Game Theoretical Decision Making Approach for a Cooperative Lane Change

Abstract

Abstract Recent advances in communications technology make it possible for vehicles to interact with each other and their environment. This allows for superior maneuvers, which open up a wide range of possibilities, which conventional vehicles without communication lack. To that end, this paper examines a decision making approach for an automated and cooperative lane change maneuver, which is based on the fundamentals of game theory. The decision making algorithm is realized with intuitive benefit functions, which are modelled similar to the semantic of human driving behavior. The used benefit functions can be classified into two sub-games: player against a single player and player against the totality of all players. By mapping four distinct driving maneuvers to their respective benefits, the problem of selecting the optimal maneuver can be solved using game theory methods. After the optimal driving maneuver has been identified, the cooperative lane change can be performed. The approach has been validated in a simulated highway scenario. Simulations have shown that a cooperative lane change does not have a significant negative effect on the traffic flow.