Iterative Tuning Strategy for Setting Two-Way Offset by Using Histogram of Traffic Flows in Traffic Signal Control

This paper introduces Iterative Tuning (IT) strategy for offset to coordinate neighbouring junctions along arterial road in urban network. Recently IT strategy for phase split is proposed with anticipation of traffic demands since traffic flows are repetitive with certain variations on working days. The offset is designed to initialize the stating time of junctions such that "green wave" are generated to allow more possible vehicles driving through junctions without stopping by red signals. Traffic flow histograms collected at upstream links are exploited here to estimate downstream queued vehicles. The objective is to minimize the number of queued vehicles for two-way links between neighbouring junctions. It is a decentralized strategy and converges to steady states via iterative actions. IT strategies for phase split and offset are integrated iteratively to generate better traffic signal schedules. Real traffic conditions in Singapore are emulated with real traffic variations on working days. Simulations results show that delay time and number of stops under IT strategies for phase split and offset outperform IT strategy for phase split only and fixed-time strategy.