T IME - D EPENDENT A* R OUTE P LANNING S YSTEM -- A N O NLINE A PPROACH FOR T AIWAN F REEWAY S YSTEM

Traditional route guidance programs often rely on spatial distance as the decision variable for their route calculations and they do not consider the actual traffic conditions. This paper used the real - time travel time of vehicles passing through road segm ents as the cost in route selection, to develop a time - dependent A* Algorithm for routing. Real - time data from a travel - time database was used as the data source for journey calculation. The travel - time database incorporated from Vehicle Detectors (VD) and Electronic Toll Collection (ETC) systems. Aside from the forward - search calculation method, backward - search capability was also added to cater for the needs of different users. The user would enter the starting point, destination and the expected departur e or arrival time. The system would then combine the road segment travel data from the travel - time database for calculation, recommend the best route and provide the total time for the journey. Finally this research showed that the A* Algorithm could effec tively and rapidly narrow the direction of search and produce a satisfactory recommended route.