中文说明：应用背景本文研究了一个不确定的潜在控制回路的在线推断稳定性植物相互关联的反馈与候选控制器使用工厂/输出对记录，而工厂可能是由不同的控制器驱动。在这样的背景下，一个方便的工具来工作的是诉诸于概念实体的虚拟参考（虚拟现实）。所采用的方法包括嵌入，在所谓的非伪自适应切换基于VR的控制方案，一个家庭的名义模型与给定的候选控制器相关联的。结果是，有监督的开关机制可以适度的机会，不稳定的控制器切换，因此，减少的幅度和时间的“学习”过程启动后，在时间中，与原有的基于多模型的方法相比，在大的稳定是保证在最小的可能假设一个稳定的候选控制器存在。关键技术不确定植物的控制自适应控制开关监控
English Description：
Application background this paper studies the on-line inferential stability of an uncertain potential control loop. The feedback and candidate controllers associated with plants are recorded using factory / output pairs, and the factory may be driven by different controllers. In this context, a convenient tool to work is to resort to virtual reference (virtual reality) of conceptual entities. The methods employed include embedding, in a so-called non pseudo adaptive switching based VR control scheme, a family of nominal models associated with a given candidate controller. The result is that the supervised switching mechanism can moderately chance unstable controller switching, so the reduced amplitude and time after the "learning" process starts in time, compared with the original multi model based method, the large stability is guaranteed at the minimum possible assumption that a stable candidate controller exists. Key technology control of uncertain plants adaptive control switch monitoring