A dynamically-constructed fuzzy neural controller for direct model reference adaptive control of multi-input-multi-output nonlinear processes (Q1864391)

The study deals with the development of a feedback fuzzy neural controller for a general class of MIMO systems governed by the expression \[ {\mathbf x}(k+1)= {\mathbf f}\bigl( {\mathbf x}(k),{\mathbf u}(k),{\mathbf d} (k)\bigr){\mathbf y}(k)= {\mathbf g}\bigl({\mathbf x}(k), {\mathbf v}(k)\bigr) \] with \({\mathbf x}(k)\), \({\mathbf u}(k)\), \({\mathbf y}(k)\) denoting system, control, and output variables, respectively and \({\mathbf d}(k)\) and \({\mathbf v}(k)\) being input disturbances and measurement noise. The authors present a generic topology for the controller followed by a detailed gradient-based learning scheme. The performance of the controller is illustrated in the case of a cold mill thickness control. It is then contrasted with the existing approaches utilizing conventional PID controllers and cascade correlation neural networks.