A pilot study of activity recognition on rehabilitation exercise of frozen shoulder using wireless inertial sensor node

This study establishes wireless sensor network (WSN) inertial sensor nodes (ISN) that comprise micro control unit, ZigBee-compatible radio frequency chip, tri-axial accelerometer, biaxial gyroscope, single-axial gyroscope, and a planar inverted-F type antenna on a four-layer printable circuit board with size of 40mm × 37mm × 2mm. Two wireless ISNs are attached on arm and wrist for measuring rehabilitation exercise of frozen shoulder. The measured data are sent wirelessly to a base node where a Matlab-based program is developed for retrieving packet, parsing packets, and recognizing motion data based on the artificial neural network (ANN) algorithm. Six rehabilitation exercises for frozen shoulder are measured by wearing two nodes on the wrist and the arm. Accelerations in three axes and the derived angle form a 4-tuple vector as the motor feature of employed recognition algorithm. As results, five of six exercises are successfully recognized with 85-90 % of accuracy rates but the complex one (i.e. the spiral rotation exercise) reached only around 60 %. The pilot study approves good feasibility of self-developed WSN ISNs to recognize rehabilitation exercises as well as contribute advanced applications for mobile or ubiquitous health care in the future.