Single layer type II Chebyshev neural network based adaptive backstepping control of DC-DC buck converter

Abstract

This article presents a novel adaptive control method based on neural networks for robust output voltage tracking in buck converters over a wide operating range. Buck converters are significantly sensitive to input, parametric and load perturbations. The intrusion of mismatched uncertainties due to load changes make the controller design task a challenging issue. Hence, a feedback control law based on the adaptive backstepping control technique integrated with a single layer type II Chebyshev neural network (CNN) is proposed. The distinctive feature of the type II CNN is its quick and accurate estimation of time varying load disturbance which is thereafter utilized for subsequent compensation in the control law. The neural networks are trained online using a Lyapunov based learning algorithm. The efficacy of the proposed control is studied for wide variations in load resistance, input voltage and reference voltage and compared against control using conventional adaptive backstepping method. Simulations are performed in MATLAB tool and experimentation is conducted using dSPACE DS1103 setup with TM320F240 DSP. The results demonstrate a good agreement between the simulation and experimental findings. Further, the proposed control achieves a remarkable reduction in settling time and peak overshoot/undershoot in the event of occurrence of unanticipated disturbances.