在线校正的无人车五次样条实时路径规划方法  

Real Time Path Planning Method of Autonomous Vehicles Based on Quintic Spline with Online Correction

作　　者：张立雄 郭艳 李宁 宋晓祥 薛端 ZHANG Lixiong;GUO Yan;LI Ning;SONG Xiaoxiang;XUE Duan(Army Engineering University,Nanjing 210001,China)

机构地区：[1]陆军工程大学,南京210001

出　　处：《火力与指挥控制》2022年第9期137-142,152,共7页Fire Control & Command Control

基　　金：国家自然科学基金(61871400);江苏省自然科学基金资助项目(BK20171401)。

摘　　要：针对时变行驶环境下对路径规划乘坐舒适性、稳定性、实时性和连续性的要求,提出了一种在线校正的无人车五次样条实时路径规划方法。基于五次样条曲线的路径生成模型,考虑速度、加速度和横摆角速度等约束,以平均曲率最小和距离最短为目标设计了目标函数,利用序列二次规划算法(SQP)优化参数。设计了基于映射查找表的动态迭代路径规划方法,通过搜索最佳分割点和在线校正参数快速规划路径。仿真结果表明,所提方法具有较好的性能。Aiming at the demand of ride comfort,stability,real-time and continuity of path planning in time-varying driving environment,a real time path planning method of autonomous vehicles based on quintic spline with online correctionan is proposed.Based on the path generation model of quintic spline curve,the objective function is designed to minimize the average curvature and the shortest distance when the constraints of velocity,acceleration and yaw rate are considered.The parameters are optimized based on the sequential quadratic programming algorithm(SQP).A dynamic iterative path planning method based on mapping look-up table is designed,which can quickly plan the path by searching the best segmentation points and online correction parameters.Simulation results showed that the proposed method has good performance.

关 键 词：无人车 五次样条 路径规划 序列二次规划 在线校正 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]