汽车工程 ›› 2022, Vol. 44 ›› Issue (6): 831-841.doi: 10.19562/j.chinasae.qcgc.2022.06.004

所属专题: 智能网联汽车技术专题-规划&控制2022年

• • 上一篇    下一篇

基于分段优化的车辆换道避障轨迹规划

唐斌1(),许占祥1,江浩斌2,蔡英凤1,胡子添1,杨铮奕1   

  1. 1.江苏大学汽车工程研究院,镇江  212013
    2.江苏大学汽车与交通工程学院,镇江  212013
  • 收稿日期:2021-11-08 修回日期:2022-01-04 出版日期:2022-06-25 发布日期:2022-06-28
  • 通讯作者: 唐斌 E-mail:tangbin@ujs.edu.cn
  • 基金资助:
    国家自然科学基金(51605199);江苏省六大人才高峰项目(2019-GDZB-084);江苏省重点研发计划(BE2019010)

Trajectory Planning of Intelligent Vehicles in Lane Change for Collision Avoidance Based on Segmented Optimization

Bin Tang1(),Zhanxiang Xu1,Haobin Jiang2,Yingfeng Cai1,Zitian Hu1,Zhengyi Yang1   

  1. 1.Institute of Automotive Engineering,Jiangsu University,Zhenjiang  212013
    2.School of Automotive and Traffic Engineering,Jiangsu University,Zhenjiang  212013
  • Received:2021-11-08 Revised:2022-01-04 Online:2022-06-25 Published:2022-06-28
  • Contact: Bin Tang E-mail:tangbin@ujs.edu.cn

摘要:

为提高复杂道路场景下智能车辆换道避障的安全性和舒适性,提出了一种基于分段优化的智能车辆换道避障轨迹规划方法。首先考虑多种换道可能,根据本车状态和多个采样点生成基于五次多项式的候选y-x曲线簇和x-t曲线簇;其次基于指数函数设计了一种障碍风险评价函数,并结合轨迹平顺性、利他性和行驶效率等构建了综合评价体系,选取出最优参考轨迹,为智能车辆换道避障提供方向和速度的参考,以防止轨迹优化时陷入局部最优;为适应障碍物运动状态时变的特点,以参考轨迹为引导,构建分段五次多项式y-x曲线和x-t曲线,并考虑与障碍车辆的碰撞风险建立了优化目标函数,将轨迹优化问题转换成带约束的非线性规划问题,通过外点法求出最优轨迹。最后基于MATLAB平台进行了仿真验证,结果表明,所提出的轨迹规划方法在满足换道平顺、舒适要求的基础上提高了车辆的环境适应性和避障调整能力。

关键词: 智能驾驶, 换道避障, 轨迹规划, 分段优化, 非线性规划

Abstract:

In order to enhance the safety and comfort of intelligent vehicle during lane change for collision avoidance under complex road scenes, a trajectory planning method of intelligent vehicle lane change for collision avoidance is proposed based on segmented optimization. Firstly, according to the state of ego vehicle and multiple sampling points, candidate y-x curve clusters and x-t curve clusters are generated based on quintic polynomials with consideration of various possibilities of lane change. Then, by designing an obstacle risk evaluation function based on exponential function, and combined with the smoothness, altruism and driving efficiency of the trajectory, a comprehensive evaluation system is constructed to select the optimal reference trajectory, providing the direction and speed references for intelligent vehicle lane change for obstacle avoidance and preventing the solution from falling into local optimum in trajectory optimization. For suiting to the time-varying features of the obstacle movement state, the piecewise quintic polynomial y-x curve and x-t curve are constructed with the reference trajectory as a guide, and the optimization objective function is also established with consideration of the risk of collision with obstacle vehicles, converting the trajectory optimization problem into a constrained nonlinear programming one to find out the optimal trajectory by exterior point method. Finally, the simulation verification is performed on MATLAB platform with a result showing that the trajectory planning method proposed enhance the environmental adaptability and obstacle avoidance adjustment ability of vehicles, while meeting the requirements of lane-change smoothness and ride comfort.

Key words: intelligent driving, lane change for collision avoidance, trajectory planning, segmented optimization, nonlinear programming