EEG Multiclass Signal Classification Based on Subtractive Clustering-ANFIS and Wavelet Packet Decomposition

The use of driving navigation via a wheelchair joystick or electromyography (EMG) controller for people with physical disabilities is still a challenge. Brain-Computer Interface (BCI) can control and be an alternative solution. BCI uses electroencephalography (EEG) signals originating from the brain, where electrodes are placed on the scalp. In this study, classification was carried out using a Motor Imagery (MI) approach based on Wavelet Packet Decomposition (WPD) and Subtractive Clustering - Adaptive Neural Fuzzy Inferences System (SC-ANFIS) using a Muse headband. The results showed that SC-ANFIS could recognize patterns with an accuracy of 68% and F1 Score of 38%.