Multivariate EMG Signal Based Automated Hand Gestures Recognition Framework for Elder Care

Electromyogram (EMG) signals obtained from muscles can provide insights into the biomechanics of human movement. EMG technology finds diverse applications including enhancing human–computer interaction, enabling muscle-controlled devices for hand gesture recognition, facilitating prosthetic control for individuals with disabilities and elder care. Hand gestures are crucial for human–computer interaction, bridging the gap between human intent and machine control. Their significance has obtained considerable attention, leading to the development of advanced detection systems. These systems facilitate effective interaction between humans and computers, thereby enhancing various applications across different domains. Current research on EMG-based hand gesture classification encounters challenges such as inaccurate classification, and limited generalization ability. To encounter these problems, an automated multi-class hand gestures identification model is proposed via machine intelligence. A publicly accessible UCI2019 EMG dataset obtained from 8-channels MYO thalmic bracelet for surface EMG data acquisition is used to demonstrate the work. Initially, the multivariate EMG channels data are pre-processed and then fed to machine learning classifiers. Six classifiers are evaluated for the proposed predictive model, with ensemble bagged tree (EBT) consistently outperforming (overall highest accuracy of 98.4%) than other classification approaches. The superior performance of EBT classifier in overall classification and class wise classification are exhibited through receiver operating characteristic (ROC) analysis. Hence, this computer-assisted decision-making model minimizes the limitations of prior studies by offering gesture based human–machine interaction and smart device control for elder care. The proposed research can also offer valuable contributions to manufacturing by facilitating tasks in the industry such as remote operation, quality control, and maintenance by enabling hands-free control and reducing the physical strain on workers.