. Our method introduces a large language model (LLM)-driven collaborative reasoning mechanism, coupled with an individual–shared dual-reward structure, to enable efficient multi-agent cooperation. It unifies hierarchical reinforcement learning, multi-agent decision-making, and joint training on real-world road network data. Evaluated on four real city-scale road networks—up to 1.6 million road segments and 430,000 intersections—CityNav significantly outperforms nine state-of-the-art baselines. It improves travel efficiency, alleviates congestion, and demonstrates strong scalability and practical deployability.

Technology Category

• Planning, Routing, and Scheduling: Planning with Language Models
• Multiagent Systems: Multiagent Planning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

The rise of Internet of Vehicles (IoV) technologies is transforming traffic management from isolated control to a collective, multi-vehicle process. At the heart of this shift is multi-vehicle dynamic navigation, which requires simultaneously routing large fleets under evolving traffic conditions. Existing path search algorithms and reinforcement learning methods struggle to scale to city-wide networks, often failing to capture the nonlinear, stochastic, and coupled dynamics of urban traffic. To address these challenges, we propose CityNav, a hierarchical, LLM-powered framework for large-scale multi-vehicle navigation. CityNav integrates a global traffic allocation agent, which coordinates strategic traffic flow distribution across regions, with local navigation agents that generate locally adaptive routes aligned with global directives. To enable effective cooperation, we introduce a cooperative reasoning optimization mechanism, in which agents are jointly trained with a dual-reward structure: individual rewards promote per-vehicle efficiency, while shared rewards encourage network-wide coordination and congestion reduction. Extensive experiments on four real-world road networks of varying scales (up to 1.6 million roads and 430,000 intersections) and traffic datasets demonstrate that CityNav consistently outperforms nine classical path search and RL-based baselines in city-scale travel efficiency and congestion mitigation. Our results highlight the potential of LLMs to enable scalable, adaptive, and cooperative city-wide traffic navigation, providing a foundation for intelligent, large-scale vehicle routing in complex urban environments. Our project is available at https://github.com/usail-hkust/CityNav.

Problem

• Scaling multi-vehicle navigation to city-wide networks efficiently
• Capturing nonlinear and stochastic dynamics in urban traffic
• Enabling cooperation between global and local traffic agents

Innovation

• Hierarchical LLM framework for multi-vehicle navigation
• Global traffic allocation with local adaptive routing
• Cooperative reasoning optimization with dual-reward structure