文献类型：期刊文献

中文题名：基于移动机器人的地图构建技术

英文题名：Mapping technology based on mobile robot

作者：杜晨[1];杜煜[2]

第一作者：杜晨

机构：[1]北京联合大学北京市信息服务工程重点实验室,北京100101;[2]北京联合大学机器人学院,北京100101

第一机构：北京联合大学北京市信息服务工程重点实验室

年份：2019

卷号：0

期号：8

起止页码：39-41

中文期刊名：传感器与微系统

外文期刊名：Transducer and Microsystem Technologies

收录：CSTPCD;;CSCD:【CSCD_E2019_2020】；

基金：国家自然科学基金资助项目(91420202)

语种：中文

中文关键词：Rao-Blackwellized粒子滤波器;重采样;粒子退化;粒子多样性

外文关键词：Rao-Blackwellized particle filter;resampling;particle degradation;particle diversity

摘要：对Rao-Blackwellized粒子滤波器算法出现的地图构建不一致性和粒子退化问题进行了研究。在Rao-Blackwellized粒子滤波器的基础上,将采样集中在观测信息的可能性区域,使得采样的粒子更加符合真实环境状态;并且引入分层重采样优化策略,通过控制阈值,维持尽可能多的粒子多样性,有效地解决粒子退化问题。最后在配有16线激光雷达传感器的Bulldog移动机器人平台上进行了实验验证。结果表明:优化的算法减少了粒子数目,增加了粒子多样性,且能创建一致性的环境地图。
The mapping inconsistency and particle degradation problems occur in Rao-Blackwellized particle filter algorithm are studied.On the basis of Rao-Blackwellized particle filter,the sampling is concentrated on the possible areas of the observed information,so that the sampled particles are more fit to the real environment state;and the layered resampling optimization strategy is introduced to maintain as many diversity of particles as possible by controlling the threshold.The problem of particle degradation is effectively solved.Finally,the experimental verification is carried out on the Bulldog mobile robot platform equipped with a 16-line lidar sensor.The results show that the optimized algorithm reduces the number of particles,increases particle diversity,and creates a consistent environmental map.