Answers are grounded in this paper and cite their sources.
Soft pneumatic actuators are widely used in soft robotics because of their compliance and ability to generate smooth, biomimetic motions. However, accurately modeling their highly nonlinear and deformable behavior, particularly under contact conditions, remains computationally intensive, limiting their integration into real-time control and digital twin frameworks. This study presents a reduced-order surrogate modeling approach to enable a real-time-capable digital twin for soft pneumatic bending actuators. High-fidelity simulations based on finite element modeling (FEM) are first used to generate full-field response snapshots, which are reduced using proper orthogonal decomposition; radial basis function interpolation is then trained to map input parameters to the reduced coefficients, enabling rapid reconstruction of the full-field actuator response. The reconstructed responses are compared with FEM simulations for error analysis and validated against experiments on a custom-fabricated pneumatic network actuator. Results demonstrate that, within the sampled operating domain, the proposed method enables accurate, near real-time (calculation time < 1 s) predictions of bending deformations, spatial stress and displacement distributions, and tip contact forces, thereby realizing a digital twin of the actuator. Furthermore, a digital twin of a multi-actuator grasping system is demonstrated, highlighting the effectiveness of the method in creating efficient digital twins for soft robotic systems with strong potential for closed-loop control, performance monitoring, and decision-making.
No takes yet. Share an insight, caveat, or question.
Pu et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: