Internal model decoupling control based Diagonal Recurrent Neural Network

Aiming at the multivariant, nonlinear, time variation and strong coupling plant, a neural network decoupling control based on the internal model control was proposed, which has the advantages both the internal model control and the neural network decoupling control. Several nonlinear PID controller based on diagonal recurrent neural network are adopted in parallel as the neural network decoupling controller. Then the stability condition of the identifier and controller is presented based on the Lyapunov theory. Simulation example are given to show effectiveness of the proposed decoupling control.