An approach to the design of discrete-time decentralized controI systems based on model-based predictive controI (MBPC) and neural estimation is proposed. The class of interconnected large-scale systems (LSS) is considered, and a model is used at each controI station to predict the corresponding subsystem output over a long time period. In the case of subsystems with m-step delay information patterns the non-locally available interaction trajectories are estimated by a multi-layer neural network trained on-line with a modified backpropagationtype algortithm. Representative computer simulation results are provided and compared for a set of illustrative examples. The proposed controI scheme shows better performance than the other schemes, and also covers the important case where the subsystems' interactions are nonlinear.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.