Self-Calibrated Multi-Floor Localization Based on Wi-Fi Ranging/Crowdsourced Fingerprinting and Low-Cost Sensors

Crowdsourced localization using geo-spatial big data has become an effective approach for constructing smart-city-based location services with the fast growing number of Internet Things terminals. This paper presents a self-calibrated multi-floor indoor positioning framework combination Wi-Fi ranging, crowdsourced fingerprinting and low-cost sensors (SM-WRFS). The parameters, such as heading altitude biases, step-length scale factor, ranging bias are autonomously calibrated to provide more accurate forward 3D performance. In addition, backward smoothing algorithm novel deep-learning model applied in order construct autonomous efficient database detected quick response (QR) code-based landmarks. Finally, adaptive extended Kalman filter is adopted combine corresponding sources different integration models precise multi-source fusion based real-world experiments demonstrate that presented SM-WRFS proven realize under environments, meter-level accuracy can be acquired supported areas.