Arterial performance measurement using probe vehicle trajectory data

Along with the development of traffic detection technologies, the probe vehicle trajectory data comes available. Most of the existing studies about trajectory data focus on traffic flow shockwave detection, whose results indicate the great potential of this new data source over traditional data sources. However, the major limitation which makes these methods infeasible in current practice is that these methods require the trajectories of the entire vehicle population. It is still not clear about the best way of utilizing sampled trajectory data to obtain real time performance measures for arterials.;An innovative approach for arterial performance measurement only using probe vehicle trajectory data is presented. As the keystone of the entire approach, Critical Points (CPs) on vehicle trajectories are defined as the points representing the changing vehicle dynamics. As the first step. CPs are extracted from raw trajectories based on this definition. Since real-time signal parameters are difficult to collect for all the signalized intersections, a signal timing detection method is developed. A shockwave based model is then presented to estimate the queue length on a cycle-by-cycle basis. Based on the queue length estimates of all the intersections, the corridor travel time is estimated. The performance of this approach is tested both by simulation and real data under different flow and signal timing scenarios. Sensitivity analysis about probe penetration rate and data detection error is also conducted. The results indicate that this trajectory based approach is very promising.