From printing to intelligence: Smart polymer composite actuators for next-generation soft robotics

Abstract

Smart polymer composite actuators are emerging as enabling materials for next-generation soft robotics by combining mechanical compliance, multifunctionality, and programmable responsiveness within lightweight architectures. Their ability to convert electrical, thermal, magnetic, optical, and chemical stimuli into controlled deformation provides advantages for adaptive interaction with complex environments. Advances in three-dimensional (3D) and four-dimensional (4D) printing have expanded their design space through multimaterial integration, spatially programmed architectures, functional gradients, and time-dependent shape transformation. Concurrently, machine learning (ML) offers new opportunities for material discovery, structure–property prediction, fabrication optimization, actuator modeling, inverse design, sensing, and intelligent control. This review critically examines the integration of polymer composites, 3D/4D printing, computational modeling, and ML across the materials-to-device development chain, from formulation and architecture to manufacturing, actuation, sensing, control, and lifecycle monitoring. Particular attention is given to printability–functionality trade-offs, multimaterial interfaces, manufacturing scalability and reproducibility, dataset limitations, uncertainty, simulation-to-reality transfer, and model generalization. Emerging strategies, including physics-informed ML, multiscale density functional theory–molecular dynamics–ML frameworks, digital twins, closed-loop design–print–test optimization, and sustainable composite systems, are discussed as pathways toward more predictive and efficient development. The review emphasizes that high ML accuracy does not necessarily ensure transferability or reliable physical performance and highlights the need for standardized validation and experimentally grounded models. Finally, opportunities for increasingly autonomous, self-adaptive, and closed-loop soft robotic systems are outlined, providing a roadmap toward reliable, scalable, sustainable, and application-ready intelligent soft robotic technologies.