EEG-Based Control of a 3D-Printed Upper Limb Exoskeleton for Stroke Rehabilitation

Abstract

Brain-computer interfaces (BCIs) have emerged as transformative tools for translating users’ neural signals into commands for external devices. The urgent need for innovative treatments to enhance upper limb motor function in stroke survivors is underscored by the limitations of traditional rehabilitation methods. The development of communication and control technology for individuals with severe neuromuscular diseases, particularly stroke patients, is centered on utilizing electroencephalographic (EEG) signals to accurately decode users’ intentions and operate external devices. Two healthy subjects and a stroke patient were enrolled to acquire EEG signals using the EMOTIV EPOC+ sensor. The experimental procedure involved recording five actions for both motor imagery and facial expression signals to control the 3D-printed upper limb exoskeleton. EEGLAB and BCILAB software were used for preprocessing and classification. The results showed successful EEG-based control of the exoskeleton, representing a significant advancement in assistive technology for individuals with motor impairments. The support vector machine (SVM) classifier achieved higher accuracy in both offline and online modes for both motor imaginary and facial expression tasks. The conclusion highlights the appropriateness of using EEGLAB for offline EEG data analysis and BCILAB for both offline and online analysis and classification. The integration of servo motors in the exoskeleton, allowing movements in five Degrees of Freedom (DOF), positions it as an effective rehabilitation solution for individuals with upper limb impairments.