A driver ego-centered environment representation in traffic behavioral simulation
Résumé
We present a multi-agent traffic simulation to improve the validity of traffic simulations in urban and suburban fields, with a better consideration of the driving context and driver behavior in terms of anticipation of positioning on the lanes and occupation of space. The demonstration intends to reproduce the observed behavior such as filtering between vehicles (two-wheels), prepositioning on lanes when approaching the road intersections, exceptional situations (stranded vehicle or improperly parked, etc.). The proposed approach considers that each driver is perceiving the situation in an ego-centered way and is readapting the road space by overriding the existing physical structure.