Enhancing Hand Gesture Recognition for Varied Arm Positions Through the Integration of Co-Located sEMG-pFMG Armband
Shen Zhang, Hao Zhou, Rayane Tchantchane, Gürsel Alıcı · 2024
This study introduces a new approach to hand gesture recognition for various limb positions by employing a 2-pair co-located surface electromyography and pressure-based force myography (sEMG-pFMG) armband. The research addresses a crucial gap in the existing literature, where the majority of studies focus on utilizing multi-modal systems to enhance hand gesture recognition performance under static arm positions. We pioneer the exploration of the co-located sEMG-pFMG sensing system under different arm positions, utilizing a dynamic arm position dataset to train the selected machine learning (ML) models. Notably, our system achieves promising results, showing a significant 6.5% accuracy increase compared to sole sEMG and an impressive 28.7% accuracy increase compared to sole pFMG, using only two of sEMG-pFMG sensing units. This notable enhancement in accuracy underscores the potential of our approach in overcoming the challenges posed by varying arm positions under real-life circumstances.