Strategy for Multiobjective Transit Signal Priority with Prediction of Bus Dwell Time at Stops

Abstract

Transit signal priority (TSP) is a vital aspect of the improvement of transit service. However, the effect of bus dwell time on TSP is often neglected, and few researchers have proposed a TSP strategy that predicts the bus dwell time and then implements bus priority. This study focused on the prediction of bus dwell time, which defined the bus arrival time at the intersection, and subsequently established a multiobjective TSP strategy that uses that prediction. The data extracted from the Changzhou, China, bus rapid transit (BRT) Line 2 were used to propose a hybrid model based on the autoregressive integrated moving average and the support vector machine to predict the dwell time. Next, the multiobjective TSP, with the real-time average passenger delay, the maximum queue length, and the exhaust emissions as its optimization objectives, was solved through the use of the fuzzy compromise approach. Finally, the strategy was evaluated with the microscopic simulation software VISSIM. The findings demonstrated that the prediction model produced satisfactory results, and the simulation results suggested that the proposed strategy could significantly reduce the intersection delay, the stop rate, and the exhaust emissions of BRT. Moreover, higher traffic flows corresponded to better benefits being achieved through this strategy. In addition, the delay, the queue length, and the exhaust emissions of general vehicle traffic would be effectively controlled. The findings of this study could be helpful to traffic managers in the development of appropriate signal timing strategies and the enhancement of operating efficiency and environmental quality at intersections.