With the emergence of the Connected Vehicle technology, estimating the queue distance using probe trajectory data has recently received considerable attention. Different from data collected at loop detectors, probe trajectory data can provide a lower bound on the queue distance even if the market penetration rate is low. In this paper, we develop an event-based method that uses both probe trajectory and signal timing data to estimate the queue length. The estimation accuracy under different market penetration rates is then examined. We also develop a data fusion method that combines probe trajectory data and loop detector data. We then investigate when the data fusion method is expected to work well. Case studies are conducted using microscopic simulation data, and some observations are made.
↧