High-performance piezoelectric yarns for artificial intelligence-enabled wearable sensing and classification

Abstract

Piezoelectric polymer fibers offer a fundamental element in intelligent fabrics with their shape adaptability and energy-conversion capability for wearable activity and health monitoring applications. Nonetheless, realizing high-performance smart polymer fibers faces a technical challenge due to the relatively low piezoelectric performance. Here, we demonstrate high-performance piezoelectric yarns simultaneously equipped with structural robustness and mechanical flexibility. The key to substantially enhanced piezoelectric performance is promoting the electroactive beta-phase formation during electrospinning via adding an adequate amount of barium titanate (BaTiO3) nanoparticles into the poly(vinylidene fluoride-trifluoroethylene) (P(VDF-TrFE)). When transformed into a yarn structure by twisting the electrospun mats, the BaTiO3-doped P(VDF-TrFE) fibers become mechanically strengthened with significantly improved elastic modulus and ductility. Owing to the tailored convolution neural network algorithms architected for classification, the as-developed BaTiO3-doped piezo-yarn device woven into a cotton fabric exhibits monitoring and identifying capabilities for body signals during seven human motion activities with a high accuracy of 99.6%.