Approximate Optimal Decoupling Control Method Based on Disturbance Observer

We propose an approximate optimal decoupling control method based on a neural network disturbance observer for a class of linear systems with strong couplings and nonlinear immeasurable disturbances. First, we use a recurrent neural network to observe the disturbance. Then, we design a feed-forward decoupling controller to convert the strong coupling system into multiple single-variable systems. Finally, based on the neural network disturbance observer, we design an approximate optimal tracking controller. Numerical simulation experiments, using a ball mill as a simulation object, verify that this method can eliminate the effects of coupling and immeasurable interference between loops and achieve an ideal control effect.