A Novel Two-Stage IMU/MAG-Aided UWB Fusion Algorithm for Pedestrian Localization in Underground Space

Nonline-of-sight (NLOS) propagation is one of the main reasons for degrading the performance of the ultrawideband (UWB) positioning technique. Aiming to improve the performance for pedestrian location estimation in an underground space environment with UWB, fuzzy comprehensive evaluation (FCE) NLOS identification algorithm based on Wilcoxon signed rank test (WSRT) and entropy weight method (EWM) is proposed, first. Then, this article presents a two-stage UWB-based fusion positioning scheme with the aid of an inertial measurement unit (IMU) and geomagnetic (MAG). In the first stage, a combination algorithm based on genetic algorithm (GA) is proposed for IMU and MAG data fusion. In the second stage, when a sufficient number of UWB ranging values are not available under NLOS conditions, an IMU/MAG aided UWB fusion improvement robust extent Kalman filtering (IREKF) for location estimation is introduced, which makes full use of the ranges value of the UWB in the residual LOS condition. Finally, the experiment is carried out in underground space and the experimental results show that using a two-stage fusion algorithm, the proposed fusion positioning scheme can effectively improve positioning accuracy and robustness, especially in the area of UWB NLOS situation.