statistical approach to non-line-of-sight BS identification

Abstract

One of the major challenges to finding accurate mobile terminal location is the non-line-of-sight (NLOS) propagation caused by the blocking of the signal's direct path by obstacles. Traditional location algorithms perform poorly in NLOS environments and this has made it necessary to develop techniques to identify NLOS transmissions and then employ suitable error mitigation techniques. It is extremely difficult to characterize NLOS error because it varies with time and is location specific. We present statistical based techniques to distinguish LOS and NLOS propagation for various scenarios and present simulation results to validate them.