Indoor Positioning With Adaptive Wavelet Denoise Enhancement and Trend Surface Analysis Based Multipath Map

The channel multipath and non-line-of-sight (NLOS) reflections introduce a bias on the time difference of arrival (TDOA) observations, which reduce the accuracy of the ultra-wideband (UWB) positioning performance in an indoor environment. In previous research, the multipath map (MPM) method is based on spatial domain modeling principle in the reverse positioning framework and is used to mitigate the multipath effect. Utilizing the mean value of the multipath residuals as the calibration value will inevitably lead to insufficient or excessive multipath modification in the grid. By combining the trend surface analysis method, this paper proposes an advanced multipath mitigation method named trend surface analysis based multipath map (TMPM). Meanwhile, this paper propose an adaptive wavelet denoise (AWD) method in order to eliminate the noise of data set in advance. We mainly demonstrate the effect of TMPM in reverse positioning scenarios, analyze the multipath correction effect of TMPM. The experimental results show that the accuracy of the tag positioning is 16.0 cm with TMPM, which is improved by 22.7% compared to that of MPM. In dynamic positioning case, results with TMPM also show benefit over MPM.