Over the past decade, soft robotics has evolved from a niche curiosity into a robust interdisciplinary domain, attracting engineers, material scientists, and computer scientists alike. Unlike rigid-bodied machines, soft robots exploit compliant structures and novel materials to navigate uneven terrain, adapt to variable tasks, and interact safely with humans. Applications span rehabilitation gloves, untethered terrain explorers, manipulators, and diverse grippers. The roots trace back to pneumatic artificial muscles of the 1950s, but momentum accelerated after 2008, with early reviews such as Trivedi et al. differentiating soft robots from hyper-redundant rigid systems.
A scientometric analysis of 10,504 publications from 2010 to mid-2021, drawn from the Web of Science Core Collection, reveals both the structural backbone and emerging trends of the field. Using CiteSpace to construct co-citation networks, the study identifies intellectual clusters, citation bursts, and thematic shifts. Publication growth was steep between 2012 and 2019, surpassing 1,000 annual papers in 2017. China leads in total output (2,966 papers), followed closely by the United States (2,911), though the U.S. ranks highest in average citations per publication. International collaboration accounts for over a quarter of top-country output and correlates with higher impact.
Engineering dominates the subject areas, with robotics, electrical engineering, and automation at the forefront. Chemistry follows, reflecting the importance of advanced materials—nanoscience, polymer science, and physical chemistry—in enabling new actuators and sensors. Journals such as IEEE Robotics and Automation Letters and Soft Robotics anchor the literature, while high-impact materials journals like Advanced Materials contribute substantially.
CiteSpace clustering highlights key research fronts. Cluster #0, “variable stiffness,” contains landmark reviews including Rus and Tolley’s 2015 synthesis, alongside work on pure soft robots and rapid pneumatic actuators. Cluster #1 centers on stretchable strain sensors, with Zhao et al.’s optical waveguides for prosthetic hands and Larson et al.’s electroluminescent skins as notable milestones. Cluster #2 focuses on multiresponsive actuators, particularly hydrogels capable of large, reversible deformations under stimuli such as temperature and pH. Younger clusters point to emergent technologies: Cluster #5 on 4D printing, Cluster #6 on liquid crystal elastomers (LCEs), and Cluster #7 on dielectric elastomer actuators (DEAs).
Citation burst analysis underscores foundational work in soft actuators and manipulators, such as Shepherd et al.’s “multigait soft robot” and Ilievski et al.’s pneu-net actuators. More recent bursts highlight magnetically actuated millimeter-scale robots (Hu et al., 2018) and 3D-printed ferromagnetic composites (Kim et al., 2019), signaling converging trends in new materials, advanced fabrication, and integrated intelligence.
Material innovation is a primary driver. Soft magnetic composites—polymers embedded with magnetizable particles—enable wireless control and deep-field actuation for applications like drug delivery. LCEs combine polymer networks with liquid crystal mesogens to produce thermally and optically driven motion, allowing multimodal locomotion. Hydrogels offer ionic conductivity, transparency, and self-healing, making them candidates for flexible sensors and actuators. DEAs deliver rapid, lightweight actuation via electrostatic membrane compression.
Fabrication advances, particularly direct 3D and 4D printing, are transforming prototyping. By combining stimuli-responsive materials with additive manufacturing, researchers produce robots that morph under heat, light, or magnetic fields. Examples include hydrogel-based grippers, ferromagnetic continuum robots, and multi-material prints integrating sensors. Techniques range from direct ink writing to multi-nozzle deposition, enabling complex architectures and embedded functionality.
Sensing and intelligence form the third major trend. Soft robots’ continuous elastic bodies demand novel proprioceptive and exteroceptive systems. Multimodal skins, integrating strain, pressure, temperature, and humidity sensing, mimic biological capabilities. Materials such as supramolecular hydrogels and ultra-thin silicon nanoribbons underpin these systems. Data-driven models, from regression to deep neural networks, process sensor input for adaptive control. Examples include LSTM-based kinematic modeling of soft actuators and CNN-enabled tactile object recognition.
The scientometric mapping illustrates a field expanding in scope and depth. New materials unlock actuation and sensing possibilities; advanced printing methods accelerate integration; and intelligent sensing systems push adaptability toward biological sophistication. Together, these trajectories define the current frontier of soft robotics and chart pathways for its continued evolution.