There are approximately 500,000 upper extremity amputees currently living in the United States;with 18,000 new upper extremity amputees added each year. The loss of an upper limb causes a person's quality of life to plummet and brings about massive physical and psychosocial challenges. The majority of amputees are hampered by restricted functionality and use a mechanical hook or a passive, cosmetic hand. Electric Grabber hands are available, but their use is limited due to a cumbersome control mechanism. Therefore, we propose to develop a noninvasive Surface EMG Decoder And Controller (SEDAC) for use in currently available Electric Grabber hands. It employs a feature extractor and an artificial neural network classifier to estimate intended hand movements. This will enable intuitive control of the prosthetic hand. Our Phase I effort is focused on the development and validation of SEDAC for real-time decoding and control of a two-function prosthesis. This offers three key advantages over current technology: 1) intuitive activation of muscle groups for each kind of movement;2) smooth transition from one movement to another;and 3) a learning capacity which transfers the burden of training from the patient to the prosthetic. Our Phase II effort builds upon this to incorporate dimensionality reducing algorithms to improve accuracy, reduce latency, and enable intuitive control of 4 additional hand functions. This will allow for actuation of next- generation dexterous prosthetic hands which are currently under development. Through these advances, we hope to bring about a much needed improvement in the quality of life for upper extremity amputees.
This project will provide trans-radial amputees with intuitive control of current generation prosthetics using surface electromyography (EMG). The technology will also provide a foundation for surface EMG control of fully dexterous prosthetics.
