The Map Merging Approach for Multi-robot Monocular SLAM

Abstract

In carrying out the movement in an unfamiliar area, a mobile robot must solve a number of important tasks. He should be able to measure, build a three-dimensional map and to determine its position on the map. With access to the map, the robot is able to plan its route. However, such an effective navigation is limited to the boundaries of the area.

If it is necessary to move beyond a well-known area, the robot needs more information. Having access to maps built by other robots that work in adjacent areas, the robot receives the missing information.

In this paper the known methods of collaborative maps created by different robots are presented. Methods for obtaining three-dimensional maps of autonomous robots based on the flow of video data, as well as methods for combining maps of adjacent areas are investigated. A modified method of connecting the maps created by different robots is proposed.