Breathable all-textile pressure sensor with conductivity-modulable polypyrrole for deep learning-enhanced sensing
Deep learning-enhanced pressure sensors that integrate signal processing with sensing capabilities offer transformative potential for wearable electronics. However, current implementations predominantly rely on petroleum-based polymers for sensing/encapsulating layers and metallic electrodes, resulting in limited biodegradability, poor biocompatibility, and insufficient breathability. This work presents an all-textile pressure sensor that combines conductivity-modulable polypyrrole (PPy) textiles for both electrode and sensing layers with real-time artificial intelligence algorithms. Eliminating metallic electrodes and petroleum-based polymers yields a device with excellent biocompatibility, biodegradability, and breathability. The textile sensing layer's structure ensures pressure-induced conductivity, contributing to high sensitivity and a wide detection range. The integrated deep learning model, a one-dimensional convolutional neural network (1D-CNN), achieves 99.6% classification accuracy on human motion datasets after 16 training epochs. Under Gaussian noise with standard deviations of 150 and 200, accuracy remains at 97.3% and 93.8%, respectively. Spraying 0.1 mL water on sensor surfaces yields 98.6% accuracy, demonstrating robustness to environmental disturbances. The system enables health monitoring, software/hardware control, and complex human motion analysis. These results confirm that the deep learning-enhanced fabric sensor can achieve accurate real-time human motion recognition, showing potential for immersive motion capture and intelligent feedback systems. This work provides a sustainable, breathable, and biocompatible platform for next-generation smart textiles.