基于虚拟安全边界的智能车辆动态换道轨迹规划  

Dynamic lane-changing trajectory planning of intelligent vehicle based on virtual safety boundary

作　　者：江浩斌[1] 洪阳珂 冯张棋 皮健 韦奇志 JIANG Haobin;HONG Yangke;FENG Zhangqi;PI Jian;WEI Qizhi(School of Automotive and Traffic Engineering,Jiangsu University,Zhenjiang,Jiangsu 212013,China)

机构地区：[1]江苏大学汽车与交通工程学院,江苏镇江212013

出　　处：《江苏大学学报（自然科学版）》2023年第6期621-629,共9页Journal of Jiangsu University：Natural Science Edition

基　　金：国家自然科学基金资助项目(51675235);泰州市重大科技成果转化项目(SCG201904)。

摘　　要：鉴于实际动态交通环境下智能车辆换道过程中可能受到周围车辆的影响,为满足其换道过程安全性和舒适性的要求,提出了一种基于虚拟安全边界的动态换道轨迹规划算法.考虑到换道结束时间和纵向位置的耦合性,采用四次和五次混合多项式换道轨迹算法分别进行纵、横向规划,确定换道时间和纵向速度范围获得候选换道轨迹.设计了车辆运动状态约束和基于车辆矩形轮廓建立的虚拟安全边界防碰撞约束,构建了车辆换道安全性、舒适性、跟车效益等代价函数,并基于安全性代价函数阈值动态调整各个代价函数的权重系数,运用fmincon函数对候选换道轨迹进行优化.考虑到周围车辆运动状态变化会对换道过程产生影响,车辆系统循环调用换道轨迹规划模块进行换道轨迹实时更新,直至车辆到达目标位置.结果表明:对于不同的换道工况,基于虚拟安全边界的智能车辆动态换道轨迹规划算法可有效避免换道过程中与周围车辆发生碰撞,保证了智能车辆换道过程的安全性.Due to the variability and complexity of the traffic environment,a dynamic lane-changing trajectory planning algorithm based on virtual safety boundary was proposed to meet the requirements of safety and comfort in lane-changing process.Considering the coupling of lane change end time and longitudinal position,the quartic and quintic mixed polynomial lane-changing trajectory algorithms were used for vertical and horizontal planning,and the lane change time and the longitudinal speed range were respectively determined to obtain the candidate lane-changing trajectory.The vehicle motion state constraint and the anti-collision constraint based on virtual safety boundary were designed,and the cost functions of vehicle lane-changing safety,comfort and car following benefit were constructed.The weight coefficients of each cost function were dynamically adjusted based on the threshold of safety cost function,and the candidate lane-changing trajectory was optimized by fmincon function.Considering the effect of the surrounding vehicle motion state change on the lane-changing process,the vehicle system circularly called the lane-changing trajectory planning module to update the lane-changing trajectory in real time until the vehicle reached the target position.The results show that for different lane-changing conditions,by the intelligent vehicle dynamic lane-changing trajectory planning algorithm based on virtual safety boundary,the collision with surrounding vehicles can be effectively avoided in the lane-changing process,and the safety of intelligent vehicle lane-changing process is ensured.

关 键 词：智能车辆 换道轨迹规划 虚拟安全边界 多项式算法 速度规划 

分 类 号：U461.99[机械工程—车辆工程]