Neural network observer for induction motor control

In many practical problems the task is the control of a nonlinear plant, under parameter uncertainty, by a controller of known structure that uses the values of the state variables. However, it frequently is the case that some state variables cannot be measured. In this article this type of problem arises from the control requirements of an induction motor and is tackled by neural network based observer techniques so that the state variables are estimated while the variations of the unknown parameters are compensated for. Extensive simulations have shown that this scheme works well in the case of motor control.<>