Abstract
In today’s world, catastrophic accidents have caused many people to lose limbs, while rapid technological advancements continue to progress. As a result, this creates a pressing need to bridge the gap between human needs and technological progress. This study focuses on the prosthetic limb user interface. This interface serves as the crucial link between the human body and the machine. The prosthetic limb's ability to replicate the original limb's function depends largely on the quality of this interface. Accordingly, this paper presents a comparative analysis of Electroencephalography (EEG) and Electromyography (EMG) in terms of performance, accuracy, sensitivity, and response time. EMG interfaces capture bioelectrical signals from residual muscles and are suitable for individuals with remaining functional musculature. On the other hand, EEG interfaces record electrical activity from the brain and are applied in cases of complete paralysis, where cortical signals are the only viable control source. We developed mathematical models for both interfaces and integrated them with machine learning techniques to enhance classification. The Support Vector Machine (SVM) algorithm, recognized for its efficiency in classification and regression, was employed as the primary classifier. The results of our MATLAB simulation, conducted in a controlled environment using synthetic signals, demonstrated the functionality and effectiveness of the approach. It is crucial to note that these high accuracy values were achieved under these idealized conditions, and performance may vary with real-world data. The simulation results demonstrated a consistent and statistically significant performance advantage for EMG- based control over EEG-based control in terms of classification accuracy and reliability, which is attributed to fundamental differences in signal physiology. This performance gap is primarily attributed to differences in physiological signals. This study aims to provide a technical reference for selecting the appropriate interface based on patient condition and contributes to guiding the future development of more intelligent prosthetic limbs.