An EMG Controlled Pointing Device Using a Neural Network

Abstract

Electromyogram (EMG) contains information on human movements such as his or her intended motions as well as force levels of muscles, and may be suitable as an input signal for a new human-computer interface tool. In the proposed method, several numbers of base directions are set on a computer display, and a posteriori probabilities that a pointer will move to the base directions are estimated from the EMG signals. Then, operator's intended directions can be estimated as a combination of the finite number of the probabilities corresponding to the base directions. From experiments, it is shown that the pointer movements can be controlled successfully by using the proposed method. Applicability of the proposed method for use of amputees is also examined.