Robotic AI-based Fiber Threading

This paper presents our research on robotic AI-fiber threading, offering a comprehensive overview of our findings. We introduce a novel robotic device and innovative approaches for fiber threading, enabling precise fiber pulling with real-time force measurement during both pulling and tensile testing. Our proposed approaches exhibit enhanced consistency, as well as superior strength and toughness when compared to traditional dry and wet spinning approaches. We explore fiber threading across a range of materials with varying viscosities, including high-viscosity materials like solvent-based polychloroprene rubber, and low viscosity materials such as
dextran, recombinant spider silk, and regenerated silk fibroin. Additionally, we introduce a novel approach for rapidly improving artificial fiber tensile properties using our robotic fiber-pulling device and Bayesian Optimization (BO) algorithm.