Real-time Travel Time Prediction on Urban Arterial Network

Travel time is one of the most desired operational variables serving as a key measure of effectiveness for evaluating the system performance of freeways and urban arterials. With accurate travel time information, decision makers, road users, and traffic engineers can make informed decisions. However, retrieving network-level travel time information has several challenges, such as traffic data collection and travel time estimation and prediction. This research addresses these challenges by developing innovative methodologies and computer applications. First, the authors developed a two-step empirical approach to effectively estimating link journey speeds using merely advance single-loop detector outputs. Second, an α–β filter is adopted to dynamically predict and smooth real-time spot speeds resulted from loop measurements. In addition to travel time estimation and prediction, a dynamic shortest path algorithm is also developed to determine the shortest travel time route based on real-time traffic condition. Furthermore, the developed algorithms are implemented in a web-based system called Real-time Analysis and Decision-making for ARterial Networks (RADAR Net). For real-time operations of RADAR Net, sensor and signal control databases are carefully designed to ensure fast query performance in a growing network-wide traffic dataset. Also, the data visualization and statistical analysis modules are added to RADAR Net to facilitate user applications. Currently, the RADAR Net system is part of the Digital Roadway Interactive Visualization and Evaluation Network (DRIVE Net) (www.uwdrive.net) developed by the STAR Lab of the University of Washington. RADAR Net is currently being operated in real-time for arterial traveler information, performance evaluation, and analysis.