Estimation of human arm movement with inertial sensors in smartphone

The inertial sensors in smartphone are an type of micro-electro-mechanical systems (MEMS). They cannot be used to estimate the movement directly because of the effect of drift, the noise and lack of calibration of the MEMS sensors causes excessive drift effect. Usually in a typical application of the inertial sensors are compensated by GPS data, stellar, solar or magnetic sensor. However, these compensations systems cannot be used for human arm movement capture. In this paper we propose a novel method to estimate movements less than one meter with an smartphone. This method is completly implemented on an Android smartphone.