四足机器人室内导航的多模块优化Fast-SLAM算法研究

打开文本图片集

中图分类号：TP242.6 文献标识码：A 文章编号：2096-4706（2025）18-0169-05

Abstract：Aiming atthe problemsof insuffcient positioning accuracy，lowmapping effciencyand weak systemrobustness ofquadrupedrobots inindoorenvironment，thispaper proposesamulti-moduleoptimizedFast-SLAMalgorithm framework. Inthe front-endodometrystage，the inter-framemotionconstraintsoflaser pointcloudsareconstructedbasedontherange flowmodel，andtheadaptiveiterativeweightedleastsquares methodisintegratedtoimprovethaccuracyandcomputational efciencyofposeestimation.Inthestateestimationstage，thefuzzyadaptiveExtendedKalmanFilterisintroducedtoenhance therobustnessofthe system bydynamicallyadjusting themeasurement noisecovariance.Intheparticlefilterstage，the poseoptimization method combining Maximum Likelihood Estimation and gradient search is usedto effectivelyreduce the resampling errorandimprovethe mappingeffciency.Thealgorithmis deployedandverifiedonarealquadrupedrobotplatform. Theexperimentalresultsshowthat theaverage trajectoryerorissignificantlyreducedandthecalculationtimeissignificantly shortenedcompared withthetraditionalmethods，showinggoodenvironmentaladaptabilityandengineringaplicationpotential.

Keywords： indoor navigation; Fast-SLAM; range flow algorithm; Extended Kalman Filter; particle filter

0 引言

随着机器人系统与自主导航技术的持续发展，四足机器人在环境监测、工业应用与室内服务等领域具有广泛的前景。（剩余5615字）