关键字：terminal ILC; data-driven control; high-order algorithm; fed-batch processes; nonlinear systems

摘要：A high-order data-driven optimal terminal iterative learning control (H-DDOTILC) is proposed for fed-batch processes, which are considered a general class of nonlinear and non-affine systems. A new dynamical linearization is introduced to the iteration domain to reveal the relationship of system terminal output and control input among batches. The proposed H-DDOTILC consists of a high-order learning control law, an iterative parameter estimator, and a rest algorithm, together. The learning control law with a high-order form is capable of utilizing more control knowledge of the previous l batches to improve control performance. The parameter updating law is used to estimate the unknown derivatives of the nonlinear system to control input, which is the main part of the nonlinear learning gain function of the control law. Essentially, the proposed approach is a data-driven control strategy, and the controller design and analysis only depend on the I/O data of the plant, which is a distinct feature for the control problems of practical nonlinear and non-affine systems. Both the rigorous analysis and the simulation results illustrate the applicability and effectiveness of the proposed approach.

卷号：93

期号：8

是否译文：否