Probelet: Continuous Link Sequence which Describes Collective Intelligence in Route Selection Behavior Analyzed from Probe Car Trajectories

摘要

In route search calculation, conventional routing algorithms consider network information such as road length, travel time, or road shape for each link. On the other hand, when drivers make route selection on familiar road network with their own intelligence, they don't handle network information separately and numerically. In their way of thinking, they deal with network information by some link sequences as units of optimized short trips and obtain comfortable route easy to drive. However, it is difficult to generate those link sequences automatically only from map data. This is one of the reasons why conventional routing algorithms sometimes give route guidance which is different from drivers' preference. In order to solve this problem and provide more suitable route guidance, we developed probe data analysis method which generates continuous link sequence based on collective intelligence in drivers' route selection behavior.