Rehabilitador de tobillo mediante identificación de señales electromiográficas clasificadas con Redes Neuronales.

Abstract

Totype to regain the mobility of the muscles around the area that involves the ankle, methodologically it is quantitative in type with a quasi-experimental design. The goal is to implement structural mechanical support and create recognition patterns so that you can identify fundamental ankle movements for muscle rehabilitation using artificial neural networks (RNAs). It is aimed at people with injuries in the ankle area, the same who were part of an activated rehabilitation at ages between 15 and 30 years, with a maximum BMI of 26. A progressive training of an RNA was used, from electrical impulses generated by muscular activity, using a team of electromyographic sensors. To validate the information of the signals present in the recovery process, movement patterns were used in an RNA-based identification architecture (multilayer three-line Madeline with red feedback). A graphic interface was considered that facilitated the visualization of the movements and the result of the red neuronal; with a hysteresis band of 0.15 - 5% movement identification error: eversion, dorsiflexion and inversion