A Dynamic Window Based UWB-Odometer Fusion Approach for Indoor Positioning

• Published in: IEEE sensors journal Vol. 23; no. 3; p. 1
• Main Authors: Zhang, Hui, Zhou, Xidong, Zhong, Hang, Xie, He, He, Wei, Tan, Xuan, Wang, Yaonan
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.02.2023; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Indoor positioning; Localization; mobile robot; particle filter (PF); ultra wideband (UWB); Ultrawideband
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2022.3228789

The positioning system based on ultra-wideband (UWB) can achieve centimeter-level positioning accuracy, which is now widely used for indoor positioning. However, UWB is subject to multipath effects and non-line-of-sight (NLOS), which can cause positioning errors. And a UWB positioning system with a minimum of three anchors is costly and difficult to deploy, which is not an economical option. This paper presents a low-cost UWB-odometer fusion method for mobile robot localization, which enables global localization using only one UWB anchor. To improve the positioning accuracy and eliminate the influence of cumulative odometer error, we propose a dynamic window based particle filter (DWBPF). It solves the problem of particle convergence and track loss. It also adaptively modifies the score weights according to the power difference of the UWB to achieve a more robust and accurate positioning. The experimental result shows that the system has a high positioning accuracy of 0.061 m.