Large Language Model-driven green wave control generation improved the average speed of arterial roads in SUMO simulation experiments.
LLMs can be used to interactively generate traffic signal control policies that improve road efficiency in simulations.
In the field of urban traffic management, optimising traffic signal control on major arterial road is crucial for reducing congestion and improving overall road efficiency. In this paper, we explore a novel approach to design and implement green wave control for urban arterials using Large Language Models (LLM), such as GPT-4. Our approach combines state-of-the-art LLM with traffic signal control policies, aiming to explore the potential of LLM for application in the field of traffic control. We design a workflow for LLM-driven green wave control generation for urban arterial road traffic signal control as an example. The experiments use SUMO simulation software to construct the traffic signal control problem of the arterial road. We verify that LLM can implement the analysis and solution process of the traffic signal control problem. The traffic signal control policy is generated interactively through natural language, which reduces the data analysis and computation pressure of traffic managers. The experimental results show that the process generates the green wave control of the arterial road that can improve the average speed of the road. The potential application of LLM in the field of traffic control is verified in this work.
Tang et al. (2024) studied Urban traffic congestion. Large Language Models (LLM) such as GPT-4 was evaluated on Average speed of the road. Large Language Model-driven green wave control generation improved the average speed of arterial roads in SUMO simulation experiments.