A major part of delay experienced by vehicles in street networks is caused by signals. Delay estimation models have several applications ranging from analysing the performance of an intersection to finding the shortest path in traffic assignment models. However, current delay-estimation models have complicated procedures for estimating the delay for different movements of each intersection approach. This complexity results in difficulties when considering signal delay in traffic assignment models. Using the data collected from five signalised intersections in the city of Tehran, Iran, the present study suggests two delay functions for signalised intersections. The first model considers traffic volume of each approach as the effective parameter for the delay at that approach. This structure is applicable for current traffic assignment models. The second model, which is calibrated separately for each movement of an intersection, suggests considering the conflicting traffic volumes to predict the delay in left turn movements, which is appropriate for analysing the performance of each movement of an intersection. (a).
Abstract