A piezoelectric micro-pyramid array sensor based on PVDF/WS2 nanosheets for high-precision deep learning-assisted pattern recognition

The piezoelectric pressure sensors, which have emerged as core components of intelligent sensing systems, exhibit significant application values in flexible electronics and human-machine interaction. In this study, we innovatively developed a micro-pyramid array sensor (MPAS) based on PVDF/WS 2 nanosheets (NSs) nanocomposites. The PVDF/WS 2 NSs nanocomposites were prepared by using the solid-state shear milling (S 3 M) technology and then elaborately integrated into a precisely designed micro-pyramidal architecture through microinjection molding (μIM) strategy. Under the very strong shear stress fields of μIM process, the spontaneous alignment of PVDF molecular chains and WS 2 NSs along the flow direction has been achieved. The fabricated sensing microdevices exhibit high sensitivity (128 mV/kPa) within a 10 N pressure range, along with 1 mm spatial resolution comparable to the level of human tactile perception. Furthermore, based on fabrication of the above microdevice, we also established an intelligent recognition system through combining multi-stage signal processing algorithms with a convolutional neural network (CNN). The resulting MPAS could achieve 100 % classification accuracy for three English letters. By setting an intensity threshold for the sensor arrays, the constructed recognition system could achieve 2 37 (≈1.37 × 10 11 ) password combinations, which is equivalent to the security level of an 11-digit decimal password lock system. Such an interesting and smart approach has demonstrated a unique security authentication methodology for intelligent electronic devices, and exhibits the promising application prospects in artificial intelligence-based precision recognition systems and advanced human-machine interaction interfaces.