EEG-EMG based bio-robotics elbow orthotics control

Abstract

Brain-computer interface (BCI) or also its advancement, hybrid brain-computer interface (hBCI), is a technology that is vastly developed. This technology has been used in many fields. BCI is a system that directly changes a human’s mind into data that can be extracted to information that can be meaningful to people. The development of this technology has applications as a rehabilitation aid for someone suffering from an inability to move his limbs, such as the arms. Through this research, it is hoped to be able to design an orthosis control system as a rehabilitation device by using a classification method with EEG and EMG signals, so that subjects who use this tool can carry out rehabilitation in upper arm movements especially in the elbow joint. The system utilized Raspberry Pi 3 B+ as the computer and ADS1299EEG-FE as analog front end for EEG and EMG. EEG frequency band power and EMG Vrms feature are extracted using Wavelet Transform and the model used for movement classification is Support Vector Machine. The results of the movement classification using both signals, using delta alpha ratio and root mean square features, obtained training accuracy for three movements namely relax, flexion, and extension of 90.3% and for testing accuracy of 85.2%. The combination of EEG and EMG signals are considered a promising approach for developing rehabilitation device of right arm limb.