Activity monitoring system using Dynamic Time Warping for the elderly and disabled people

We present a new system that tracks human movements and behaviors for rehabilitative purposes. We applied Dynamic Time Warping (DTW) to recognize human activities of daily living. Ten different movements, standing, sitting, lying on the left side, lying on the right side, lying supine, lying prone, walking, running, stand-to-sit and sit-to-stand are considered and were kept to reference databases signals. Our system consists of two parts: transmitter and receiver. A transmitter part is the device mounted at the user's waist within a pager case measuring 90times40times20 mm. The whole device weighs approximately 50 g including batteries. A sensor used in this device is a 3-axial accelerometer (Hitachi H48C). The signals from the accelerometer are transmitted wirelessly to a personal computer in receiver part using Zigbee Pro 2.4 GHz. A personal computer only requires MATLAB program to recognize our system. DTW is used to match the signals from different behaviors in online with the databases. DTW will find a minimal path between two time series: the test signal and the reference database signal. This minimal value can classify a kind of activity of that test signal. The experiment shows 91 percent accuracy in recognizing these behaviors.