Classification of EEG Signals Using a Common Spatial Pattern Based Motor-Imagery for a Lower-limb Rehabilitation Exoskeleton

This article presents a lower-limb exoskeleton rehabilitation integrated with an electroencephalogram (EEG) based brain-computer interface (BCI). The BCI is used to record signals collected using the EEG interface during resting, a motor imagery (MI) and a motor execution (ME) tasks. The MI data is generated when a subject imagines the movement of a limb. Therefore, the signals were characterized by Common Spatial Pattern (CSP), Power Spectrum Density (PSD) and Discrete Wavelet Transform (DWT) in the specific Mu band of Auto Regressive model (AR). In the case of the lower-limb representation, there is a problem of reliably distinguishing leg movement intentions. The study shows how the combined use of multi-model signals can improve the accuracy and reliability of the human-machine interface. The signals induced by CSP, PSD, and DWT+AR are used for the lower-limb exoskeleton control commands to drive the movement in real time. With using the support vector machine (SVM), the signals are tested the classification precision. The classification of the control system permits one to drive the lower limb rehabilitation exoskeleton effectively.