Design and Verification of Model-based Nonlinear Controller for Fluidic Soft Actuators

Abstract

Soft actuators are of great interests in recent years due to inherent compliance and adaptability. However, most of current studies focus on the physical actuation phenomenon and their manufaction, few results can be found on the modeling and closed-loop control due to their complicated elastic deformation. This work studies physical modeling, parameter identification, and nonlinear robust backstepping control for fluidic soft bending actuators governed by high-speed on-off solenoid valves. A second-order transfer function is used to describe the dynamic behavior from driving pneumatic pressure to soft actuator bending angle, and the model parameters are identified by with different operating frequencies. The nonlinear dynamical model of pneumatic valves is also built and identified by using least square method. Based on the identified high-order model, a robust backstepping control algorithm is designed for soft actuator to deal with nonlinearities and parameter uncertainties. Experimental results show that the closed-loop stability is guaranteed and the good tracking performance is achieved by the proposed method.