Wireless smart gloves with ultra-stable and all-recyclable liquid metal-based sensing fibers for hand gesture recognition

Smart gloves are being extensively studied as supplementary solution for vision and voice interaction interfaces. Despite extensive efforts, smart gloves usually exhibit an obvious trade-off between functionality, performance and cost because of the limitations of static and dynamic complex hand gesture recognition capability and accuracy, hindering their application prospects. Here, we realize an intrinsically all-recyclable, ultra-stretchable, highly compliant and scalable resistive sensing fiber based on liquid metal and thermoplastic materials. The sensing fibers enables high skin compliance, resulting in highly accurate static and dynamic hand gesture signals and reduces the design complexity and cost of the subsequent integrated system. The study of the interface behavior between liquid metal core and SBS shell contributes to further understanding of individual roles that cohesive and adhesive forces play in the electromechanical stability of other liquid metal-based fiber electronics. Given the ultra-stable electrical properties of the sensing fiber (-10,000 cycles remain stable and regular) and the ability to continuously capture real-time static and dynamic somatosensory signals, we design a wireless smart glove-based interactive interfaces with cost-effectiveness, high integration, broad adaptability and crosstalk-free recognition ability of both static and dynamic hand gestures. Combined with machine learning and self-adaptive algorithms, the successful construction of three static and dynamic hand gesture recognition systems (11 hand gestures with a recognition accuracy of -93.6 %) demonstrates the potential of the proposed wireless smart gloves to have widespread practical applications prospects.

Automated summary

Smart gloves are being researched as additional solutions for vision and voice interaction interfaces. However, there is a trade-off between functionality, performance, and cost due to limitations in hand gesture recognition capability. A recyclable and stretchable sensing fiber based on liquid metal and thermoplastics has been developed, enabling highly accurate hand gesture signals and reducing design complexity. Combined with machine learning, three hand gesture recognition systems have been successfully constructed, demonstrating the potential of wireless smart gloves for widespread practical applications.