Luenberger, Kalman and neural network observers for sensorless induction motor control

Abstract

This paper presents a comparison between the sensorless vector control schemes of the induction motor using the Luenberger observer, the Kalman filter and a neural network observer. The first two methods have been implemented on a digital signal processor (DSP). Different possibilities for reducing the complexity of their implementation are discussed. This is of particular relevance for industrial applications based on DSP microcontrollers. The performance for the third method is appreciated by simulation tests.