Simultaneous localization and mapping in multipath environments

This paper presents and extends the idea of multipath assisted positioning, named Channel-SLAM. Channel-SLAM uses multipath propagation to allow positioning for an insufficient number of transmitters or increase the accuracy otherwise. Channel-SLAM treats multipath components (MPCs) as signals from virtual transmitters (VTs) which are time synchronized to the physical transmitter and fixed in their position. To use the information of the MPCs, Channel-SLAM estimates the receiver position and the position of the VTs simultaneously using simultaneous localization and mapping (SLAM) and does not require any prior information such as room-layout or a database for fingerprinting. This paper investigates mapping, where we derive a probabilistic map representation based on the receiver positions. Thus, if the receiver knows its current location, the information in the probabilistic map helps to estimate the trajectory of further receiver movement. Hence, as soon as the receiver returns to an already mapped position, information of the probabilistic map can be used for the movement to obtain better estimations of the receiver position. The algorithm is evaluated based on measurements with one fixed transmitter and a moving pedestrian which moves on partially overlapping loops.