Real-Time Detection of Stop-and-Go Waves on Freeways Using Probe Trajectories

With the development of sensing and communication technologies, real-time information can be collected and exchanged by connected and automated vehicles via on-board sensors and V2X communication. This paper presents the development and application of a critical points (CPs) approach for real-time detection of stop-and-go waves on freeways. Using CPs related to changes in vehicle dynamics with a new filter algorithm, the stop-and-go wave speed on freeways can be estimated in real-time. Compared with previous studies on estimation of stop-and-go wave speed using the similar I-80 NGSIM dataset, there exist errors in the real-time detection of stop-and-go waves using only one probe vehicle. Thus, the different penetration rate of probe vehicles is set to observe the accuracy of detection. The result shows that the real-time detection will be more accurate, when the penetration rate exceeds 15%. The findings have a potential application for congestion prediction and control.