Map-based Antenna Adaptation for Multipath-assisted Positioning

In this paper, a radio environment map is formulated leveraging a multipath-resolved radio channel model in order to improve the robustness of a multipath-assisted positioning system equipped with directional antennas. A simplified position error bound is derived from the channel model and used as a basis for the map formulation. The map is evaluated in a lab environment on the basis of ultra-wideband (UWB) channel measurements. As an example application, the maps are used in a maximum likelihood based positioning algorithm in order to guide an antenna selection process with the intention to reduce the number of measurements. The algorithm is evaluated on simulated and real measurement data, showing consistent sub-meter accuracy (90% error of 30 cm) while using only a subset of the antennas.