Cooperative autonomous driving for traffic congestion avoidance through vehicle-to-vehicle communications

In this paper, we consider the problem of cooperative driving for connected autonomous vehicles to reduce traffic congestion. The connected autonomous vehicles are able to exchange information as well as driving intentions with surrounding vehicles through vehicle-to-vehicle (V2V) communications. We define three driving conditions for a given vehicle, namely, maximum, deadlock, and freerun, in which a deadlock condition implies that the vehicle is in a congestion condition that cannot be resolved alone. For the purpose of improving overall traffic flow, we design a strategy called Altruistic Cooperative Driving (ACD), in which a connected autonomous vehicle can automatically identify a deadlock condition and form a multi-vehicle coordination group to resolve the deadlock through a cooperative maneuver procedure. The proposed ACD is evaluated in our traffic simulator in terms of speed efficiency and communication traffic between the connected autonomous vehicles. The results show that ACD can improve speed efficiency and successfully resolve traffic congestion caused by deadlock conditions.