Optimal Fuzzy Logic Enabled EEG Motor Imagery Classification for Brain Computer Interface

Abstract

Brain-computer interface BCI) is a technology that assists in straight link among the human brain as well as external devices like computers or robotic systems, without including muscles and peripheral nerves. BCI allows individuals with motor disabilities to manage external devices with the aid of brain signals such as motor imagery detected from electroencephalography (EEG) signals. An EEG Motor Imagery Classification for BCI is a specific application of EEG in which brain signals directly related to motor imagery tasks are analyzed and classified to control external devices or applications, namely robotic systems or computers. In this regard, the study introduces a Jellyfish Optimization with Fuzzy Logic Enabled EEG Motor Imagery Classification for Brain Computer Interface (JFOFL-MICBCI) technique. The JFOFL-MICBCI technique aims to exploit the fuzzy logic system with metaheuristics for classifying EEC motor imagery signals. It initially executes Continuous Wavelet Transform (CWT) for transforming 1D-EEG signals into 2D time-frequency amplitude ones. For feature extraction, the JFOFL-MICBCI technique uses the SqueezeNet method, and its hyperparameters can be adjusted by the employ of the JFO system. The JFOFL-MICBCI method exploits the adaptive neuro-fuzzy inference system (ANFIS) approach for performing the classification process. A comprehensive range of experiments has been accompanied to demonstrate the higher efficiency of the JFOFL-MICBCI technique. The obtained results inferred the better of the JFOFL-MICBCI technique with other recent systems.

Keywords

• EEG signals
• human–computer interaction
• fuzzy logic
• brain–computer interface
• metaheuristics