A comparison for capturing arm kinematics using solutions of Wahba's problem and ordinary data fusion mechanisms

Rehabilitation is becoming one of critical and essential part in medicine due to fast aging population, the people with various disabilities and increasing health care expenditure. With the advancement of wearable sensor technologies, inertial sensors with magnetometers are well known as a worthy solution for these challenges since it is affordable and precise solution. However, the measurements from IMU devices are engaged with accuracy deficiencies in attitude estimations due to noisiness in measurements. The accuracy improvement of estimation using IMUs has investigated in various studies in literature, though there is a room for further discussions of such improvements with less heavy and complex algorithms. This paper is focused to investigate the applicability of some typical solutions of Wahba's Problem and ordinary filtering mechanism with inertial sensor measurements to accurately capture human arm kinematics. Additionally, these algorithms are compared and validated with real-time sensor measurements.