Design of Multivariable Self-Tuning PID Controllers via Quasi-diagonal Recurrent Wavelet Neural Network

Multivariable PID controllers have recently emerged as a kind of convenient yet very powerful control technique for solving coupling nonlinear system. This article describes a new method for design of multivariable PID based on Quasi-Diagonal Recurrent Wavelet Neural Network (QDRWNN). Firstly, Due to the advantages of Wavelet Neural Network (WNN) and Diagonal Recurrent Neural Network (DRNN) such as the good learning ability, generalization of wavelet transform, dynamic mapping and converges quickly, we present a novel Neural Network QDRWNN. Secondly, the new Neural Network is used to identify the coupling nonlinear system on line and tune parameters of multivariable PID controllers automatically. Finally, an illustrative example is given to demonstrate the feasibility and validity of the proposed method.