Classificador Neural para Intenção de Movimento do Braço e Antebraço via Extreme Learning Machine

: The myoelectric signals are electrical potentials that represent the dynamics of muscle contrac-tion and its study has shown to be relevant in applications in physiotherapy and rehabilitation engine for the improvement of the quality of life of amputated individuals or with some type of motor deficiency. The article proposes a neural classifier that consists of two feedforward neural networks in parallel, con-structed with the aid of the MATLAB ® computational tool to identify the movement of the arm and fore-arm, being able to be used as a command source for myoelectric prostheses and robotic arms. Numerical simulations from real data prove the efficacy of the developed neural classifier. a partir de dados reais comprovam a efi-cácia do classificador desenvolvido.