TDOA Based Indoor Positioning with NLOS Identification by Machine Learning

Abstract

Indoor positioning technology has broad application prospects, such as intelligent indoor navigation and security monitoring. However, due to the complex and varied indoor scattering environment, non-line-of-sight (NLOS) propagation seriously affects the accuracy of positioning. In this paper, we investigate the NLOS propagation identification for time difference of arrival (TDOA) based positioning in indoor environment. We analyze the correlations between measured distances based on TDOA and select measured distances to identify NLOS data. A new NLOS classification method is proposed based on machine learning to improve positioning accuracy. Numerical results demonstrate the performance of the proposed positioning method.