Cell-based traffic estimation from multiple GPS-equipped cars

We propose a cell-based algorithm used to estimate road traffic condition from multiple Global Positioning System (GPS)-equipped taxis. The GPS data is received intermittently at a central server from approximately 4,000 taxis running within Bangkok metropolitan area. The algorithm divides the Bangkok area into 6,400 square cells. Traffic condition considered here is travel time between two cells. The A* algorithm is used to predict the travel path between any pair of origin-destination. The incremental weighted update algorithm is used to update and estimate travel time between cells. The algorithm can estimate traffic condition of all areas anytime, even if there is no GPS data sent from that area. The result shows that the cell-based algorithm can achieve a good traffic estimation covers all areas and times, compared to the traditional average space-speed method, which can only give traffic condition when there are at least two data points in that area.