Multi-agent Radio SLAM: A Point Cloud-based Approach

Abstract

This paper proposes a multi-agent radio simultaneous localization and mapping (SLAM) employing a point cloud-based approach to enhance the accuracy, speed, and stability of the mapping. Traditional radio SLAM relies on measurements from a single sensing source, resulting in low reliability and limited sensing coverage. To overcome these challenge, the proposed algorithm fuses local maps of individual user equipments (UEs) through a point cloud-based approach. In a scenario with two moving UEs, the simulation results demonstrate that the angular error converges to a smaller value at a faster rate, and the Chamfer distance reaches zero nearly twice as fast compared to the single-agent cases.