Intelligent Estimation of Control Parameters of a Flexible Robot by Using Non-Dominated Sorting Genetic Algorithm (NSGA-II)

Due to the inherent limitations of feedback linearization in fully eliminating nonlinear dynamics and handling system uncertainties, numerous strategies have been proposed, including parameter adaptation and integration with advanced nonlinear controllers. In this study, we present an online hybrid adaptive robust control framework founded upon the Non-Dominated Sorting Genetic Algorithm. The control process begins by linearizing the nonlinear system equations using feedback linearization. To address the persistent nonlinear behavior in the output states, an adaptive robust sliding mode control is applied. This control is enhanced by a novel mathematical framework that updates controller parameters via the gradient descent method, utilizing the chain rule of derivation. Comparative evaluation of the proposed controller against conventional methods demonstrates superior robustness and stability, with the system achieving rapid convergence and improved performance.