A Traffic Signal Optimization Strategy Considering Both Vehicular and Pedestrian Flows

Current signal control strategies tend to ignore the pedestrian delays that may be imposed by reducing traffic delays. Such an objective is reasonable for motorways and rural roads where vehicular traffic is dominant over pedestrian traffic. However, it is not the case in metropolitan cities with large volume of pedestrian demands. This paper developed a traffic signal optimization strategy that considers both vehicular and pedestrian flows. The objective of the proposed model is to minimize the weighted vehicular and pedestrian delays. The deterministic queuing model is used to calculate vehicular traffic delay and pedestrian delay on sidewalk. Pedestrian delay on crosswalk is calculated based on an empirical pedestrian speed model, which considers interactions of pedestrian platoons and their impacts on average walking speed. The proposed model is first implemented at a Japanese intersection. MATLAB is utilized to solve the optimization problem and to output a set of MOEs. The results show that the proposed model improved average person delay (APRD) by 10% without changing the existing cycle length. Moreover, the model can optimize the cycle length and further improve APRD by as much as 44%. In order to demonstrate the applicability of the proposed model for general cases, this paper also conducted sensitivity analysis. According to the results, the proposed model is most significant and necessary for two circumstances: (1) metropolitan areas with high pedestrian demands and (2) major urban arterials with high pedestrian demands crossing the major streets.