Hierarchical Network Identification for Large-Scale Dissipative Systems

在本研究项目中，我们考虑了分层网络系统，首先推导了系统辨识的充分条件，并给出了一种基于子空间辨识方法的辨识方法（称为分层辨识方法）。作为该方法的扩展，我们还提出了一种基于分布固有正交分解的模型约简方法。此外，我们将分层网络辨识方法推广到耗散系统，并推导出一种辨识供给率的辨识方法。我们还应该强调，在得出本项目所需的初步结果时，我们得出了一些关于互联和分布参数系统的李雅普诺夫稳定性、耗散率和有限频率特性的原始结果。另一方面，通过应用本项目和工具箱开发的结果进行验证仍然是未来的工作。

In this research project, we considered hierarchical network systems and first derived a sufficient condition for identification of the system and gave an identification method (called hierarchical identification method) based on subspace identification method. As an extension of the method, we also developed a model reduction method based on distributed proper orthogonal decomposition. Moreover, we extended the hierarchical network identification method to dissipative systems and also derived an identification method for identification of supply rates. We should also stress that we derived some original results on Lyapunov stability, dissipativity and finite frequency properties of interconnected and distributed parameter systems in deriving preliminary results necessary to this project. On the other hand, it remains a future work to give a validation by application of the results of this project and toolbox developments.