Stroke pattern classification during manual wheelchair propulsion in the elderly using fuzzy clustering

The purpose of this study was to analyse the kinematic pattern of elderly group during manual wheelchair propulsion. Fourteen elderly persons propelled manually in a wheelchair ergometer. A new objective method based on metrical and topological aspect of the contour of hand center of mass is proposed. A geometric mapping transforms the original time-hand trajectory to a normalized couple of features (R1 and R2). Fuzzy clustering was used to classify wheelchair propulsion pattern based on their features R1 and R2. Four classes were found in order to represent different propulsion pattern. Significant differences were found between classes for fraction of effective force and the biomechanical effectiveness. It was also found that classes are posture dependent and this can help in developing rehabilitation programmes for different groups of patients.