UNMANNED VEHICLE LANE CHANGING DECISION-MAKING METHOD AND SYSTEM BASED ON ADVERSARIAL IMITATION LEARNING

摘要

An unmanned vehicle lane changing decision-making method based on adversarial imitation learning, and a system for implementing the method. The method comprises: first, describing an unmanned vehicle lane changing decision-making task as a partially observable Markov decision-making process; then, using an adversarial imitation learning method to perform training from examples provided by professional driving teaching to obtain an unmanned vehicle lane changing decision-making model; and in an unmanned driving process of a vehicle, obtaining a vehicle lane changing decision-making result by means of the unmanned vehicle lane changing decision-making model by taking currently obtained vehicle environmental information as an input parameter of the unmanned vehicle lane changing decision-making model. According to the method, a lane changing strategy is learned by means of adversarial imitation learning from examples provided by professional driving teaching; a task reward function is not required to be manually designed, and direct mapping from a vehicle state to a vehicle lane changing decision can be directly established, thereby effectively improving the correctness, robustness and adaptivity of the lane changing decision of the unmanned vehicle under a dynamic traffic flow condition.