BRAIN COMPUTER INTERFACE BASED ON EEG SIGNALS FOR MOTION CONTROL OF HAND PROSTHESIS USING ANFIS

A large number of people in the world who have amputated limbs that are usually replaced by mechanical prostheses. Moreover, electromechanical prostheses have been gaining strength and are supported by different types of interfaces as brain interfaces computer that improve the functionality of these, despite show representative results for the control of prostheses, it is still an open field research that seeks to improve its effectiveness and efficiency. In this paper, we present a methodology for classification of electroencephalographic signals (EEG) to control movement of a prosthetic hand, based on the adaptive neurofuzzy inference system (ANFIS) applied to features derived from the wavelet transform is presented (TW) and the rough fuzzy sets (FRS) to EEG signals obtained in 10-10 system. Thus the performance of the proposed system was measured using cross-validation with 30 repetitions 70-30 obtaining high performance in terms of accuracy, which means that this model has potential as a classifier for the detection of EEG changes for the command generation for the control of hand movement.