Detección de la intención de movimiento de extremidades superiores e inferiores a través del procesamiento de señales EEG

Abstract

This project arises from the need to satisfy the requirements of individuals who, have lost the ability to control their limbs and are in rehabilitation. The aim of this project is to develop an algorithm capable of identifying the intention of movement in the upper and lower extremities, starting from electroencephalographic signals (EEG). The acquisition of signals EEG is performed in subjects without apparent pathology and subjects with different pathologies that impede their mobility, for which the Emotiv EPOC+ acquisition device is used. A stage of conditioning and pre-processing of the signal is implemented, in order to improve the signal-to-noise ratio (SNR). Subsequently, frequency and time-frequency analyzes are applied by the fast Fourier transform (FFT) and the discrete wavelet transform (DWT), respectively; to extract characteristics related to the intention of movement. For the classification of signals, we experiment with artificial neural networks (ANN), fuzzy inference systems (FIS) and adaptive neuro-fuzzy inference systems (ANFIS). As a result of this project we have a data set with 40 records of upper extremities and 6 records of lower extremities, corresponding to 10 subjects (4 records per subject) and 3 subjects (2 records per subject), respectively. In addition, it is inferred that the characterization with DWT provides more information, which facilitates the recognition of patterns in the signals EEG. Finally, performance analysis yields 92.39 %, 86.52 %, 72.13 % and 94.73 % for the neural, fuzzy, neuro-fuzzy and parallel neural-fuzzy classifiers, respectively.