Afolayan, Blessing Itoro (University of Deusto), Ghosh, Arka (University of Deusto), sarwar, samra (University of Deusto), Masegosa Arredondo, Antonio David (University of Deusto), calderin, jenny fajardo (University of Deusto)

Optimization of Traffic Signals Using Bayesian Optimization in a Multi-Step Prompt Engineering Method

Scheduled for presentation during the Regular Session "S32b-AI-Driven Traffic Monitoring, Safety, and Anomaly Detection" (FR-EA-T32), Friday, November 21, 2025, 13:30−13:50, Southport 2

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 18, 2025

Abstract

The use of artificial intelligence (AI) in traffic management has grown significantly in the last several years. There is increasing interest in using large language models (LLMs), such as ChatGPT, to improve decision making across a range of fields, including traffic control through signal optimization. To improve average speed (AS) and average waiting time (AWT), two key traffic performance indicators (KPIs), this study investigates the application of ChatGPT in traffic signal optimization along with model-based optimization. To improve the language model's output, this study presents a multi-step prompt engineering technique (MSPE). Additionally, model-based optimization is used to enhance ChatGPT's recommendations for traffic signal timing. Tested using the RESCO benchmark and two comprehensive case studies with the SUMO traffic simulation platform, the framework showed notable improvements, and sensitivity analysis validated the results' robustness.