Cramer-rao lower bound for localization in environments with dynamical obstacles

Abstract

Cramer-Rao Lower Bound (CRLB) of location estimation under Gaussian distribution is widely used in localization applications. However, under the environments with dynamical obstacles, the existing CRLB does not represent the effect of the non-line-of-sight (NLOS) bias caused by dynamical obstacles. In this paper, based on received signal strength (RSS) measurements, a uniform random variable is used to model the NLOS bias effect. Furthermore, The corresponding maximum likelihood estimator (MLE) and CRLB under the joint distribution of Gaussian distribution and uniform distribution are derived. Numerical results validate that the proposed MLE and CRLB are effective in environments with dynamic obstacles.