High-Accuracy Gesture Classification Using Biocompatible, Low Impedance Dry Electrodes

Epidermal electronics match the mechanical properties of human skin, enhancing the coupling between electronics and biological tissues. Flexible functional materials and additive fabrication methods are used in developing dry electrodes for skin applications to enhance wearable devices for monitoring muscle activity and electrophysiological signals. Maintaining a consistently low impedance electrode-skin interface is crucial. Traditional wet electrodes use hydrogels and skin abrasion, which are impractical for self-administration and in-home care. These devices utilize machine-learning models for gesture classification and other monitoring tasks, achieving high accuracy with adaptive learning capabilities. A wearable surface electromyography (sEMG) biosensing system based on a screen-printed conformal electrode array implements a neuro-inspired hyperdimensional computing algorithm for real-time gesture classification and model training, achieving 97.12% accuracy for 13 hand gestures and 92.87% for 21 gestures, with adaptive updates enhancing performance under variable conditions. Using electroless copper and gold plating with 3D printing, biocompatible, low impedance dry electrodes that do not require hydrogels were demonstrated. These electrodes, tested alongside clinical-standard devices, show an average electrode-skin impedance of 66.7 kΩ at 50 Hz, comparable to wet electrodes. Strategies that use blown aerosol chemical vapor deposition to produce free-standing, conformal carbon nanotube (CNT) dry films will also be discussed.