Enhanced urban mobile localization via partitioned NLOS bias model approach

In this paper, we present an alternative to the enhancement of urban microcellular wireless localization based on aggregate received range and signal power along with the proposed partitioned NLOS bias models. Specifically, we preprocess received measurements as validated measurements and propose a state-space system framework which covers validated NLOS biases by Markov-transitioned multiple models. This leads to a location estimation algorithm, an improved selective fuzzy-tuned IMM-EKF (SFT-IMM-EKF) method. The proposed algorithm can noticeably enhance the SFT-IMM-EKF since the proposed framework is distinct from existing system models, in which the NLOS bias dynamics is either ignored or modeled by a random walk process. This is shown by simulation studies.