A new neural network-based approach for self-tuning control of nonlinear multi-input multi-output dynamic systems

This paper presents a new neural network-based approach for self-tuning control of nonlinear MIMO dynamic systems. According to the approach, a neural network ARMAX (NN-ARMAX) model of the system is identified and continuously updated, using an online training algorithm. Control design is accomplished by solving an optimal discrete-time linear quadratic tracking problem using an observable block companion form Kalman innovation model, which is built from the parameters of a local linear version of the NN-ARMAX model. The state-feedback control law is implemented using the Kalman estimated state, which is calculated without estimating the noise covariance properties. The effectiveness of the proposed control approach is illustrated using a simulation example.

I. Introduction

SELF-TUNING control refers to approaches in which the design of the control law is based on a model which is identified and continuously updated, using measured input-output data. Hence, self-tuning control strategies are well suited for systems whose model is unknown and/or time-varying systems.