We propose a dynamic vehicular routing algorithm with traffic prediction for improved routing performance.The primary idea of\nour algorithmis to use real-time aswell as predictive traffic information provided by a central routing controller. In order to evaluate\nthe performance, we develop a microtraffic simulator that provides road networks created from real maps, routing algorithms, and\nvehicles that travel from origins to destinations depending on traffic conditions. The performance is evaluated by newly defined\nmetric that reveals travel time distributions more accurately than a commonly used metric of mean travel time. Our simulation\nresults show that our dynamic routing algorithm with prediction outperforms both Static and Dynamic without prediction routing\nalgorithms under various traffic conditions and road configurations.We also include traffic scenarios where not all vehicles comply\nwith our dynamic routing with prediction strategy, and the results suggest that more than half the benefit of the new routing\nalgorithm is realized even when only 30% of the vehicles comply.
Loading....