Magnetic field based indoor pedestrian positioning using self-contained sensors

Abstract

The geomagnetic field exists everywhere and has been demonstrated to be local disturbed and stable over time. In this paper, we present an approach for indoor positioning which does not rely upon any additional support equipment, using magnetic fingerprint collected by magnetic field sensors in smartphone and walking information of user measured by self-contained inertial sensors. In order to improve the practicability of our positioning method, we choose the Z-axis magnetic of world coordinate system as feature information, which can reach no restrain of phone’s gesture and orientation when pedestrians walk. Moreover, to improve the discernibility of geomagnetic information, we measure the geomagnetic signals as a sequence, and we match the geomagnetic signals with the established fingerprint using dynamic time warping (DTW) algorithm. Meanwhile, a hidden Markov model (HMM) based positioning process is adopted to strengthen the robustness of the algorithm. Extensive field tests have been conducted in an office building to verify the performance of proposed algorithm. Test results show that the proposed algorithm can achieve less than 1.11 ms of mean error and 2.13 ms of 95th percentile error.