ABSTRACT Skin‐interfaced electronic devices based on impermeable substrates and stacked architectures often suffer from poor sweat permeability and severe mechanical mismatch between layers, presenting significant barriers to long‐term wearing comfort and stability. To address this, we have developed a multifunctional Janus structured strain sensor based on thermoplastic polyurethane/MXene/carbon nanotubes with a stimulus‐responsive 1H, 1H, 2H, 2H‐perfluorooctyltrimethoxysilane‐titanium dioxide exterior. The asymmetric architecture facilitates reversible surface wettability switching through light or thermal stimulation, enabling dynamic environmental adaptation, which includes efficient sweat removal under UV exposure and reliable water repellency after heat‐induced recovery of hydrophobicity. The sensor exhibits excellent overall sensing performance, with a strain detection range of 0%–150%, a gauge factor of 199.685, and maintains stability over 8000 consecutive testing cycles. Molecular dynamics simulations and micro‐structure analysis elucidate the sensing mechanism of the interaction between force‐sensitive materials and the conductive network. Integrated with a deep learning algorithm, the sensor achieves 97.53% accuracy in sign language recognition. To validate its practical potential, a smart glove system integrating the sensor is designed to wirelessly control a robotic car and provide real‐time voice feedback, demonstrating its potential for wearable human–machine interaction systems.
Chu et al. (Fri,) studied this question.