Autoregressive integrated moving average modeling for short-term arterial travel time prediction

Travel time information is a good operational measure of the effectiveness of transportation systems and can be used to detect incidents and quantify congestion. The ability to accurately predict freeway and arterial travel times in transportation networks is a critical component for many Intelligent Transportation Systems (ITS) applications (e.g., advanced traffic management systems, in-vehicle route guidance systems). This paper focuses on arterial travel time prediction by studying the travel time data, modeling and diagnostic checking so that short-term travel time can be predicted with reasonable accuracy. A 3.7-mile corridor on Minnesota State Highway 194 is chosen as our test site. The Global Positioning System (GPS) test vehicle method is used in our data collection. The time series analysis techniques are then used in our travel time modeling, in particular, we focus on the autoregressive integrated moving average (ARIMA) model due to the non-stationary property of the observed data. The models established for the corridor are verified via both the residual analysis and portmanteau lack-of-fit test. Finally, based on the models developed we present our prediction results. The method proposed in this paper can be easily modified and applied to short-term arterial travel time prediction for other urban areas.