复杂环境下基于RRT\*引导MPPI的移动机器人路径规划

打开文本图片集

关键词：移动机器人；快速随机探索树算法；模型预测路径积分；路径规划；权重函数；动态避障中图分类号：TP18；TP242 文献标志码：A 文章编号：1001-3695（2025）12-019-3675-08doi：10.19734/j.issn.1001-3695.2025.04.0127

Path planning of mobile robots based on RRT* guided MPPI in complex environments

Zheng Jianxiao1，Huang Zijian’，Liu Jinsongl，Tong Xiaoyang1，Chen Gang² （1.Schoolfical&EcricalnXi'Uiesitcure&i chanical Engineering，JiaxingUniversity，JiaxingZhejiang3140o1，China）

Abstract：This paper proposedahybrid path planning algorithmtoaddresstheisues ofpath planning eficiencyand dynamic obstacleavoidanceincomplexenvironments.Thisalgorithmfusedaglobalrapidlyexploringrandomtreewithalocalmodel predictive path integralcontroller.Intheglobal plannngstage，animprovedRRT*algorithm evaluated local environmental complexityviaanawarenessmodel.Italsointroducedadirectionalbiasfactortooptimizepathconvergence.Furthermore，it proposed adaptiverewiring radius and shear interpolation strategies.These strategies improved search eficiencyand provided highqualitynavigationnodes.Inthelocalplanning stage，theMPPIcontrolerutilizedtheinitialpathfromtheRRT*Itemployedanadaptive targetpointstrategy to search forpathnodes within adynamicradius and selected theoptimalone basedon weightevaluation.Thedesigincorporateddynamicobstacleconstraints toenablereal-timepathadjustments，whichenhanced avoidancecapabilities.Experimental results showthe proposed algorithm’sadvantages.ComparedtoastandardRRT*guided MPPI algorithm in complex dynamic environments，the proposed algorithm increased the planning success rate by 12.37% and reduced planning time by 29. 20% .It also achieved a shorter total travel distance.The algorithm demonstrates significant improvements in path planning eficiency and smoothness.It also exhibits excellent robustness andadaptability.

KeyWords：mobilerobot；RRT*；modelpredictivepath integral；pathplaning；weightfunction；dynamicobstacleavoi dance

0 引言

随着移动机器人技术的快速发展，移动机器人已广泛应用于扫地清理和巡检等多种场景[1]。（剩余17614字）