Design of an electromyographic mouse

This paper shows the design of a PC mouse commanded by electromyographic (EMG) signals from two muscles of the forearm, palmar longus and extensor digitorum, there were classified four hand movements (including rest): flexion, extension and closure. To fulfill this objective the problem was divided into two stages, the first consisting of the hardware design to allow the acquisition of EMG signal on the PC. The second stage was the design of the software that would classify EMG signals and thus give the mouse cursor movement. For the classification of the EMG signals two types of artificial neural network were used, the Multilayer Perceptron and Radial Basis Function networks. The creation of the dataset and performing tests were done with random healthy people with an age range between 20 and 25 years. The results were higher than 87% in tests conducted in real time, which shows that the system designed had good response rates in real time and a good level of success.