基于局部博弈的单交叉口路径规划研究  

Single Intersection Path Planning Based on Local Game

作　　者：姜楠 赵清海[1,3] 徐冲 杜思雨 JIANG Nan;ZHAO Qinghai;XU Chong;DU Siyu(School of Mechanical and Electrical Engineering,Qingdao University,Qingdao 266071,China;School of Automation,Qingdao University,Qingdao 266071,China;Electric vehicle intelligent power integration technology national local joint engineering research center,Qingdao 266071,China)

机构地区：[1]青岛大学机电工程学院,山东青岛266071 [2]青岛大学自动化学院,山东青岛266071 [3]电动汽车智能化动力集成技术国家地方联合工程研究中心,山东青岛266071

出　　处：《复杂系统与复杂性科学》2024年第4期126-133,共8页Complex Systems and Complexity Science

基　　金：国家自然科学基金(52175236)。

摘　　要：针对智能车辆在单交叉口环境下存在因冲突风险引起的碰撞率高和通行效率低的问题,提出一种基于局部博弈算法的路径规划算法。通过建立路口冲突风险模型分析存在的冲突风险,对车辆与行人的冲突提出消解方法;引入局部博弈理论,建立收益函数评价可行决策,求出在约束条件下纯策略纳什均衡的解,并对车辆速度进行规划,最终实现智能车辆的路径规划。仿真结果表明,提出的路径规划算法在不同车流量下相比于演化博弈和合作博弈算法交互成功率提升9.32%,通行效率平均提升33%,能有效缓解交通压力。Aiming at the problems of high collision rate and low traffic efficiency caused by conflict risk of intelligent vehicles in single intersection environment,a path planning algorithm based on local game algorithm is proposed.By establishing the intersection conflict risk model,the conflict risk is analyzed,and the conflict resolution method between vehicles and pedestrians is proposed.The local game theory is introduced to establish the profit function to evaluate the feasible decision,and the solution of the pure strategy Nash equilibrium under the constraint condition is obtained.The vehicle speed is planned,and the path planning of the intelligent vehicle is finally realized.The simulation results show that the proposed path planning algorithm improves the interaction success rate by 9.32%compared with the evolutionary game and cooperative game algorithm under different traffic flows,and the traffic efficiency increases by 33%on average,which can effectively alleviate the traffic pressure.

关 键 词：智能车辆 人车冲突 局部博弈 纳什均衡 路径规划 

分 类 号：TB391[一般工业技术—材料科学与工程] U491.23[交通运输工程—交通运输规划与管理]