# TrafficClaw: A Generalizable LLM Agent in the Unified Physical Environment for Urban Traffic Control

arXiv: https://arxiv.org/abs/2604.17456 · HTML: https://mlsanity.com/paper/2604.17456/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Multimodal and embodied agents, one of 16 directions in LLM agents and tool use.
- Builds on 20 of its 46 references that are in the map.
- First posted on arXiv 19 Apr 2026.

## Placed in

- LLM agents and tool use / [Multimodal and embodied agents](https://mlsanity.com/agents/multimodal-and-embodied-agents/)

## Builds on (most cited first, 20 of 20)

- [Playing Atari with Deep Reinforcement Learning](https://mlsanity.com/paper/1312.5602/) (2013)
- [HybridFlow: A Flexible and Efficient RLHF Framework](https://mlsanity.com/paper/2409.19256/) (2024)
- [WebThinker: Empowering Large Reasoning Models with Deep Research Capability](https://mlsanity.com/paper/2504.21776/) (2025)
- [JudgeBench: A Benchmark for Evaluating LLM-based Judges](https://mlsanity.com/paper/2410.12784/) (2024)
- [AIDE: AI-Driven Exploration in the Space of Code](https://mlsanity.com/paper/2502.13138/) (2025)
- [CoLight: Learning Network-level Cooperation for Traffic Signal Control](https://mlsanity.com/paper/1905.05717/) (2019)
- [AgentGen: Enhancing Planning Abilities for Large Language Model based Agent via Environment and Task Generation](https://mlsanity.com/paper/2408.00764/) (2024)
- [A Survey on Traffic Signal Control Methods](https://mlsanity.com/paper/1904.08117/) (2019)
- [LLMLight: Large Language Models as Traffic Signal Control Agents](https://mlsanity.com/paper/2312.16044/) (2023)
- [TrafficGPT: Viewing, Processing and Interacting with Traffic Foundation Models](https://mlsanity.com/paper/2309.06719/) (2023)
- [VillagerAgent: A Graph-Based Multi-Agent Framework for Coordinating Complex Task Dependencies in Minecraft](https://mlsanity.com/paper/2406.05720/) (2024)
- [CitySim: Modeling Urban Behaviors and City Dynamics with Large-Scale LLM-Driven Agent Simulation](https://mlsanity.com/paper/2506.21805/) (2025)
- [Air Quality Prediction with Physics-Guided Dual Neural ODEs in Open Systems](https://mlsanity.com/paper/2410.19892/) (2024)
- [AgentMove: A Large Language Model based Agentic Framework for Zero-shot Next Location Prediction](https://mlsanity.com/paper/2408.13986/) (2024)
- [CoRide: Joint Order Dispatching and Fleet Management for Multi-Scale Ride-Hailing Platforms](https://mlsanity.com/paper/1905.11353/) (2019)
- [Expert Level control of Ramp Metering based on Multi-task Deep Reinforcement Learning](https://mlsanity.com/paper/1701.08832/) (2017)
- [BrowseMaster: Towards Scalable Web Browsing via Tool-Augmented Programmatic Agent Pair](https://mlsanity.com/paper/2508.09129/) (2025)
- [MM-Path: Multi-modal, Multi-granularity Path Representation Learning -- Extended Version](https://mlsanity.com/paper/2411.18428/) (2024)
- [ARROW: An Adaptive Rollout and Routing Method for Global Weather Forecasting](https://mlsanity.com/paper/2510.09734/) (2025)
- [Large Language Model Powered Intelligent Urban Agents: Concepts, Capabilities, and Applications](https://mlsanity.com/paper/2507.00914/) (2025)

## Abstract (arXiv)

Large language model (LLM) agents have shown strong capabilities in long-horizon reasoning, tool use, and decision-making in digital environments, yet extending them to physically grounded systems remains challenging. Unlike web, code, or game environments, where objectives are often weakly coupled, physical systems evolve through tightly coupled dynamics in which local interventions propagate across interacting subsystems over time. Urban traffic control exemplifies this challenge, as traffic signals, freeways, public transit, and taxi systems continuously interact through shared spatial infrastructure and temporal mobility demand. Existing optimization, reinforcement learning (RL), and LLM-based approaches are largely designed for isolated subsystems, limiting coordinated reasoning and system-level optimization. We propose TrafficClaw, a LLM-based generalizable traffic control agent for physical urban systems. TrafficClaw operates within a unified traffic environment that exposes coupled urban dynamics and feedback, performs executable spatiotemporal reasoning with persistent memory for long-horizon adaptation, and leverages multi-stage agentic RL for coordinated system-level optimization. Experiments across three metropolitan regions and six traffic-control tasks demonstrate strong generalization, robustness, and cross-subsystem coordination. Our project is available at https://github.com/usail-hkust/TrafficClaw.