Hybrid indoor localization scheme with image sensor-based visible light positioning and pedestrian dead reckoning

In this paper, we propose a hybrid indoor localization scheme with image sensor-based visible light positioning and pedestrian dead reckoning. The architecture and algorithm of the proposed positioning scheme are analyzed. We build and demonstrate two different positioning prototypes with single/multiple cells, which are based on a commercial light-emitting diode lamp and off-the-shelf mobile phone equipped with an image sensor and inertial measurement unit. Experimental results show that the proposed visible light positioning scheme can achieve both cell recognition and three-dimensional positioning with a single captured image of a single light source, and the positioning error is only several centimeters. Moreover, with the proposed hybrid positioning algorithm consisting of visible light positioning and pedestrian dead reckoning, we achieved real-time decimeter-level indoor location tracking continuously and robustly with sparse light-source beacons in an actual indoor scenario.