An Improved Location Calibration Method for Indoor Pedestrian Positioning

Pedestrian dead reckoning (PDR) has emerged as an effective technique employed in location services for the Internet of Things. However, PDR confronts with cumulative errors, necessitating the location calibration. Conventional location calibration on the basis of received signal strength indicators (RSSIs), encounters the difficulty of selecting an appropriate threshold to reduce significant calibration errors. To address the difficulty of RSSI-based location calibration, we propose a novel distance approaching detection (DAD) method: 1) DAD improves the detection process by replacing RSSI with the distance between pedestrian and Bluetooth station. 2) DAD refines the detection rule by suggesting the location calibration epoch based on the minimum distance between pedestrian and reference Bluetooth station. Compared to conventional RSSI threshold detection method (CRT), DAD minimizes location calibration errors and enhances the precision of PDR. Furthermore, evaluation results across multiple routes demonstrate following improvements of DAD: 1) PDR trajectories using DAD closely align with the reference routes, while significant deviations are observed when employing CRT. 2) The mean of distance errors for PDR with DAD is reduced 0.36 m compared with CRT, resulting in an improvement of precision by 45%. 3) The easy configuration of DAD is helpful to cope with the challenge of adeptly and accurately setting the parameters in complex indoor environments for CRT, and Extended Kalman Filter with PDR/Bluetooth.