Large-Scale Nanofiber Manufacturing|

Prof. Kun Dai (Zhengzhou University): Multilayer Bioinspired Tunable Strain Sensor for Machine Learning-Assisted Gesture Recognition

Challenge: While flexible strain sensors capable of detecting subtle mechanical signals and conforming to irregular surfaces are increasingly vital for physiological monitoring, soft robotics, and human-machine interaction, their development has been constrained by the inherent trade-off between sensitivity and detectable signal range.

Approach: Inspired by the multilayer architecture of natural nacre, Prof. Kun Dai's team fabricated a carbon nanotube (CNT)/graphene (GR)/thermoplastic polyurethane (TPU) mat (CGGTM) through integrated electrospinning and high-pressure spraying technologies.

Innovation 1: The designed multilayer structure with mutually non-interfering conductive networks endows CGGTM with:
• Ultra-low detection limit (0.025% strain)
• High sensitivity (GF=14.85 at 130% strain)
• Wide detection range (0-130% strain)
• Fast response/recovery (180ms/210ms)
• Exceptional cyclic durability (>10,000 cycles)

Innovation 2: When assembled as triboelectric nanogenerators, CGGTM demonstrates:
• Stable triboelectric output (Voc≈45V)
• Successful integration with ML algorithms for:

• Biometric signal acquisition

• Multi-gesture motion recognition (98.7% accuracy)