Dynamic Task Space Control Enables Soft Manipulators to Perform Real‐World Tasks

Abstract

Dynamic motions are a key feature of robotic arms, enabling them to perform tasks quickly and efficiently. A majority of real‐world soft robots rely on a quasistatic control approach, without taking dynamic characteristics into consideration. As a result, robots with this restriction move slowly and are not able to adequately deal with forces, such as handling unexpected perturbations or manipulating objects. A dynamic approach to control and modeling would allow soft robots to move faster and handle external forces more efficiently. In previous studies, different aspects of modeling, dynamic control, and physical implementation have been examined separately, while their combination has yet to be thoroughly investigated. Herein, the accuracy of the dynamic model of the multisegment continuum robot is improved by adding new elements that include variable stiffness and actuation behavior. Then, this improved model is integrated with state‐of‐the‐art system identification and dynamic task space control and experiments are performed to validate this combination on a real‐world manipulator. As a means of encouraging future research on dynamic control for soft robotic manipulators, the source code is made available for modeling, control, and system identification, along with the recipes for fabricating the manipulator.