A hybrid wavelet-wiener noise reduction algorithm for geomagnetic signals in dynamic positioning

Indoor positioning technology is a hotspot in current research, and geomagnetic combined positioning is one of the mainstream directions. Currently, indoor or underground positioning techniques are all passive positioning such as RFID, UWB, WIFI and Zigbee. All these passive position techniques are dependent on network and electrical. Our research is focused on active position based on magnetic and PDR position techniques. However, in the case of dynamic walking, the measured geomagnetic data has large disturbance and noise, especially when using low-cost MEMS devices, which significantly impairs the accuracy of geomagnetic positioning. This paper presents an experimental analysis of magnetic disturbance in dynamic measurements and the derivation of the characteristics of magnetic noise during dynamic measurements. Furthermore, a hybrid wavelet-Wiener noise reduction algorithm model is constructed for such noise to improve the dynamic positioning accuracy. The experimental both in indoor corridor and underground tunnel results demonstrate that the hybrid wavelet-Wiener noise reduction algorithm has a pronounced impact on the removal of magnetic noise from dynamic measurements. In comparison to wavelet denoising, it has been observed that the relative root mean square error (RRMSE) can be reduced by 1.5 maximum and the denoising ratio (DNR) by 38 maximum. Furthermore, the geomagnetic DTW (Dynamic Time Warping) matching path value has been found to be similarly reduced from 197954 to 134734. These findings provide a theoretical basis for improving the accuracy of dynamic positioning.