An adaptive routing system for location-aware mobile devices on the road network

As congestion problems become a greater concern and negatively impact society, solutions which alleviate them are needed to improve the performance of the transportation system. Routing systems which take into account the travel-time experienced by the driver have been largely unexplored in the domain of adaptive routing. In this article, we present a system which enables users of smartphones to obtain directions generated using an algorithm which provides an optimal routing policy for reliable on-time arrival; that is, directions which seek to maximize the probability of arriving to the destination within a given time budget, rather than to minimize the travel time based on posted speed limits. Our work leverages the geolocation capabilities of smartphones to provide optimal routing directions along the route dependent on the realized (experienced) travel time. The adaptive routing scheme we implement allows for significant power savings and improved driver safety compared to classical routing algorithms; special attention is paid to minimizing driver distraction by emphasizing aural and graphical components over textual elements during route guidance. Finally, we illustrate system performance and design choices on synthetic examples and real traffic data from the Mobile Millennium system in San Francisco.