Extracción de características y métodos de clasificación para reconocimiento de movimientos de mano a partir de señales de EMG y EEG: Revisión

Abstract

Today, there are various methods for fabricating upper limb or hand prostheses, from simple signal processing techniques to more complex neural network systems. In general, the process is based on stages of data acquisition, signal processing, extraction and classification of signal characteristics, and prediction functions. Each of the variants depends on the design parameters, such as manufacturing cost and percentage of assertiveness. In low-cost systems, fewer captures of the biosignals are used, resulting in simpler signal processing, obtaining a 75% assertiveness percentage. In systems that require greater assertiveness, up to 99.5%, the number of sampling channels must be increased, which requires more complex processing involving neural networks. In this article, several of the techniques used in this process are exposed, which are obtained from the review of the state of the art of the last 10 years.