Geomagnetic/PDR Fusion Localization Method Based on Smartphones

Abstract Indoor pedestrian positioning is an important foundation for location services, and because geomagnetic signals are readily perceived, positioning methods based on geomagnetic signals have been a hotspot in indoor pedestrian positioning research. Addressing the issues of large cumulative error and low positioning accuracy in current particle filter fusion-based positioning methods, this paper proposes Geomagnetic/PDR Fusion Localization Method indoor pedestrian positioning method based on smartphone. This method combines the particle swarm optimization algorithm with the traditional particle filter algorithm to perform optimal position finding and improve real-time positioning accuracy. Then, the DTW-A* (DTW, Dynamic Time Warping) algorithm is established to obtain the variable-length geomagnetic sequence to correct the cumulative error over a period of time, in order to solve the cumulative error problem of the particle filter-based positioning method. The proposed method is compared with existing mainstream positioning methods through experiments, showing that the average errors of the proposed method are 0.90m and 0.72m in two typical scenarios, significantly lower than those of the comparative methods, PDR, MaLoc, and Magicol.Additionally, experimental validation is conducted on different types of mobile phones and with users of varying heights, demonstrating that the method is not only applicable but also stable. This robustness suggests that the method is expected to support indoor positioning across different devices.