A method of personal positioning based on sensor data fusion of wearable camera and self-contained sensors

In this paper, we propose a method of personal positioning that combines images taken from a wearable camera with data from self-contained sensors attached to the user through a Kalman filter as a data integration mechanism. The proposed method estimates the user´s position and direction by image registration between the input images from the camera and a set of images captured at known positions and directions beforehand as a database. It updates the estimation of the user´s position and direction with pedestrian dead-reckoning by detecting walking behavior of the user and by estimating the heading direction of the body with the self-contained sensors.