3D indoor layered localization of multi-information fusion via intelligent terminal

Abstract

Aiming at the problem that that pedestrian can only get low accuracy and plane coordinate of current indoor localization from intelligent terminal, a layered indoor 3D positioning technology based on WiFi positioning, pedestrian dead reckoning (PDR) and architecture information is proposed. In order to estimate height changes of pedestrian, the two pedestrian behavior patterns of horizontal walking and up-down stairs can be distinguished by detecting the pitch-rate mutation. Height changes are got by using the effective steps and stairs information. The equations of the two different behavior patterns are constructed and three-dimensional PDR is realized. A particle filter algorithm combining multi navigation information is proposed to eliminate the cumulative errors of PDR. Fusing the architecture information and behavior patterns into the weight calculation and the sampling of the particle, the localization precision is effectively improved. The results show that the average error of the horizontal location is 2.1 m and the average error of the height estimation is 0.8 m. The accuracy could satisfy the demand of pedestrian location service using individual intelligent terminal.