Cooperative Maneuver Planning for Highway Traffic Scenarios based on Monte-Carlo Tree Search

Future automated vehicles must conquer the challenge of anticipating the intentions of other traffic participants in order to ensure the safety and efficiency of road traffic. Therefore, the research field of cooperative driving deals with maneuver planning algorithms that enable vehicles to behave cooperatively in interactive traffic scenarios. The published work in this field lacks comparability due to the use of various differing approaches regarding the description of the guidance layer of the driving task and the optimization method. Thus, we present a cooperative planning algorithm based on Monte Carlo Tree Search (MCTS), which optimizes arbitrary sequences of highway traffic scenarios with respect to a given cost function. Although it might not be applicable in real traffic due to real-time capability, communication protocols, mixed traffic or oversteering by the driver, it shall serve as a comparison for other algorithms that fulfil these requirements at the expense of solution quality. Exemplary simulations show that the algorithm solves challenging traffic scenarios with multiple vehicles successfully.