Posture Recognition and Adaptive Step Detection Based on Hand-held Terminal

A postures recognition and adaptive step detection algorithm based on low-cost microelectromechanical system (MEMS) sensors are proposed in this paper. The traditional Pedestrian Dead Reckoning (PDR) algorithm relies on the fixed sensor orientation, is unable to adapt to the high degree of freedom of hand-held terminal in navigation. The recognition method can recognize the postures of smartphone, such as texting, phoning, swinging and placed in the pocket, the algorithm consists of posture definition, feature extraction and class determination. An adaptive step detection algorithm is also proposed according to the detected device posture, the algorithm includes median threshold crossing detection, effective peak detection and adaptive step interval detection. The experiment results show that the proposed algorithm is able to identify the device postures effectively, and detect the pedestrian steps accurately. The accuracy of recognition can be above 95%, the accuracy of step detection can be above 97%. The research can provide important foundation for pedestrian dead reckoning based on hand-held terminals, and also be used in health management, behavior surveillance and other fields.