Dynamic performance optimization techniques for Buck–Boost converters based on adaptive neural network control

Buck–Boost converters are widely employed in modern power electronic systems to provide flexible voltage regulation across varying load and input conditions. However, their dynamic performance is often constrained by nonlinear characteristics, parameter uncertainties, and external disturbances, posing challenges for achieving fast and stable control under complex operating environments. To address these limitations, this paper proposes a dynamic performance optimization approach for Buck–Boost converters based on an adaptive neural network control framework. The method leverages an online-learning neural network to compensate for model uncertainties and nonlinear behaviors, while a baseline controller ensures fundamental stability. An integrated adaptation mechanism is introduced to adjust control parameters in real time, enabling enhanced transient response and improved steady-state accuracy. Experimental validation using a digitally controlled Buck–Boost prototype demonstrates that the proposed scheme maintains stable operation under wide input variations and load disturbances. Trend-level comparisons with conventional control strategies indicate superior dynamic performance, enhanced robustness, and reduced overshoot, highlighting the potential of adaptive neural network control for next-generation intelligent power conversion systems.