A Coral-Inspired Dual Modal Hydrogel Sensor with Deep Learning-Assisted Decoupling of Force-Thermal Stimuli

Multimodal perception is at the forefront of flexible electronics, yet its core challenge lies in signal crosstalk induced by multiple environmental stimuli. Inspired by the integrated perception capability of stony corals to thermal, force, and other environmental stimuli, this study fabricated a multi-stimuli-responsive visualizable hydrogel sensor (PCSC hydrogel). It synergistically integrates the thermochromic unit (SP) and the force-induced resistance response unit (CQDs) within a PVA/CMCS dual network. It not only possesses thermochromic properties and a sensitive force-induced resistance response (gauge factor = 4.14) simultaneously but also exhibits a significant difference in the response times between its force-induced and thermal-induced resistance responses (0.4 and 7.1 s, respectively). On the basis of the multimodal signals inherently output by the material, we constructed a deep learning processing framework that fuses image and resistance signals. Utilizing a lightweight ResNet network to extract spatial features from thermochromic images and an multilayer perceptron to analyze resistance time series signals, we achieved not only temperature recognition (97.5% accuracy) and gesture recognition (98.5% accuracy) but also the successful real-time synchronous decoupling and identification of concurrently applied force and thermal stimuli, with an overall accuracy exceeding 95%. This work provides a new strategy for developing a new generation of multimodal flexible perception technologies capable of adapting to complex environmental interactions.