Matching Maps of Physical and Virtual Radio Transmitters Using Visibility Regions

Abstract

Channel-SLAM is a multipath assisted positioning algorithm that treats multipath components as line-of-sight (LoS) signals from virtual transmitters. It maps the physical and virtual transmitters’ locations simultaneously with estimating the user position with simultaneous localization and mapping (SLAM). Since Channel-SLAM is a relative localization system, the coordinate systems of transmitter maps from different users are subject to an unknown relative rotation and translation. In this paper, we present a new algorithm to estimate this rotation and translation, which we call map matching. Map matching is a requirement for collaborative Channel-SLAM, where users share transmitter maps to improve their positioning performance. Our idea is to augment maps of transmitter locations in Channel-SLAM with knowledge on from which locations there is a LoS condition to a transmitter in order to increase the robustness of map matching. We evaluate our algorithm by simulations in an indoor scenario.