An EMG controlled pointing device using a neural network

In this paper, we argue that myoelectric signals (EMG) measured from surface electrodes can be utilized as an interface tool for the handicapped. The EMG signal contains information about the operator´s intended motion as well as the force level of the muscles and may be suitable as an input signal for a new interface tool. This paper proposes an EMG controlled painting device using a neural network and develops a prototype system. In the proposed device, an operator´s intended direction of the pointer is estimated using the finite number of base directions which are set on the computer display. The neural network has to estimate the probability that the pointer will move to each base direction, so that the heavy learning calculation and the huge network structure are not neccessary. Through experiments, it is shown that the direction and the velocity of the pointer movement can be controlled by using the EMG signals