Closed-Loop System Identification by Using Multi-Rate Sampling Scheme

本文提出了一种新的直接闭环辨识方法，该方法仅使用通过输出互采样方案获得的对象输入/输出数据。通过对系统输出进行比控制间隔更快的采样，我们可以澄清，间采样的对象模型也可以由SIMO模型结构来描述，这可以提供闭环的可辨识性。该辨识方法的优点之一是，利用互采样输出观测值可以消除通常的可辨识性限制条件，并且不要求参考输入具有持续激励（PE）特性。在广义预测控制的数值仿真和PID控制器稳定磁悬浮系统的实验研究中，证明了该方法的有效性。我们还澄清，

A new direct closed-loop identification method only using the plant input/output data acquired through an output inter-sampling scheme is presented in this study. By taking the faster sampling of the system output than the control interval, we can clarify that the inter-sampled plant model can also be described by the SIMO model structure, which can provide the identifiability of the closed-loop. One of the advantages is that the proposed identification method using the inter-sampled output observations can remove the ordinary restrictive identifiability condition, and it is not required that the reference input holds the persistently exciting (PE) property. Its effectiveness is demonstrated in numerical simulation treating a generalized predictive control, and an experimental study of a magnetic suspension system stabilized by PID controller. We also clarified that the