基于改进RBPF的激光SLAM算法研究  

Research on Laser SLAM Algorithm Based on Improved RBPF

作　　者：王馨阁 田宗强 陈文涛 张雨 杨光 王伟 陈孟元 WANG Xinge;TIAN Zongqiang;CHEN Wentao;ZHANG Yu;YANG Guang;WANG Wei;CHEN Mengyuan(School of Electrical Engineering,Anhui Polytechnic University,Wuhu 241000,China;School of Electrical and Electronic Engineering,Anhui Institute of Information Technology,Wuhu 241000,China)

机构地区：[1]安徽工程大学电气工程学院,安徽芜湖241000 [2]安徽信息工程学院电气与电子工程学院,安徽芜湖241000

出　　处：《安徽工程大学学报》2020年第6期18-25,共8页Journal of Anhui Polytechnic University

基　　金：国家级大学生创新创业训练基金资助项目(201810363105,201910363017,202010363025);安徽工程大学大学生科研基金资助项目(2020DZ28);安徽省高校自然科学重点基金资助项目(KJ2018A0627)。

摘　　要：针对基于RBPF-SLAM的传统算法在地图构建过程中会存在由于粒子退化等导致定位精度不足的问题,提出了一种改进的RBPF-SLAM方法。为了构建精确的地图,将运动里程计与激光测量数据相结合,调整粒子权重,并通过自适应重采样的方法重新估计粒子的权重,得到精确的栅格地图,改进了粒子滤波在地图估计中的不足。仿真结果表明,通过将运动里程计信息和激光测量数据相结合并通过自适应重采样的方法,解决了粒子权重退化问题,优化后的激光SLAM算法构建的环境地图效果更好且定位精度得到提升。Aiming at the problem of insufficient positioning accuracy due to particle degradation in the map construction process of traditional algorithms based on RBPF-SLAM,an improved RBPF-SLAM method was proposed.In order to build an accurate map,the paper combines the motion odometer with laser measurement data,adjusts the particle weight,and re-estimates the particle weight through the adaptive re-sampling method to obtain an accurate grid map,and finally improve particle filtering in map estimation of inadequacy.The simulation results show that by combining the motion odometer information with the laser measurement data and adopting the adaptive resampling method,the particle weight degradation problem was solved,and the environment map constructed by the optimized laser SLAM algorithm has better effects and improved positioning accuracy.

关 键 词：激光雷达 SLAM 粒子滤波 自适应重采样 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]