A Multistep Corridor Travel-Time Prediction Method Using Presence-Type Vehicle Detector Data

Abstract

Travel time is an important measure for transportation system performance evaluation. In particular, it is an essential input to the advanced traveler information systems and route guidance applications that requires reliable traffic information in real time. Therefore, numerous studies have been conducted to predict segment and corridor travel times on the basis of data from loop detectors. Focusing on freeway corridor travel-time prediction that incorporates the effect of traffic propagation, this article presents a dynamic prediction model using the traffic measurements from presence-type vehicle detector data. The authors designed a multistep predictor in the form of a seasonal autoregressive integrated moving average model with an embedded adaptive Kalman filter that allows for continued adjustment of prediction results when new data become available in real time. The authors evaluated the corridor travel-time prediction method against 2 alternative methods using the same procedure as proposed, but one does not consider the effects of traffic propagation along the corridor and the other one does not perform the multiple-step traffic prediction. Along with the performance metrics, this study demonstrates the advantages of the model in capturing the dynamics of traffic and reveals the significant effects of traffic propagation on corridor travel-time prediction for producing more accurate prediction results. The authors also developed an online implementation framework.