A developmental framework towards dynamic incident rerouting using vehicle-to-vehicle communication and multi-agent systems

Abstract

Non-recurring roadway traffic incidents are a part of everyday life that impedes normal traffic flow. One of the major challenges is to reroute traffic around incidents in order to maximize throughput. This often comes up when an incident occurs on a freeway and alternate routes need to be utilized on city and residential streets. Today's navigational devices can be used to find alternate routes based on minimum-distance or minimum-duration algorithms. However, these navigational devices tend to use the same greedy routing algorithms. This often results in moving the congestion from the incident location to another location in the network and perhaps causing further incidents due to the increased traffic that the roads are not designed to handle. In order to tackle this problem, we present an architectural framework that borrows concepts from multi agent systems and recent activity in vehicle-to-vehicle communication to model and implement behaviors that maximize overall system throughput.