Investigation of wireless rehabilitative assessment data using singular spectrum analysis

To assist people with disabilities caused by accidents or illnesses regain their social life and activities it is important to assess their condition. As well, it is also important to monitor their progress in rehabilitation, in order to determine the regimen and intensity of therapy. At present many objective assessments require therapists to subjectively grade various tasks the patients perform. The administration of these assessments is manual, making it error prone and tedious. In this paper, we make use of wireless body sensors using Inertial Measurement Units (IMUs) worn by the patients to accurately measure and assess the ability of a patient to perform various tasks which are part of clinically established tests of functionality. We also are able to capture the nuances of movements which are not easily obtainable by other means. We focus on the use of accelerometers and gyroscopes in these tests and their use in detecting abnormalities in the performed tasks. Using Singular Spectrum Analysis in a sub-space based analysis gives useful results which will infuse intelligence into automated systems to aid and guide therapy.