基于回波强度的AGV重定位方法研究

Research on AGV relocalization method based on echo intensity

自动引导车(Automated Guided Vehicle,AGV)在具有坡道的多楼层环境导航时,搭载2D激光雷达的AGV仅仅只能通过激光同时定位与建图(Simultaneous Localization and Mapping,SLAM)进行单个楼层建图,无法直接生成多楼层环境的地图,只能通过建立多楼层地图之间的拓扑关系,才能实现不同楼层地图的转换。为了解决多楼层地图导航之间的重定位问题,提出一种基于回波强度的AGV重定位方法。首先基于激光雷达的回波强度,过滤非反光板信息;然后基于DBSCAN算法对激光扫描数据进行聚类,并通过最小二乘法拟合数据,获得不同反光板质心的局部坐标集合;最后依据反光板布局信息匹配反光板数据,采用几何关系法实现AGV在新地图的重定位。实验结果表明,该方法能有效实现AGV的重定位,定位精度高且稳定,不影响后续的导航任务,具有实际应用价值。

AGV based on 2D lidar can only map single floor through laser SLAM when navigating multi-floor environment with ramps,and cannot directly map multi-floor environment.The conversion of different floor maps can only be completed by establishing a topological relationship between multi-floor maps.In order to solve the problem of relocalization between multi-floor map navigation,an AGV relocalization method based on echo intensity is proposed.First of all,non-reflective board information is filtered based on the echo intensity of the lidar.Then,the laser scan data is clustered based on DBSCAN algorithm,and the local coordinates of different reflector centroids are obtained by fitting the data by least squares.Finally,AGV relocalization of the new map is realized based on the known reflector layout information matching the reflector data.Experimental results show that the method can effectively complete AGV relocalization with high localization accuracy and stability,avoid affecting the subsequent navigation tasks,and have practical application value.