A data-driven sigmoid-based PI controller for buck-converter powered DC motor

Abstract

This paper presents a novel data-driven sigmoid-based PI for tracking of angular velocity of dc motor powered by a dc/dc buck converter. A global simultaneous perturbation stochastic approximation (GSPSA) is employed to find the optimum sigmoid-based PI parameters such that the angular velocity error is minimized. The merit of the proposed approach is that it can produce fast PI parameter tuning without using any plant model by measuring the I/O data of the system. Moreover, the proposed PI parameters that are varied based on sigmoid function of angular velocity error has great potential in improving the control performance compared to the conventional PI controller. A well-known buck converter powered DC motor model is considered to validate our data-driven design. In addition, the performances of the proposed method are examined in terms of angular velocity trajectory tracking and duty cycle in comparison with other existing approaches. Numerical example shows that the data-driven sigmoid-based PI approach provides better control performances as compared to existing methods.