Step count and classification using sensor information fusion

Abstract

In order to suppress the GNSS (Global Navigation Satellite System) limitation to track persons in indoor or in dense environments, a pedestrian inertial navigation system can be used. However, this type of systems have huge location estimation errors due to the Pedestrian Dead Reckoning (PDR) characteristics and the use of low-cost inertial sensors. To suppress some of these errors we propose a system that uses several sensors spread in person's body combined with information fusion techniques. Information fusion techniques provide lighter algorithms implementations, to count and classify the type of step, to run in mobile devices. Thus, improving pedestrian inertial navigation systems accuracy.