Mobile robot repositioning method and device based on whale algorithm and ensemble learning

The invention provides a mobile robot relocation method and device based on a whale algorithm and ensemble learning, and the method comprises the steps: loading three-dimensional point cloud data of a global map in a working scene, and dividing the global map into a plurality of sub-maps; obtaining a motion trail of the mobile robot after real-time local planning, mapping the motion trail to a plurality of sub-graphs, and then matching to obtain a matching sub-graph meeting a preset matching rule; constructing an inter-frame depth residual image, and obtaining real-time point cloud data of which a dynamic target is filtered by taking pixel information of the depth residual image as a judgment condition; performing whale algorithm adaptation on the real-time point cloud data field and the matched sub-graph to determine the accurate pose of the mobile robot; and inputting the accurate pose into the integrated learning model, and outputting the real-time optimal pose of the mobile robot. In a dynamic complex lim