Fluorescent Hydrogel‐Based Strain Sensor With Machine Learning‐Augmented Performance

Hydrogels are ideal matrices for bio‐integrated wearable sensors due to their biocompatibility and skin‐like mechanical properties. Compared with conventional electronic strain sensors, optical strain sensors offer electrode‐free operation, resistance to electromagnetic interference, and direct real‐time visual readouts. However, simultaneously achieving mechanical robustness and high optical sensitivity in hydrogel‐based strain sensors remains challenging. Herein, a new fluorescent hydrogel has been developed by homogeneously integrating carbon quantum dots (CQDs) into a poly(2‐hydroxyethyl methacrylate‐co‐2‐aminoethyl methacrylate) (p(HEMA‐co‐AEMA)) network. The effects of CQD concentration on the hydrogel's Young's modulus and fluorescence intensity have been systematically investigated. The results reveal a strongly non‐monotonic dependence on CQD concentration; that is, Young's modulus peaks at 0.1–0.2 mg mL ⁻¹ and decreases thereafter; fluorescence intensity reaches a maximum at 0.5–0.6 mg mL ⁻¹ , while higher concentrations induce inner‐filter effects increasing fracture propensity. A machine‐learning (ML) framework based on a Random Forest model with leave‐one‐out validation was employed to capture the nonlinear, non‐monotonic relationships between CQD loading, mechanical properties, and optical response. As a proof of concept, the fluorescent hydrogel‐based strain sensor is demonstrated for real‐time finger‐bending monitoring. This work highlights ML‐assisted design of mechanically flexible, user‐friendly fluorescent hydrogel strain sensors for wearable applications.