Smartphone-based Indoor Navigation Using PDR and Magnetic Matching

摘要: This paper presents an integrated pedestrian deadreckoning (PDR) /Magnetic matching (MM) navigation algorithm with off-the-shelf sensors in smartphones. method maximizes the advantages and minimizes disadvantages of each technology. As magnetic is a profile-matching method, we calculate rough length measured profile using steps detected by accelerometers. Because accurate real-time steplength estimation still open issue for navigation, utilize dynamic time warping (DTW) inaccurate length. Additionally, detect roll pitch angles accelerometers smartphones extract vertical horizontal intensities, order to increase fingerprint dimension without extra sensors. Furthermore, use gradient fingerprints, instead intensity mitigate effect diversity run-to-run biases MEMS magnetometers. Tests different indicated that matched MM results were had small fluctuations. Nevertheless, although algorithm, it mismatches (i.e., matches incorrect point more than 20 m away from true value) on several occasions. The integration PDR has potential provide continuous reliable indoor environments having frequent interferences, under natural device motion modes including handheld, at ear, dangling hand, pocket.