A tool for linguistic assessment of rehabilitation exercises

In this paper, human motion analysis is performed by modeling a physical complex exercise in order to provide feedback about the patient’s performance in rehabilitation therapies. The Sun Salutation exercise, which is a flowing sequence of twelve yoga poses, is analyzed. This exercise provides physical benefits as improving the strength and flexibility of the muscles and the alignment of the spinal column. A temporal series of measures that contains a numerical description of this sequence is obtained by using a wearable sensing system for monitoring, which is formed by five high precision tri-axial accelerometer sensors worn by the patient while performing the exercise. Due to the complexity of the exercise and the huge amount of available data, its interpretation is a challenging task. Therefore, this paper describes the design of a computational system able of interpreting and generating linguistic descriptions about this exercise. Previous works on both Granular Linguistic Models of Phenomena and Fuzzy Finite State Machines are used to create a basic linguistic model of the Sun Salutation. This model allows generating human friendly reports focused on the assessment of the exercise quality based on its symmetry, stability and rhythm.