In classic radar localization scenarios the transmitter and the receiver are in the same spot localizing a passive target by illuminating it with EM waves. However, more and more localization applications arise, where there is an active beacon/transmitter which is to be localized by some multiple-antenna base station, e.g. a forklift in a warehouse, a car in an urban scene, and many more. For those localization applications, electromagnetic waves in the microwave spectrum are well suited, especially when exploiting the knowledge about the phase relations of the receive signal at different antennas of the receiver. The applied localization algorithms generally consider the EM wave to travel directly from the beacon to the base station, which is not true in a reflective environment (warehouse, city, etc.). In such environments the wave also travels along non-direct paths from the beacon to the receiver due to reflections causing the so called “multi-path propagation”, which causes heavy degradation of the localization accuracy especially in phase-based localization algorithms.
In this talk an algorithm is presented which uses compressed sensing and a high antenna count and physically large multi-channel receiver in order to estimate and subtract the impact of the non-direct propagation paths from the receive signal. Based on measurements in a 5.8 GHz radio setup, it will be shown that this algorithm is well-suited to improve the position estimations in such scenarios.