Identification and control of brushless DC motors using on-line trained artificial neural networks

This paper proposes high performance with simultaneous online identification and control designed for brushless DC motor drives. The dynamics of the motor/load are modeled online and controlled using an artificial neural network (ANN) based identification and control scheme incorporating three multilayer feedforward neural networks that are trained online using the gradient descent training algorithm. The control of the direct and quadrature components of the stator current successfully tracked a wide variety of trajectories. The control strategy adapts to the uncertainties of motor/load dynamics, and, in addition, learns their inherent nonlinearities. The use of feedforward neural networks makes the drives system robust, accurate and insensitive to parameter variations.