Robust Urban Wireless Localization: Synergy Between Data Fusion, Modeling and Intelligent Estimation

In this paper, we present a viable Bayesian estimation alternative to mobile localization enhancement in mixed line-of-sight (LOS)/non-LOS (NLOS) urban areas. The development of the proposed approach relies on a synergistic combination of valid aggregate measurements, NLOS bias modeling and estimation, and computational intelligence. For reliable wireless positioning, we first introduce valid range measurements in which the effect of the NLOS range bias due to small-/large-scale multipath fading is limited. Subsequently, we propose a hybrid system framework with Markovian state transitions, data fusion of valid range and signal power, NLOS bias modeling, and fuzzy inferences for modeling the dynamics of a mobile station (MS) with respect to each base station (BS).

The proposed framework enables us to develop a selective fuzzy-tuned extended Kalman filtering based interacting multiple-model (SFT-IMM-EKF) algorithm for each BS to perform mobile location estimation. We show that due to the synergistic effects, the proposed SFT-IMM-EKF can remarkably improve the IMM, the SFT-IMM and the IMM-EKF. The result is substantiated by numerical simulations. As well, it is demonstrated that the proposed algorithm can robustly leverage against the adverse impacts of severe NLOS errors and MS mobility variations.