Development and Validation of Computational Modeling Framework for Tendon-based Soft Robotics

Tendon-driven soft robotic systems are extensively recognized for their intrinsic flexibility, exceptional dexterity, potential for safe human interaction, and adaptability to unstructured environments. Modeling such systems is challenging because of their highly nonlinear continuum structure, coupled with complex tendon interactions, and friction. Traditionally, modeling has been dominated by continuum Cosserat rods. While mathematically robust, the underlying assumptions of the Cosserat model restrict its use to simple, slender geometries. Moreover, the Cosserat rod models do not fully consider the complex tendon interactions in the system. Herein, we propose an alternative high-fidelity finite-element model for tendon-driven soft robotics systems, incorporating a hyperelastic material model, contact definitions within and between the bodies, and frictional forces, to accurately capture the deformation of tendon-driven locomotion and manipulation systems, achieving a mean tip position error of less than 7.91 percent. In addition, the FEM model is verified by comparing the complete shape of the actuators using curvature and torsion metrics. The efficacy of the proposed FEM model is demonstrated through a parametric analysis of tendon-driven soft actuators and by modeling nonslender actuator geometries subjected to significant contact interactions. This research paves the way for scalable and generalizable computational models to enhance the practicality of tendon-driven soft robotics.