Neuro-controller for high performance induction motor drives in robots

Abstract

Presents an approach to the speed control of an induction motor (IM) as a robust high performance drive (HPD) using an online self-tuning adapted artificial neural network (ANN). Based on motor dynamics and nonlinear unknown load characteristics such as robot systems, a neuro speed controller is developed. The proposed controller is very simple and serves as an identifier and a controller at the same time. The combination of the adaptive learning rate with the epochs used through the online training offers a unique feature of system identification and adaptive control. The performance of the controller was evaluated under various operating conditions to track different speed trajectories. The results validate the efficacy of the ANN for the precise tracking control of IM. Furthermore the use of the ANN makes the drive system robust, accurate, and insensitive to parameter variations. Also the drive system is implemented in real-time using a digital signal processor (DSP) TMS320C31.