Evaluating Effectiveness of Volume- and Trajectory-Based Methods to Enhance Traffic Signal Retiming and Coordination

Abstract

Optimizing traffic signals often involves a trade-off between complexity, data availability, and real-world feasibility. Existing simulation-based studies lack validation using field-collected data, limiting their real-world applicability. Additionally, few studies evaluate cost-effective hybrid optimization strategies that leverage varying levels of data availability in areas with data constraints. Signal timing plans based on static volume data are outdated and often fail to capture real-time traffic dynamics, numerous agencies rely on these plans due to resource restrictions. Scalable, cost-effective solutions are needed to help bridge the gap between theory and practice. Trajectory data offers more accuracy, but its utilization is limited by cost and infrastructure needs. This study explores and compares different strategies for improving signal timing on eight intersections along Avenue B, a congested arterial corridor in Yuma, Arizona. The strategies include: a volume-based method, a trajectory-based method using real-time probe vehicle data, and a hybrid method that integrates both. Five scenarios were developed and evaluated using PTV VISSIM, with performance measured at both the corridor and intersection levels, using metrics such as average travel time, delay, number of stops, stop delay, and percentage of arrivals on green. The baseline simulation model was calibrated using field data to ensure realistic outputs.The assessment of several scenarios indicated that those including trajectory data surpassed the others. The combination of volume-based splits with trajectory-informed offsets (VSTO) exhibited the most balanced and effective performance, achieving a decrease of 13.17% in average corridor travel time and 53.99% reduction in stop delay. It enhanced corridor-wide progression while also reducing intersection-level delay by 25.36%. The trajectory-based offset coordination method (TOC) also improved signal coordination and vehicle progression with a travel time and stop delay reduction of 17.17% and 46.85%, respectively. However, these enhancements came at the expense of side street performance, with intersections experiencing an average control delay increase of 33.65%. The volume-based optimization with speed-informed offsets (VSSO) effectively reduced localized delay at individual intersections yet did not provide consistent improvements on the corridor-level. Intersection delay declined by 23.81%, while average travel times and stop delay decreased by 1.3% and 9.54%, respectively. While not significant, VSSO slightly improves on platoon progression whereas the scenario that optimizes splits only using volumes (VSO) does not improve on the corridor level, still, VSSO reduces intersection delay by the same amount as the former. Overall, the hybrid method VSTO which integrates trajectories with volumes, proved to be the most adaptable, offering both improved progression and localized delay reduction, highlighting the value of integrating even limited real-time data into traditional signal timing workflows. This study demonstrates that effective signal optimization does not always require complex infrastructure, but rather effectively utilizing even limited availability of datasets and tools can deliver scalable, context-sensitive solutions for cities like Yuma. The findings aim to support transportation agencies in adopting realistic strategies that bridge the gap between theory and implementation.