Session: 09-03-01: Applied Mechanics, Dynamic Systems, and Control Engineering

Paper Number: 95524

95524 - Benchmarking Various Nonlinear Control Design Techniques for a Two-Link Planar Robot Arm 

The aim of this work is to provide an in-depth study of the kinematics, dynamics, motion planning, and different nonlinear control design techniques applied for a simple, two-link planar robot arm. In an era with increased attention and interest on robotics, the two-link arm still provides the foundation for the design and control of more sophisticated serial robotic manipulators and/or humanoid robots. The joints of the two-link robot considered in this study are assumed to be equipped with motors for providing the torque inputs, encoders for measuring the joint positions, and tachometers for measuring the joint velocity. Through motion planning and inverse kinematic analysis, the motion in the task space is converted into the joint space, and the desired joint variables are then extracted for feedback control design. To facilitate the control design, a control-oriented dynamic model of the system is obtained through Lagrange’s equations in nonconservative form. The resulting nonlinear dynamic model is then incorporated into feedback control design based on three different nonlinear design techniques: First, the feedback linearization technique is implemented through an input transformation, which leads to the standard “computed torque” method commonly used in robotics. The performance of this first approach is validated in a closed-loop numerical simulation in MATLAB, in which the robustness of the algorithm is tested by simulating the disturbed model, in which some parametric uncertainties of known upper bounds are added on the link inertia. Next, a more sophisticated, robust control approach is followed based on the nonlinear Sliding-Mode Control (SMC) strategy. To facilitate the nonlinear SMC design, a performance index is first defined through a rated combination of the position tracking error and the velocity tracking error, and the nominal Lagrangian dynamic model is incorporated into the SMC design to minimize this performance index. Input chattering is reduced to avoid unnecessary control switching action, and the tracking performance of the SMC algorithm is then tested on the same disturbed model as in the computed torque simulations to reflect the performance improvement under robust design approach. Finally, an adaptive nonlinear controller is designed based on the Model-Reference Adaptive Control (MRAC) scheme, in which the controller has no a priori knowledge on the limits of the parametric uncertainty in the system. The tracking performance of the adaptive controller is also tested on the same disturbed model to provide a comparison with the performance of the computed torque and SMC methods. Design challenges and possible improvement areas are addressed based on these initial results.

Presenting Author: Zeki Ilhan Midwestern State University

Presenting Author Biography: Dr. Ilhan received his PhD in Mechanical Engineering from Lehigh University in 2016. He joined Midwestern State University, Texas (MSU Texas) in Fall 2017 as a Visiting Assistant Professor, and was hired to a tenure-track position in Fall 2018. Prior to joining MSU, he taught one semester at the U.S. Coast Guard Academy as a part-time Lecturer in mechanical engineering. Dr. Ilhan’s research interests include control-oriented modeling, optimal, nonlinear and distributed parameter system controls with emphasis on control of plasma profiles in nuclear fusion, and nonlinear control-mechanisms design integration.

Authors: