An Adaptive IMU/UWB Fusion Method for NLOS Indoor Positioning and Navigation
Daquan Feng,Junjie Peng,Yuan Zhuang,Chongtao Guo,Tingting Zhang,Yinghao Chu,Xiaoan Zhou,Xiang-Gen Xia
DOI: https://doi.org/10.1109/jiot.2023.3245144
IF: 10.6
2023-01-01
IEEE Internet of Things Journal
Abstract:Indoor positioning system (IPS) plays an important role in the applications of Internet of Things (IoT), including intelligent hospital, logistics, and warehousing. Ultrawideband (UWB)-based IPS has shown superior performance due to its strong multipath resistance and high temporal resolution. However, the non-line-of-sight (NLOS) situations noticeably degrade both the positioning accuracy and the communication reliability. To address this issue, we first propose a support vector machine (SVM)-based channel detection method to distinguish the line-of-sight (LOS) and NLOS conditions. Then, one base station (BS)-based distance and angle positioning algorithm with extended Kalman filter (DAPA-EKF) in NLOS environment is proposed. For the LOS environment, least squares (LSs) with EKF processing of acceleration (LS-AEKF) and velocity (LS-VEKF) are developed. To further improve the performance, the combination of time difference of arrival (TDOA) and KF in LOS environment is proposed. Simulation results show that the positioning accuracy of the proposed algorithm is improved in various environments. Finally, validated using more than 1000 testing positions, the positioning accuracy of LS-AEKF is 73.8%–74.1% higher than that of LS-VEKF among the two proposed algorithms in terms of three or four BSs metrics.