• CN:11-2187/TH
  • ISSN:0577-6686

机械工程学报 ›› 2022, Vol. 58 ›› Issue (14): 276-287.doi: 10.3901/JME.2022.14.276

• 运载工程 • 上一篇    下一篇

扫码分享

基于采样区域优化的智能车辆轨迹规划方法

张利鹏1,2, 苏泰1,2, 严勇1,2   

  1. 1. 燕山大学车辆与能源学院 秦皇岛 066004;
    2. 河北省特种运载装备重点实验室 秦皇岛 066004
  • 收稿日期:2021-08-03 修回日期:2021-12-08 出版日期:2022-07-20 发布日期:2022-09-07
  • 通讯作者: 张利鹏(通信作者),男,1979年出生,博士,教授,博士研究生导师。主要研究方向为智能车辆动力学与控制、新能源汽车复合传动、驾驶员认知与人机共驾。E-mail:evzlp@ysu.edu
  • 作者简介:苏泰,男,1996年出生,硕士研究生。研究方向为智能车辆动力学与控制;严勇,男,1996年出生,硕士研究生。研究方向为智能车辆动力学与控制。
  • 基金资助:
    国家自然科学基金资助项目(51775478)。

Trajectory Planning Method of Intelligent Vehicle Based on Sampling Area Optimization

ZHANG Lipeng1,2, SU Tai1,2, YAN Yong1,2   

  1. 1. School of Vehicle and Energy, Yanshan University, Qinhuangdao 066004;
    2. Hebei Key Laboratory of Special Delivery Equipment, Qinhuangdao 066004
  • Received:2021-08-03 Revised:2021-12-08 Online:2022-07-20 Published:2022-09-07

摘要: 针对现有智能车辆轨迹规划方法在结构化道路中、高速场景中因均匀采样产生的计算时间浪费问题,提出一种基于采样区域优化的轨迹规划方法。该方法综合考虑道路环境信息和障碍物信息,将采样区域划分为基础代价区和障碍代价区,分别计算两个区域内采样点的代价值并依据该代价值对采样点进行筛选,高代价点将被忽略,低代价点将用于计算最优轨迹,从而避免均匀采样造成的计算量浪费。为进一步缩短规划时间并提高轨迹选择合理性,将候选轨迹按轨迹代价值进行排序,依次对候选轨迹进行碰撞检测,将不满足检测的轨迹剔除并将第一条通过检测的轨迹选为最优轨迹。为检验算法的可靠性,通过构建仿真道路及设计多个场景,对规划算法进行仿真验证。研究结果表明,所提出方法在有效降低单步轨迹规划时间的同时,保证了最优轨迹的安全性、合理性和可靠性。

关键词: 智能汽车, 轨迹规划, 结构化道路, 采样区域, 多场景仿真

Abstract: In order to solve the problem of computational time waste of existing intelligent vehicle uniform sampling trajectory planning methods in structured road, medium and high speed scene, a trajectory planning method based on sampling area optimization was proposed. The road environment information and obstacle information are considered, the sampling area is divided into the base cost area and obstacle cost area, the cost value of the sampling point in the two regions is calculated and the sampling point according to the value is screened. The high cost points will be ignored, and the low cost points will be used to calculate the optimal trajectory, so as to avoid the calculation waste caused by the uniform sampling. In order to further reduce the planning time and improve the rationality of trajectory selection, the candidate trajectory is sorted according to the cost value, and then carried out collision detection on the trajectory in turn. The trajectory that did not meet the detection were removed and the first trajectory that passed the detection was selected as the optimal trajectory. To test the reliability of the algorithm, the simulation path is constructed and multiple scenes are designed to simulate the planning algorithm. The simulation results show that the proposed method can effectively reduce the single-step planning time and ensure the security, rationality and reliability of the optimal trajectory.

Key words: intelligent vehicle, trajectory planning, structured road, sampling area, multi-scene simulation

中图分类号: