Residual Shoulder Motion Mes Classifier

Abstract

The integration of multiple control strategies into a hybrid system could allow a prosthesis user to perform smoother, multi-joint reaching movements while reducing the necessary mental effort. This work illustrates the usefulness of the shoulder myoelectric signal as an input source to a control strategy by extracting time domain features from these signals and using a linear discriminant analysis classifier to recognize the user’s intended motion. This work is an important first step for future development of hybrid systems that will enable simultaneous control of multiple degrees of freedom used for reaching tasks in a prosthetic limb.