Network-Level Traffic Signal Coordination Using Field-Calibrated Simulation and Optimisation: A Case Study of Gandhinagar City

Abstract

Urban traffic networks with mixed vehicle types are often afflicted by poor signal coordination, leading to long delays and longer travel times. In this study, we developed and tested a traffic signal coordination framework at six closely spaced intersections in Gandhinagar, India. We used Passenger Car Unit (PCU)-based demand to model mixed traffic and treated some left-turn movements as free-flow based on field observations. We built a detailed traffic simulation using the SUMO platform and calibrated it against peak-hour data, including traffic volumes, speeds, signal timings, and travel times. We then used a Genetic Algorithm to optimise signal coordination by adjusting cycle lengths, green splits, and offsets across the network. The model matched observed travel times, with average RMSE values of 12 seconds in the morning and 11 seconds in the evening. The optimised strategy cut total network travel time by 33,943 seconds and total control delay by 7,180 seconds, improving performance by 5.42% in the morning and 2.91% in the evening. These results show that combining field-calibrated simulation with evolutionary optimisation can help improve traffic flow in complex urban networks.