测绘通报 ›› 2021, Vol. 0 ›› Issue (2): 44-48.doi: 10.13474/j.cnki.11-2246.2021.0041

• 学术研究 • 上一篇    下一篇

面向LiDAR/Radar松组合的迭代加权IEKF-BP组合算法精度分析

宋宝1, 柯福阳2, 赵兴旺1   

  1. 1. 安徽理工大学空间信息与测绘工程学院, 淮南 安徽 232001;
    2. 南京信息工程大学遥感与测绘工程学院, 南京 江苏 210044
  • 收稿日期:2020-07-06 修回日期:2020-12-24 出版日期:2021-02-25 发布日期:2021-03-09
  • 通讯作者: 柯福阳。E-mail:Ke.fuyang@qq.com
  • 作者简介:宋宝(1995-),男,硕士生,主要研究方向为传感器导航与定位。E-mail:bsong20@126.com
  • 基金资助:
    江苏省“六大人才高峰”高层次人才项目(XYDDX-045);西宁市科技计划(2019-Y-12);无锡市科技发展资金(N20201011)

The accuracy analysis of iterative weighted IEKF-BP combination algorithm for LiDAR/Radar loose combination

SONG Bao1, KE Fuyang2, ZHAO Xingwang1   

  1. 1. School of Geomatics, Anhui University of Science & Technology, Huainan 232001, China;
    2. School of Remote Sensing & Geomatics Engineering, Nanjing University of Information Science & Technology, Nanjing 210044, China
  • Received:2020-07-06 Revised:2020-12-24 Online:2021-02-25 Published:2021-03-09

摘要: 为了验证目前高精度定位中多传感器组合定位模型性能的优越性,以更好地解决自动驾驶场景下自主定位中出现的预测精度标准不一致、预测不及时及误预测率高等问题,本文利用LiDAR与Radar数据,建立了一种基于迭代加权的IEKF-BP组合算法的松组合模型,并对两种传感器组合定位结果精度进行了分析。试验表明,迭代加权的IEKF-BP组合算法的组合结果精度优于单一的IEKF算法和BP神经网络算法组合定位精度,其中,在XY方向上的均方根误差分别为0.028、0.028 m,平均误差分别为0.023、0.014 m,能准确反映载体的运动状态,满足未来无人驾驶中定位需求。

关键词: 组合定位与导航, LiDAR/Radar松组合定位, 迭代拓展卡尔曼滤波, BP神经网络, 迭代加权的IEKF-BP组合定位算法

Abstract: In order to verify the superiority of the multi-sensor combined positioning model in high-precision positioning, and to solve the problems of inconsistent prediction accuracy standards, untimely prediction, and high misprediction rates in autonomous positioning and navigation, this article proposes a novel LiDAR/Radar integrated positioning model based on iterative weighted IEKF-BP combination algorithm using its data, and the accuracy of the combined positioning results of the two sensors is analyzed. The experiment shows that the combined result accuracy of the iteratively weighted IEKF-BP combined algorithm is better than the combined positioning accuracy of the single IEKF algorithm and the BP neural network algorithm. Among them, the root mean square errors in the X and Y directions are 0.028 and 0.028 m. The average errors are 0.023 and 0.014 m. The result can accurately reflect the movement state of the carrier and meet the future positioning needs of unmanned driving.

Key words: integrated positioning and navigation, LiDAR/Radar loose combination positioning, IEKF, BP neural network, iterative weighted IEKF-BP combined positioning algorithm

中图分类号: