Artificial Neural Network Controller for DC-DC Boost Converter: A Design and Performance Analysis

This study introduces an artificial neural network (ANN)-based machine learning controller for the DC-DC boost converter. The primary controller is a Disturbance Observer based Feedback Linearization controller, which serves as an expert to provide training data for the proposed ANN. After fine-tuning the ANN is seamlessly integrated into the feedback loop, directly facilitating boost converter control. The key advantage is in the ANN ability to enhance system identification, reduce model errors, and accommodate uncertain parameters. MATLAB/Simulink simulations validate the high performance of the ANN controller, showcasing its capability to follow dynamic reference commands fast, maintain output stability amidst input voltage variations, and effectively handle constraints on maximum duty-ratio and current.