Real-time implementation of a neural block control using sliding modes for induction motors

Author

Elena Antonio-Toledo, M. ; Sanchez, Edgar N. ; Loukianov, Alexander G.

Author_Institution

CINVESTAV Unidad Guadalajara, Zapopan, Mexico

fYear

2014

fDate

3-7 Aug. 2014

Firstpage

502

Lastpage

507

Abstract

In this paper, a controller for induction motors is proposed. A recurrent high order neural network (RHONN) is used to identify the plant model, which is trained with an Extended Kalman Filter (EKF) algorithm. The control scheme is based on discrete-time block control technique using sliding modes (SM) for tracking position trajectory. The effectiveness of the proposed control scheme is verified via real-time implementation using a three-phase induction motor.

Keywords

Kalman filters; discrete time systems; identification; induction motors; learning (artificial intelligence); machine control; neurocontrollers; nonlinear filters; recurrent neural nets; variable structure systems; EKF algorithm; RHONN training; SM; discrete-time block control technique; extended Kalman filter algorithm; neural block control; plant model identification; position trajectory tracking; real-time implementation; recurrent high-order neural network; sliding modes; three-phase induction motor controller; Covariance matrices; Induction motors; Neural networks; Rotors; Stators; Training; Vectors;

fLanguage

English

Publisher

ieee

Conference_Titel

World Automation Congress (WAC), 2014

Conference_Location

Waikoloa, HI

Type

conf

DOI

10.1109/WAC.2014.6936017

Filename

6936017