A Single‐Channel Humidity‐Pressure Dual‐Modal Decoupling Sensor for Assistive Geriatric Rehabilitation

ABSTRACT Against the background of global population aging and the growing need for elderly care, developing multi‐modal sensing platforms with high interference immunity, specifically for complex respiratory dysfunctions, has become a critical demand in the field of precision medicine. However, conventional multimodal sensors suffer from performance limitations and signal crosstalk due to excessive reliance on complex material synthesis and suboptimal structural design. A humidity–pressure dual‐modal sensor (HPDS) was designed by integrating polyimide, electrodes, and a heterogeneous conductive elastomer into a sandwich‐structured device. The HPDS exhibited good sensitivity over a wide pressure range (0.58–392 kPa), high mechanical stability exceeding 50 000 cycles, and a fast humidity response of 200 ms. An intelligent assistive training scheme combined with a lightweight one‐dimensional convolutional neural network (1D‐CNN) was further constructed for respiratory pattern discrimination and physical therapy assistance, achieving an accuracy of 98.9% in classifying the six core respiratory states. This work provided a universal design paradigm for next‐generation low‐coupling and high‐performance AI devices, showing great promise in the development of intelligent healthcare and home‐based rehabilitation.