Fuzzy-EMG-based Assistive interface for children with Spinal-Muscular-Atrophy

Abstract

Spinal Muscular Atrophy (SMA) is a progressive neuromuscular disorder. Usually, this condition is considered genetically induced with no known cure to date. Children are born with the condition and develop muscular weakness progressively as they grow. The weakness ultimately encompasses the whole muscular function rendering the limbs dysfunctional or paralyzed. Many children with SMA, if they do not have the weakness from the beginning, will start having the disease manifesting itself on the legs first and then the arms and, in due time, they will become quadriplegic and even more disabilities can follow including speech impairment. Assistive Technology support for people with such disabilities often requires identification of the best residual muscular function so that this can be utilized as a means of voluntary control. Electromyography (EMG) is a popular clinical procedure to monitor muscular function in a large number of healthcare and other clinical measurements. It translates muscular activity into proportional voltage signals which can then be used for analysis and other applications. Most of the existing assistive applications are based on amplitude thresholding of the EMG signals, which can drift over time due to fatigue on part of the patient and partly due to changes at the electrode interface over the period of use. This requires that a care-giver must re-calibrate the signal threshold making the process both impractical and prone to errors. In this paper, a new approach has been presented that alleviates the need for re-calibration of thresholds for such applications development. Fuzzy classifier has been used on pattern-related features from the signal samples and based on that appropriate computer signals can be generated to be adapted in an Assistive application such as playing a computer game, using serial keyboard interface, controlling the wheelchair/other-hardware, or even being able to generate text.