Control of Dielectric Elastomer Soft Actuators Using Antagonistic Pairs

The inherent compliance and resilience allow soft robots to deal with unstructured environments. Among the various soft actuators explored for soft robotic applications, dielectric elastomer actuators (DEAs) stand out due to their intriguing advantages of large deformation, high energy density, fast response, and low cost. However, the viscoelasticity and the electromechanical coupling of these actuators increase the difficulty of the control design, especially for large deformation and/or high-speed operations. In this paper, the control issue of the DEAs is investigated. Inspired by the similarities between DEAs and biological muscles, an antagonistic actuation mechanism is proposed for DEAs control. Owing to the actions of the agonist can be compensated by the antagonist, the viscoelastic effects of the DEAs in antagonistic pairs can be compensated by each other. Moreover, a cerebellum-inspired adaptive learning controller for achieving accurate movements of the DEAs is designed and analyzed. To study the proposed actuation mechanism and controller, a simple 1-DOF manipulator based on two DEAs in an antagonistic pair is developed as a platform. Finally, several experiments have been conducted to test the developed system, and the results show that the proposed approach can achieve good performance even in large deformation and/or high-speed operations as well as strong robustness in disturbance rejection.