A Gaussian Unscented Kalman Filter algorithm for indoor positioning system using Ultra Wide Band measurement

Abstract

In order to further improve the accuracy of the non-linear positioning model in the research of ultra wide band (UWB) indoor positioning, a Gaussian unscented Kalman filter (GUKF) algorithm is proposed in this paper. This localisation algorithm first uses a Gaussian function to design a Gaussian smoothing filter template to process the smoothing of experimental data in the GUKF algorithm, and then the filtering algorithm is used to obtain higher positioning accuracy. This paper utilises simulations and actual experiments to verify and analyse the GUKF algorithm, and the actual experiment environment was divided into line-of-sight (LOS) and non-line-of-sight (NLOS) experimental environments. The measured experimental results indicate that in the static test of location tags in LOS and NLOS experimental environments, the root mean square error (RMSE) of the GUKF algorithm is reduced by 15.88% and 14.10%, respectively; in the dynamic test, the RMSE of the GUKF algorithm is reduced by 16.67% and 17.89%, respectively, compared with the unscented Kalman filter algorithm. In addition, the positioning performance evaluation method of the mean error and cumulative distribution function curve also verifies that the GUKF algorithm has a higher positioning accuracy than the UKF, Least Squares, and Time of Arrival algorithms.