Interactive and Wearable MIP Recognition Technique Combine Pattern recognition and Spatial Tracking based on SPG Sampling

摘要

Health monitoring based on sphygmogram (SPG) intelligent analysis is a promising alternative for cardiovascular diseases pre-diagnosis. Unfortunately for a wearable SPG intelligent analyzer the qualified SPG sampling is a stern challenge due to user wrist's any movement or incorrect posture (MIP) might result in distorted SPG morphology and analysis failure. Hence, using micro inertial measurement unit (MIMU) and pattern recognition technology in our previous research could solve the problem generally [3]. However, on account of the limitation of MIP database, a more powerful method on the base of MIP pattern recognition, which could recognize any movement of user's wrist (like paddling geometric graphics) and give machine-voice-guide as man-machine feedback, is desired. In this paper, by using acceleration integral calculation to infer the spatial relative displacement and Kalman filter technology to reduce null-drift effect, user wrist's any exact MIP and continuous movement trajectory could be detected well. The testing results show that such a technique is valid in reflecting various spatial movements accurately and fast, which provides solid basis for realizing movement control and make it available in any wearable sensor application.