Dual Design Iterative Learning Controller for Robotic Manipulator Application

• Phichitphon Chotikunnan

  Department of Control and Instrumentation Engineering, King Mongkut’s University of Technology Thonburi, Bangkok 10140, Thailand

• Benjamas Panomruttanarug

  Department of Control and Instrumentation Engineering, King Mongkut’s University of Technology Thonburi, Bangkok 10140, Thailand

• Poramate Manoonpong

  Bio-Inspired Robotics and Neural Engineering Laboratory, School of Information Science and Technology, Vidyasirimedhi Institute of Science and Technology, Rayong 21210, Thailand Embodied AI and Neurorobotics Laboratory, SDU Biorobotics, The Mærsk Mc-Kinney Møller Institute, University of Southern Denmark, Odense 5230, Denmark

Iterative learning control enables high precision performance through observed historical data in previous iterations. Several techniques for designing iterative learning controllers have been developed in the existing literature. However, evidence to support the design’s efficiency in real applications is, unfortunately, missing in some designs. This paper presents a practical iterative learning controller design, so-called the dual design, combining two existing controller designs using the weighted sum technique. The two controllers are designed using data-driven and frequency response approaches, distinctively selected to take benefits from each. A single gain controller designed from the gain adjustment mechanism usually has slow learning behavior but can be very robust to external uncertainty. The other design imitating the inverse of the frequency response of the system can learn extremely fast. However, its performance may not be as effective as desired when the frequency response of the system is incorrectly perceived. By taking advantage of both controllers, the dual design can achieve fast-learning behavior as well as robustness to external disturbances. Simulation and experiments were carried out to demonstrate the design efficiency.

• PDF