Study on active disturbance rejection controller based on Neural Network in press overprint system

Based on the characteristic of large parameter-variation range in the overprint correction system of gravure press, Neural Network (NN) was embedded in active disturbance rejection controller. An NN-based active disturbance rejection control strategy was proposed and applied to control the overprint correction system of gravure press in this paper. The controller structure as well as training samples and methods for the applied NN were analyzed and discussed, and positive conclusions were drawn. Simulation results show that compared with single active disturbance rejection controller in application, the NN-based one has the advantages of higher parameter adaptability and better dynamic characteristics, etc.