Applying an MSVR Method to Forecast a Three-Degree-of-Freedom Soft Actuator for a Nonlinear Position Control System: Simulation and Experiments

Abstract

In this article, a method used for tip-position coordinate control of a three-degree-of-freedom (DOF) soft actuator is proposed.In general, the behavior of pneumatic soft actuators is simple. However, the actuator, which consists of three artificial muscles, is capable of more complex motions compared to conventional soft actuators. By designing a model and control system that can handle multiple input patterns, various motions are possible. In addition, a machine learning technique called multioutput support vector regression (M-SVR) is used as a method to compensate for the complexity of multiple-input, multiple-output systems. First, a model that can be used to design a control system is offered. Then, a control system is designed, using the recommended model and machine learning approaches. Furthermore, the effectiveness of the proposed system is verified by experiments.