Robust-Optimal Output-Voltage Control of Buck Converter using Fuzzy Adaptive Weighted Combination of Linear Feedback Controllers

Abstract

This paper presents a computationally-intelligent adaptive weighted controller combination scheme to optimize the output-voltage regulation capability of a low power DC-DC buck convertor. The proposed scheme beneficially combines two linear feedback controllers; namely, Proportional-Integral-Derivative (PID) controller and Linear-Quadratic-Regulator (LQR). The PID controller provides control effort based on the error-dynamics of output-voltage. Wherein, the term regarding the error-derivative is replaced with the information of capacitor-current to nullify the effects of noise injected by the derivative action during transients. The LQR provides optimal control decisions by utilizing the state-feedback of inductor-current and output-voltage. The outputs of PID controller and LQR are linearly combined by computing their weighted sum. The fixed weightages associated with each controller cannot compensate the parametric uncertainties and load-step transients. Therefore, the weightages are adaptively self-tuned via a hyperbolic tangent function of error in output voltage. The performance of weighted control scheme is also investigated by augmenting it with a fuzzy inference system that directly captures the variations in output-voltage and capacitor-current to adaptively self-tune the weightages. The performances of aforementioned weighted controllers are comparatively analyzed via credible real-time experiments. The fuzzy weighted controller yields time-optimal control effort during step-reference tracking and offers minimum-time transient recovery during load variations.