Mobile Station Localization Emitter in Urban NLoS using Multipath Ray Tracing Fingerprints and Machine Learning

摘要

A hybrid technique is proposed to enhance the localization performance of a mobile station in an urban scenario in a Dense Multipath Components. The idea is to use the Ray-tracing simulation tool to build a “radio frequency map” of the Channel Impulse Response of every point in the simulation domain and match the multipath components estimated to a defined location. Conventional localization techniques mitigate errors trying to avoid Non-Line-of-Sight (NLOS) measurements in processing emitter position, while the proposed method uses the multipath fingerprint information produced by RT simulation together with calibration emitters feeds a Machine Learning engine, which refines the target localization embedding all the reflection and diffraction in the propagation scenario. The simulations done showed the feasibility of the proposed method, provided that the buildings can be appropriately included in the estimation of the emitter position.