A Novel Deep Learning Approach to 5G CSI/Geomagnetism/VIO Fused Indoor Localization

• DOI: 10.3390/s23031311
• Publication Date: 2023-01-24T06:29:45Z
• Authors: Chaoyong Yang; Zhenhao Cheng; Xiaoxue Jia; Letian Zhang; Linyang Li; Dongqing Zhao
• Keywords: Odometry; Fuse (electrical)

For positioning tasks of mobile robots in indoor environments, the emerging technique based on visual inertial odometry (VIO) is heavily influenced by light and suffers from cumulative errors, which cannot meet requirements long-term navigation positioning. In contrast, techniques that rely signal sources such as 5G geomagnetism can provide drift-free global results, but their overall accuracy low. order to obtain higher precision more reliable positioning, this paper proposes a fused 5G/geomagnetism/VIO localization method. Firstly, error back propagation neural network (BPNN) model used fuse geomagnetic signals results; secondly, conversion relationship VIO local results coordinate system established through least squares principle; finally, state extended Kalman filter (ES-EKF) constructed. The experimental show 5G/geomagnetism fusion method overcomes problem low single sensor accurate results. Additionally, after fusing average robot two scenarios 0.61 m 0.72 m. Compared with VINS-mono algorithm, our approach improves environments 69.0% 67.2%, respectively.