AI-empowered multimodal intelligent sensors integrated with decoupling and advanced functional materials

With the continued development of emerging technologies and devices such as flexible electronics, wearable systems, and electronic skins, multimodal intelligent sensors have become a key technological foundation for continuous health monitoring, human − machine interaction, and intelligent robotics. Compared with single-modal devices, multimodal intelligent sensors can simultaneously acquire multidimensional information, including mechanical, thermal, humidity, gaseous, and biochemical signals. However, their performance remains constrained by overlapping material response windows, signal crosstalk caused by structural coupling, and the difficulty of decoding complex datasets. Herein, recent progress toward high-performance multimodal intelligent sensors is summarized, discussed, and evaluated through a structured narrative-review framework centered on advanced material design, perception-decoupling strategies, artificial intelligence-driven data analysis, and system deployment. First, the major transduction pathways for pressure, strain, temperature, humidity, gas, and biochemical signals are outlined, and the design principles of sensor devices based on hydrogels, carbon-based and two-dimensional composites, Janus heterogeneous structures, and biomimetic textiles are summarized, together with their advantages in flexibility, conductivity, interfacial regulation, and multifunctional integration. Next, key decoupling strategies are discussed in depth, including structural spatial decoupling, orthogonal responses of functional materials, microstructure and interface engineering, the construction of independent signal channels, and feature-representation-based assisted unmixing, thereby clarifying the major routes toward low-crosstalk and high-fidelity perception under concurrent multi-stimulus conditions. Furthermore, this review examines the role of data-driven models in multimodal signal recognition, fusion-based decision-making, and scenario understanding. By linking these advances with applications in wearable health monitoring, electronic skins, and human–machine interaction, this review highlights the evolution of multimodal systems from device-level sensing toward system-level cognition. Finally, the major bottlenecks in the current field are identified, including material stability, decoupling capability, sensing accuracy, long-term reliability, system integration, and scalable manufacturing, and future perspectives are provided for next-generation multimodal sensing platforms featuring high selectivity, self-powering capability, manufacturability, interpretability, and edge-intelligence-enabled collaboration. This review aims to provide a cross-layer analytical framework and practical design guidelines for material selection, structural engineering, algorithm configuration, and system-level applications of high-performance multimodal intelligent sensors.