A unified analysis for Cramer-Rao Lower Bound for non-line-of-sight geolocation

摘要

Non-line-of-sight (NLOS) geolocation becomes an important issue with the fast development of mobile communications in recent years. Several methods have been proposed to mitigate NLOS effects in geolocation. However, there has been no systematic study reported on how the existing methods are related to one another and what is the best geolocation accuracy that we may possibly achieve. As first step to answer these questions, here we present a unified treatment to obtain Cramer-Rao Lower Bound (CRLB) for time-of-arrival (TOA), maximum likelihood estimation (MLE) and signal strength (SS) based positioning methods (to be defined below). Its physical interpretation is discussed and is very helpful to establish connection among the three methods. It can be shown that the CRLB of MLE is equivalent to that of first estimating time delay with a matched filter and then perform TOA positioning. One interesting observation is that the CRLB depends only on the signals obtained by line-of-sight (LOS) base stations and NLOS signals are completely discarded. Furthermore, we notice that a similar conclusion can be draw without the explicit expressions of NLOS geolocation. Thus we extend the current problem to a broader formulation of the CRLB of parameter estimation with unwanted parameters.