Hand rehabilitation training system integrating non-contact and contact triboelectric nanogenerators for enhanced gesture and handwriting recognition

The human hand is one of the most adaptable and versatile organs due to its complex anatomy and functionality. However, this very adaptability makes the hand highly susceptible to injury, highlighting the need for effective hand rehabilitation programs. Current rehabilitation methods are often limited by location and lack of personalized approaches, necessitating significant improvement. In this study, a fun and engaging hand rehabilitation training game is developed. A gesture recognition sensor based on non-contact triboelectric nanogenerator is designed to enhance the overall coordination and strength of the arm, wrist, and hand. Additionally, a handwriting signal recognition sensor based on contact triboelectric nanogenerator is designed to strengthen and improve finger coordination. The gesture recognition sensor, integrated with deep learning algorithms, accurately identifies six directional movements with 97.33 % accuracy, while the handwriting signal recognition sensor successfully identifies 26 uppercase English letters with 99.5 % accuracy. Utilizing these sensors, a game simulating a supermarket purchase scenario is created, providing a flexible and convenient approach to hand rehabilitation. This system offers a potential solution to improve the design of hand rehabilitation products, making the training process more enjoyable and accessible. [Display omitted] •Design of a gesture recognition sensor utilizing a non-contact TENG for detecting hand movements.•Development of a handwriting signal recognition sensor based on a contact TENG for precise character identification.•Application of deep learning algorithms for recognition of six directional movements and 26 uppercase English letters.•Creation of a game-based rehabilitation training solution designed for hand therapy across different environments.