A Helmet-Mounted Pedestrian Dead Reckoning System

Since most persons in emergency service wear some sort of headgear, we investigate how basic navigation sensors could be incorporated into a safety helmet and thereby be made truly wearable. It is well known that the key to solving the generic ubiquitous localization problem is the combination of different positioning techniques and sensor modalities. In this paper, a combined approach of pedestrian dead reckoning (PDR) and GPS positioning is followed. An novel combination of neural-network step length predictions and helmet-mounted sensors is presented. The experimental system shows low accumulated error over an extended walk and indoor/outdoor navigation is demonstrated.