【作者】 王永胜； 古谚谌； 赵超； 杨皓宇； 罗禹贡；

【Author】 Wang Yongsheng;Gu Yanchen;Zhao Chao;Yang Haoyu;Luo Yugong;School of Vehicle and Mobility,Tsinghua University;Audi China;

【机构】 清华大学车辆与运载学院； 奥迪中国；

【摘要】 本文面向停车场十字路口和停车位前方两类场景,为解决自主泊车过程中自车与其他交通参与车辆交互时的决策问题,提出了自主泊车博弈决策方法。首先判断自主泊车过程中与交通参与者产生交互的类型,建立交互冲突模型;然后在冲突模型基础上,定义车辆策略集并设计博弈代价函数,基于非合作完全信息静态博弈方法对车车交互进行建模;最后通过求解纯策略纳什均衡解,获得最优通行策略。仿真结果表明,所提方法能够应对停车场中的不同车车交互场景,为自主泊车车辆提供合理决策。更多还原

【Abstract】 For the two types of scenarios,the parking lot intersection and the front of the parking space,in order to solve the decision-making problem when the self-vehicle interacts with other traffic participating vehicles in the process of autonomous parking,a game decision method is proposed for autonomous parking.Firstly,the interaction conflict model is established by judging the type of interaction with traffic participants during the autonomous parking process.Then,on the basis of the conflict model,the vehicle strategy set is defined,and a game cost function is designed,and vehicle-vehicle interaction is modelled based on non-cooperative complete information static game method.Finally,the optimal general strategy is obtained by solving the pure strategy Nash equilibrium solution.Simulation results show that the proposed method can cope with different vehicle-vehicle interaction scenarios in the parking lot,and provide reasonable decision-making for autonomous parking vehicles.更多还原