A practical route guidance approach based on historical and real-time traffic effects

Abstract

Implementing convenient traveling information service is a crucial task for deploying intelligent transportation system applications and location-based services. Traditional traveling information service systems, such as car navigation systems or web maps, only provide relatively static information which doesn't truly reflect the dynamic changes of traffic situation, and result in very limited practical use. Although there have emerged some car navigation products and other applications involving dynamic traffic information, considering the rapid change of city traffic situation, these applications still face practical difficulties for all the information received real-timely will get outdated within a few minutes, which makes the so called dynamic applications basically time-slice limited static ones. Aiming at such a problem, a short-term traffic prediction approach and a consequent real-time route guidance process are presented in this paper which integrates historical traffic based statistical reasoning, real-time traffic and events processing, with a BP neural network based analytical model, to forecast the situation and evaluate the influence of traffic during the traveling process. Then a collaboration working framework is set forward to implement dynamic route guidance, with the combination of a GIS server, a traffic forecasting server and a database management system. The traffic forecasting server, integrating with historical statistics reckoning continuously receives real-time traffic information obtained from floating vehicles, traffic events described in natural language, and achieves short-term forecasting results for the whole road networks, then fed the results back into the database management system and GIS server, so that a time-dependant optimal routing can be conducted through a dynamic least traveling time algorithm developed in this study. A prototype navigation system fulfilling the above aspects has been developed and the dynamic route choice approach demonstrated on road networks in the downtown area of Beijing city. The approach presented in this paper is argued to provide a practical solution for real-time public traveling information service and dynamic web maps.