Association of Transmitters in Multipath-Assisted Positioning

A huge variety of services require a precise localization. While global navigation satellite systems may show accurate positioning results in good view-to-sky conditions, their performance decreases drastically in case of shadowing and multipath propagation, such as indoors or in urban scenarios. Our approach is therefore to use terrestrial signals of opportunity for positioning. We exploit multipath propagation in a multipathassisted positioning approach: each multipath component is regarded as being emitted by a virtual transmitter in a line-ofsight condition. Since the locations of the virtual transmitters are unknown, they are estimated in addition to the user position. This results in a simultaneous localization and mapping (SLAM) problem, where physical and virtual transmitters are considered as landmarks. This paper discusses our approach named ChannelSLAM, and extends it by a solution to the data association problem. We present and compare two different methods to decide for associations among virtual transmitters. By means of simulations, we show that data association can increase the positioning performance of Channel-SLAM remarkably.