부산대학교 도서관

Soft robots show significant potential for use in search and rescue, human-robot interaction, and other emerging fields due to their ability to easily conform, deform, and interact with their environment. However, precise control of these soft robots is still being explored. In this letter, we investigate a potential solution to address the limitations of precise control for soft robots. We experimentally verify the accuracy of a general analytical formulation of a continuous kinematic model using a custom soft growing manipulator. Next, we provide an experimental verification of its inverse kinematic model. With this precise model, most soft continuum kinematic models, whether tendon-driven or with a payload, can be represented. Our robot fits the proposed generalized curvature function for $n=2$, with an average error relative to the robot's overall length of 5.01%, based on four robot lengths of 0.5 m, 0.8 m, 1.0 m, and 1.2 m. The inverse kinematic model was verified using three positions, resulting in errors of 2.91%, 7.91%, and 2.14%. We also showed that the shape can be recovered based on using the tip position in the inverse kinematic model. Our future work will involve verifying this model in 3D space and incorporating it into a model-based feedback loop controller to enhance position control accuracy.