The intrinsic compliance of soft continuum robots makes them well‐suited for gentle and adaptable grasping in dynamic and unstructured conditions. However, common actuation mechanisms—such as cables or pneumatic pumps—often involve bulky powering components that do not scale well in number, practically hindering their applicability in real‐world applications. This work presents a computational framework for the optimal tendon routing in cable‐driven soft manipulators, so as to match a set of target grasps while reducing the number of actuators. The design of an octopus‐inspired soft robotic arm for underwater manipulation is taken as a case study. The framework employs a genetic algorithm combined with the finite element method to identify the optimal tendon number and routing to reach desired bio‐inspired grasping poses. Numerical results are validated by fabricating two optimal solutions and comparing them with simulations. Results show good reproducibility across multiple simulations and experimental results are in agreement with the macroscopic configurations obtained in simulation (average error of 15%, 9.63%, and 11.33% for fetching, reaching and twisting poses, respectively, relative to the arm's length). This work provides a pipeline for designing underactuated soft arms, thereby facilitating their application in a real‐world environments.
Martini et al. (Mon,) studied this question.