Speed Classification of Upper Limb Movements Through EEG Signal for BCI Application

Brain-Computer Interface (BCI) systems have obtained remarkable results in rehabilitation and robot control processes by converting brain signals into commands. The quantity of movement speed is the fundamental issue BCI that requires additional research. This paper investigated classification slow fast speeds eight different upper limb movements through electroencephalogram (EEG) information about values maximum angle from MPU6050 module. Datasets were recording EEG 10 subjects module connected on their right hand during movements. study used Filter Bank Common Spatial Pattern (FBCSP) Wavelet-Common (W-CSP) methods to extract features In both methods, selected Mutual Information (MI) sent Convolution Neural Network (CNN) various machine learning classifiers. Due subject-independent classification, FBCSP-CNN method highest accuracy 90% with a Kappa coefficient 0.8 for flexion/extension shoulder. Results our proposed demonstrate ability introduce refined set commands system recognizing associated parameters.