基于改进水母搜索优化的FastSLAM算法  

FastSLAM Algorithm Based on Artificial Jellyfish Search Optimization

作　　者：王林 杨光永[1] 刘福康 徐天奇[1] WANG Lin;YANG Guangyong;LIU Fukang;XU Tianqi(School of Electrical Information Technology,Yunnan Minzu University,Kunming 650000)

机构地区：[1]云南民族大学电气信息工程学院,昆明650000

出　　处：《计算机与数字工程》2025年第2期297-302,共6页Computer & Digital Engineering

基　　金：国家自然科学基金项目(编号:61761049,61261022)资助。

摘　　要：针对传统FastSLAM算法中存在的proposal分布跟实际分布相差较大,因此需要大量粒子才能较好地表示后验分布,造成内存爆炸,提出了用改进水母搜索优化FastSLAM算法,首先将适应度值不佳的粒子进行混沌处理,其次在水母位置更新时加入小波变异,最后用改进水母优化算法对FastSLAM粒子采样进行更新,提升proposal分布采样的位姿质量。实验分析表明:该方法能够有效减少机器人定位建图的误差,提高其工作效率,可以应用于改进FastSLAM算法的研究。Aiming at the fact that the propertyal distribution existing in the traditional FastSLAM algorithm is quite different from the actual distribution,so a large number of particles are required to better represent the posterior distribution,causing a mem⁃ory explosion,the FastSLAM algorithm is proposed to optimize the FastSLAM algorithm with improved jellyfish search,first of all,the particles with poor adaptability values are chaotic processed,and then the wavelet variation is added when the jellyfish position is updated,and finally the fastSLAM particle sampling is updated with the improved jellyfish optimization algorithm.The pose quali⁃ty of property distribution sampling is improved.Experimental analysis shows that the proposed method can effectively reduce the er⁃ror of robot positioning mapping and improve its work efficiency,and can be applied to the study of improving FastSLAM algorithm.

关 键 词：FASTSLAM proposal分布 水母优化 定位建图 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]