Precision in Action: Harnessing Non-Redundant EMG Signals for HCI Evolution

In the ever-evolving landscape of Human-Computer Interaction (HCI), the need for more efficient and user-friendly interfaces has become increasingly apparent. This study is dedicated to enhancing the recognition of hand-based Electromyography (EMG) signals in HCI scenarios by addressing interference issues and optimizing data processing. Our approach centers on identifying redundant electrodes in hand gesture recognition using EMG signals, a vital step in improving the overall efficiency of EMG signal equipment. We introduce a variance-based method for categorizing redundancy levels across five distinct action signal groups into 16 levels, with a visual representation for clarity. By rigorously evaluating hand motion recognition results under various configurations of redundant channels, we pinpoint the optimal combination of channels for precise recognition. Furthermore, we select and compare two robust classifiers for EMG signals, analyzing their classification results to determine the most effective hand motion pattern recognition approach. This research underscores the significance of HCI in modern technology and presents a method to harness non-redundant EMG signals, thus contributing to the evolution of human-computer interaction. By reducing interference and enhancing data processing, we aim to pave the way for more precise and efficient interactions between humans and technology.