Exchanging transmitter maps in multipath assisted positioning

Abstract

While multipath propagation has commonly been regarded as a drawback for wireless localization technologies, the spatial information contained in multipath components (MPCs) can be exploited for positioning a user. In multipath assisted positioning, each MPC arriving at a receiver is regarded as a line-of-sight signal from a virtual transmitter. We assume the locations of the physical and virtual transmitters to be unknown and estimate them jointly with the user position with simultaneous localization and mapping (SLAM). In a setting where multiple users move in the same scenario, maps of physical and virtual transmitters can be exchanged among them. However, these maps are in different local coordinate systems with unknown relative rotation and translation. The distances among transmitters within each map are exploited in order to find correspondences among transmitters in different maps. Based on the correspondences, the unknown rotation and translation parameters are estimated. This allows a user to exploit the information in a transmitter map received from other users, and hence extends our multipath assisted positioning approach from a single user to a cooperative radiolocation algorithm. In simulations in an indoor scenario we show that using a prior transmitter map decreases the user positioning error although the map is in an unknown coordinate system different from the user's.