Wireless Sensor Network based movement classification using wrist-mounted 9DOF sensor boards

In this paper, a new, linear discriminant analysis (LDA) based, movement recognition system is introduced, which uses wrist-mounted Wireless Sensor Network (WSN) motes equipped with 9 degree of freedom (9DOF) sensor boards. The 9DOF sensor board is built up from a tri-axial accelerometer, a gyroscope, and a magnetometer. Different arm movements can be used to detect emergency situations. In order to recognize specific arm movements in stationary positions and also during the movement of the body, eleven movement classes were constructed. Measurement data for all classes were collected from nine subjects, and used for training and validation of the system. Different time-domain features (TDF) were calculated in the processing windows. The recognition rates were tested with different sensors, features, window sizes and window shifts.