Evaluation of Aligning Methods for Landmark-Based Maps in Visual SLAM

摘要

This work has been focussed on the alignment problem of visual landmark-based maps built by the robots. The scenario presented is that of a team of robots that start their navigation tasks from different positions and independently, i.e., without knowledge of other robots’ positions or observations. These robots share a common area of a typical office building. The maps built with the FastSLAM algorithm are initially referred to the reference system of each robot, located at their starting positions. In this situation, we consider the possibility of merging these maps into a global map. However, in order to do that, the landmarks of these