NARMA-L2 controller for stepper motor used in single link manipulator with low-speed-resonance damping

Abstract

Stepper motors are used in robotic applications, industrial machines, and many different application areas. Mechanical resonance problems occur in the control of stepper motors at low speeds. In the control of robotic systems with nonlinear characteristics, conventional control methods are inadequate. Single link manipulators are the most basic structure used in the identification and control of the nonlinear load in robotic systems. In this study, the control of single link manipulator was performed using the stepper motor. As in a real system, the single link manipulator is modelled by connecting to the stepper motor via the gearbox. An Artificial Neural Network (ANN)-based NARMA-L2 controller was preferred for the control of the stepper motor. The NARMA-L2 controller offers superior performance especially in the control of nonlinear loads. Firstly, simulation of the single link manipulator and the stepper motor, which also included resonance moments, was carried out in the MATLAB/Simulink environment. Then the training data of the NARMA-L2 controller was created and the training process was realized. Position control of the stepper motor was performed in MATLAB/Simulink environment using conventional control method and NARMA-L2 controller. The simulation results were obtained at three different speeds and for different angle values. By comparing the conventional control method with the proposed control method, the superior performance of the NARMA-L2 controller in the speed and position control of the motor was demonstrated.