Robust Control Performance in Uncertain Environments: A Semi-Heuristic Gradient Descent Approach

Designing a robust controller for systems with parameter uncertainties is a complex and demanding task. Traditional deterministic and probabilistic approaches may fall short in providing efficient and satisfactory solutions. To address this challenge, we propose a semi-heuristic approach that exploits the Kharitonov theorem to establish an initial point for the gradient descent algorithm. Through an iterative optimization process that incorporates user-defined performance criteria, our approach provide a robust controller with respect to presence of system uncertainties. Numerical simulations validate the effectiveness of our proposed method and highlight its superiority in addressing robust control problems compared with auto-tuned ΡΠ) controller.