Research on pedestrian indoor positioning technology integrating UWB and PDR

Abstract

For pedestrian indoor positioning in a complex environment, a single positioning technology cannot achieve a good positioning effect. For example, ultra wide band (UWB) positioning technology cannot reduce the influence of non-line-of-sight (NLOS) error by itself. Therefore, this paper uses extended Kalman filter (EKF) algorithm to integrate UWB technology with pedestrian dead reckoning (PDR) technology. PDR positioning technology has the advantages of strong autonomy and high short-time positioning accuracy, which is complementary to UWB positioning technology. In order to facilitate the fusion of UWB and PDR and facilitate the use of pedestrian, the PDR device was fixed on the waist of pedestrian, and a gait detection method based on multiple constraints was proposed for pedestrian gait recognition. Aiming at the problem of NLOS error of UWB in the process of fusion, an adaptive noise variance method is proposed to dynamically adjust the measurement covariance of UWB. Through experimental verification, the average error of the positioning results of the proposed method is 0.21m, which is significantly more accurate than that of pure UWB positioning or pure PDR positioning.