Non-Line-of-Sight Error Analysis Model Based on Fitting

Abstract

This paper proposed a new method, non-line-of-sight(NLOS) error analysis model based on fitting, in view of the problem that the NLOS error has serious influence on the positioning accuracy during the current indoor positioning process, which has good positioning accuracy after NLOS compensation. In this paper, we establish the NLOS analysis model based on fitting by simulating the propagation path of the signal and analyzing the propagation degree of the signal. Thereby we calculate the indoor NLOS error distribution by the model. The NLOS error value of any position in space is calculated by least squares fitting, which reduces the complexity of the analysis model and improves the prediction accuracy of non-line-of-sight error. Verified by testing, the accuracy of the non - line - of - sight error is less than 0.24m by using the NLOS error analysis model, which has good prediction effect. The root mean square error of the NLOS error by fitting is only 7.031e-09, which shows that this method can better fit the NLOS error. The method we proposed is also applied to the localization, whose positioning accuracy can attain 0.96m. Compared with the quadratic linear programming algorithm based on the Taylor series expansion algorithm, the minimum weighted quadratic positioning technique based on the maximum likelihood estimation algorithm, and the improved Kalman filter algorithm, positioning accuracy and operation speed are improved.