Neuro-self tuning adaptive controller for non-linear dynamical systems

Abstract EN

In this paper, a self-tuning adaptive neural controller strategy for unknown nonlinear system is presented.

The system considered is described by an unknown NARMA-L2 model and a feed forward neural network is used to learn the model with two stages.

The first stage is learned off-line with two configuration serial-parallel model & parallel model to ensure that model output is equal to actual output of the system & to find the Jacobin of the system.

Which appears to be of critical importance parameter as it is used for the feedback controller and the second stage is learned on-line to modify the weights of the model in order to control the variable parameters that will occur to the system.

A back propagation neural network is applied to learn the control structure for self-tuning PID type neurocontroller.

Where the neural network is used to minimize the error function by adjusting the PID gains.

Simulation results show that the self-tuning PID scheme can deal with a large unknown nonlinearity.