Research Project
8643937
SUMMARY: Upper limb amputation is experienced by approximately 18,496 persons annually in the U.S., with most of these amputations transradial or more distal. Existing commercial prosthesis devices only allow such users to control one device at a time (e.g., choose between hand actuation and wrist actuation), commonly via EMG-based inputs. Users then switch between devices sequentially. It is reported that direct control of two joints simultaneously-two degrees of freedom of proportional control-is the greatest desired prosthesis improvement described by these users. Several recent research advancements have demonstrated independent, proportional association between 7-128 electrodes and 2-5 degrees of freedom in the fingers or in the hand-wrist. However, the use of high-channel electrode systems will not be practical in the near future in most commercial systems. Other research systems apply multifunction pattern recognition so that amputees can select between a limited set of pre-programmed multi-joint movements. While useful, all functions are still only comprised of one degree of freedom of movement at a time. But, if two degrees of freedom of independent, proportional control could be provided by as few as four conventional electrodes, existing prosthesis hardware could immediately implement the technique. Substantial improvement in function would result. Therefore, we propose the demonstration of an EMG-based prosthesis controller that applies four conventional electrodes on the forearm and provides two independent, proportional degrees of freedom. We propose demonstration of the technique via an EMG-force task on ten able-bodied subjects and five unilateral transradial amputees in which EMG is related to joint torque/force. For amputees, force will be measured from the sound limb during a force mirroring task. Because the best locations of the four electrodes for our proposed approach are not known a priori, we will apply up to 16 electrodes during experimental data collection and then reduce down to four electrodes during offline processing. For the two degrees of freedom, we will investigate three dimensions of contraction: wrist flexion-extension, ulnar-radial deviation and pronation-supination. Offline analysis will determine which two dimensions of contraction provide the most separable control signals for two degrees of freedom in a prosthesis.
