A Fastest Route Planning for LBS based on Traffic Prediction

This paper proposes a new route plan on the basis of traffic prediction for finding a fastest route to a given destination in traffic network. So far, lots of traffic prediction systems were introduced to help drivers. Previous works were done mainly on providing restricted route services which depend on only cumulative traffic velocities. For this reason, we consider both real-time and cumulative traffic information together to obtain more accurate future traffic information. In location-based services, the traffic network is needed to solve certain constrains, such as turns problems and provide method for avoiding traffic congestions. To guide a fastest route service in such a complicated network, we first construct a linear dual graph from a traffic network. Then, we propose main algorithmic approaches which are developed by Kalman Filter and cumulative traffic patterns to predict a much better quality of future traffic information by combining real-time with cumulative traffic conditions. Finally, we adopt Dijkstra’s shortest path algorithm to minimize the travel time with generating a fastest cost function. Experimental results show that this approach is highly efficient in route plan than previously used ways by only cumulative approaches. This approach is supposed to proceed convenience for drivers and develop a quality of navigation service in telematics. Key-Words: Fastest Route Planning, Traffic Prediction, Kalman Filter, Cumulative Traffic Patterns, Linear Dual Graph, LBS