System Performance Optimization at Urban Rail-Highway Grade Crossings Using Online Adaptive Priority Strategy

The transit priority systems at grade crossings have been well studied over years. In urban areas, these systems often cause too much traffic delays due to the lack of real-time coordination between trains and the other traffic. The purpose of this study is to develop an online adaptive priority strategy to optimize traffic operation at successive grade crossings and take into account the impact on cross-street traffic. A mathematical model is proposed to minimize intersection delays for light rail transit (LRT) meanwhile keeping the impact on cross-street traffic as little as possible. To achieve this goal, we use overall passengers’ delay as the performance index, which is defined as the weighted online delay of the light rail vehicle (LRV) along successive intersections and the cross-street traffic delay based on statistics of historical data. The optimization of the overall delay is enabled by real-time arrival time prediction with GPS-instrumented vehicles. Green bands for both approaches of LRVs are formed to cover the online predicted arrival time at the first intersection, being wide enough to accommodate prediction errors while not too much to reduce impact on cross streets. To validate the proposed strategy, both numerical analysis and simulation test are conducted with trolley operation along a section of C Street in San Diego downtown area. With sensitivity analysis, we get further insight into the impact of user-defined parameters and prediction errors on the system performance. Compared with current signal timings, the obtained optimal timings demonstrate promising results.