Data-driven algorithms for real-time adaptive tuning of offsets in coordinated traffic signal systems

A real-time adaptive control algorithm for tuning traffic signal offsets in a coordinated traffic signal system is presented. The algorithm described here is used in the Adaptive Control Software-Lite (ACS-Lite) adaptive control system. The algorithm uses a statistical average flow profile of traffic on the coordinated approaches to an intersection to assess when vehicles are arriving during the signal cycle. Alternative offset adjustments are evaluated by calculating how much of this flow profile is being captured by the green phase serving each coordinated approach. The algorithm considers the impact of the offset adjustment on traffic at the local intersection as well as on traffic at adjacent intersections that are also under ACS-Lite control. Simulation tests have quantitatively shown that this tuning approach can improve arterial progression performance relative to the quality of the initial baseline fixed offsets used in a traffic pattern. Several issues and areas for improvement of the algorithm are also identified and discussed.