移动机器人FASTSLAM算法研究

本科毕业设计题目：移动机器人FastSLAM算法研究学院:专业:学号:学生姓名:指导教师:日期:武汉科技大学本科毕业设计I摘要移动机器人同时定位与地图创建是实现未知环境下机器人自主导航的关键性技术，具有广泛的应用前景，也是目前机器人研究的热门课题之一。基于卡尔曼滤波器的SLAM算法有计算的复杂性以及对数据融合误差非常敏感的缺点，使其不能再实际环境中得到广泛应用。为解决传统SLAM算法的缺陷，介绍了一种基于Rao-Blackwellized粒子滤波器的FastSLAM方法，该方法将SLAM问题分解为对机器人姿态和路标在地图中的位置的递归算法。每一粒粒子都有对应的地图，再将地图估计分解成N个独立的特征估计，路径估计采用粒子滤波器，地图估计采用扩展卡尔曼滤波器。FastSLAM有机地将粒子滤波器与卡尔曼滤波器集成在一起，鲁棒性地解决数据关联和多目标跟踪问题，其时间消耗与路标的数量成对数关系，计算量小，用时短。基于Rao-Blackwellized粒子滤波器的FastSLAM算法是一种高效的机器人同步定位和绘制地图的算法，其具有高效性和准确性，该方法使用提高了机器人地图创建的实时性，增强了避障能力。关键词：移动机器人；FastSLAM算法；路径估计；地图估计武汉科技大学本科毕业设计IIAbstractMobilerobotlocalizationandmappingthekeytechnologiesoftherobotautonomousnavigationinunknownenvironment,hasbroadapplicationprospects,iscurrentlyahotresearchtopicoftherobot.TheKalmanfilter-basedSLAMalgorithmtocalculatethecomplexityandtheerrorisverysensitivetotheshortcomingsofthedatafusion,sothatitcannotbewidelyappliedintheactualenvironment.FastSLAMbasedonRao-BlackwellizedparticlefilterSLAMproblemisdecomposedintothepositionoftherobotposeandlandmarksonthemaprecursivealgorithmtosolvethedefectsofthetraditionalSLAMalgorithm.Aparticlehasacorrespondingmap,andthenmaptheestimateddecomposedintoNindependentcharacteristicsestimatedpathestimationusingparticlefilter,mapestimatedusingextendedKalmanfilter.TheFastSLAMorganicparticlefilterandKalmanfilterintegrated,robustsolutiontodataassociationandmulti-targettracking,thetimeconsumptionandthenumberoflandmarkslogarithmicsmallamountofcalculation,withshort.FastSLAMalgorithmsbasedonRao-Blackwellizedparticlefilterisanefficientrobotsimultaneouslocalizationandmappingalgorithm,itsefficiencyandaccuracy,themethodtoimprovetherobotmapcreatedreal-time,andenhancetheabilitytoobstacleavoidance.Keywords:Mobilerobot;TheFastSLAMalgorithm;Thepathestimation;Themapestimation武汉科技大学本科毕业设计III目录1绪论......................................................................................................................................11.1移动机器人定位和地图创建问题...................................................................................21.1.1移动机器人国内外发展状况.................................................................................21.1.2移动机器人的地图构建问题.................................................................................31.1.3机器人的定位方法.................................................................................................52基于粒子滤波器的SLAM算法.........................................................................................72.1SLAM的通用框架和理论模型................................................................................72.2粒子滤波器定位的基本原理....................................................................................82.3扩展卡尔曼滤波器算法............................................................................................92.3粒子重采样..............................................................................................................102.4移动机器人SLAM问题描述.................................................................................102.4.1SLAM计算复杂度........................................................................................102.4.2SLAM的联合估计........................................................................................112.4.3SLAM的后验估计表示................................................................................112.4.4SLAM公式推导............................................................................................132.4.5有效的数据关联............................................................................................142.4.6FastSLAM的粒子表示形式.........................................................................152.4.7FastSLAM的计算时间复杂度.....................................................................163模型建立............................................................................................................................173.1运动模型..................................................................................................................173.2观测模型..................................................................................................................184FastSLAM算法步骤.........................................................................................................194.1FastSLAM算法步骤...............................................................................................194.2新位姿采样..............................................................................................................204.3环境特征估计的更新..............................................................................................205算法流程图和代码............................................................................................................215.1FastSLAM算法伪代码...........................................................................................215.2FastSLAM算法流程...............................................................................................216仿真环境建立和仿真结果................................................................................................236.1仿真环境介绍.................................................