自动驾驶参考轨迹分段多项式平滑方法  

Smoothing method with piecewise polynomial fitting for reference trajectory of autonomous driving

作　　者：王猛[1,2] 崔鹏 陈珏璇 WANG Meng;CUI Peng;CHEN Juexuan(School of Artificial Intelligence and Automation,Huazhong University of Science and Technology,Wuhan 430074,China;K+Lab Laboratory,Wuhan Kotei Informatics Co.,Ltd.,Wuhan 430073,China;Department of Autonomous Driving,Wuhan Kotei Technology Corporation,Wuhan 430200,China)

机构地区：[1]华中科技大学人工智能与自动化学院,湖北武汉430074 [2]武汉光庭信息技术股份有限公司K+Lab实验室,湖北武汉430073 [3]武汉光庭科技有限公司智能驾驶部,湖北武汉430200

出　　处：《现代电子技术》2021年第15期154-158,共5页Modern Electronics Technique

基　　金：2019年度湖北省博士后创新岗位项目:无人驾驶决策与规划方法研究。

摘　　要：自动驾驶系统从地图数据获取一系列轨迹点作为寻迹行驶的参考轨迹,由于地图数据误差的存在,导致输入的参考轨迹不够平滑,影响到后续决策控制模块的性能,最终造成自动驾驶车辆行驶的舒适性差。针对这一问题,提出一种分段多项式拟合的参考轨迹平滑处理方法,对原始的参考轨迹点做分段处理,利用带约束条件的最小二乘法求解分段轨迹点的多项式拟合曲线,最后将分段拟合曲线等间距采样作为平滑处理后的参考轨迹。该方法能够保证分段曲线高阶连续的同时,在分断点处具有一阶连续。实验结果表明,采用分段多项式拟合方法获得的拟合曲线连续光滑,处理后的参考轨迹具有较好的平滑性,能够进一步消除地图数据误差造成的影响,提升自动驾驶的舒适性。Usually,an autonomous driving system obtains a series of trajectory points from map data as the reference tracks for tracing driving.Due to the data error of the map,the input reference trajectory is not so smooth,which will affect the performance of subsequent decision control module and finally lead to poor driving comfort of autonomous vehicle.In view of this,a reference trajectory smoothing method with piecewise polynomial fitting is proposed.The original reference trajectory points are divided into several segments.The polynomial fitting curve of the segmented trajectory points is solved by the least square method with constraint conditions.The piecewise fitting curve samples with equal distance are taken as the reference trajectory after smoothing processing.The experiment results show that the proposed method can guarantee the high⁃order continuity on each segmented curve and the first⁃order continuity at each segmented point,the fitting curve obtained by the piecewise polynomial fitting method is continuous and smooth,and the reference trajectory after processing is of smoothness.Therefore,it can further eliminate the influence caused by data error of the map and improve the comfort level of autonomous driving.

关 键 词：平滑处理 分段多项式拟合 自动驾驶 参考轨迹 分段处理 约束条件 一阶连续 

分 类 号：TN911.1-34[电子电信—通信与信息系统]