Adaptive nonlinear control of induction motor using neural networks

To avoid the various constraints related to the feedback linearisation control (FBLC), in this papers we propose a new control approach for the induction motor control based on artificial neural networks (ANN) trained on-line. The two ANN are used for the on-line reconstitution of the state feedback necessary for the FBLC. The training rules used result from a combination between the ANN properties, the adaptive nonlinear control propriety and the nonlinear adaptation rules. Via these three techniques a training rules were extracted, these last transform the tracking errors into a means to adjust the used ANN behavior so that they adapt with the various operation modes of induction motor.