Artificial neural networks for decentralized control of large-scale adaptive structures

In this paper, based on the concept of design of informationally decentralized controllers, a decentralized non-parametric identification and control strategy with artificial neural networks for large-scale structures is proposed. A six-cable-stayed bridge model is treated as a representation of large-scale structures here. The cable-stayed bridge is divided into several subsystems. Each individual actuator for a subsystem is controlled by a decentralized neurocontroller that only uses local state feedback from the subsystem. The effectiveness and adaptability of the proposed decentralized algorithm basing on artificial neural networks is evaluated through numerical simulations.