ABSTRACT The scope of output-only/blind identification is restricted to stochastic/statistical processes, but for the first time in this study, the detectability conditions for general output-only subspace identification are investigated. This aids the range of input sources to be extended in a much realistic manner, beyond the only stochastic inputs. For this purpose, the subspace framework is assigned to make a connection between the output signal contents and the LTI system order. A few substantial hypotheses and algebraic statements are propounded affirming the sufficiency of the genuine output sequences for the identification purpose. This can be perceived as the cornerstone of state-space model reconstruction. In order to consolidate the notions according to reality, several examples are studied and examined for different input classes with stochastic disturbance.

中文翻译：

---

仅输出子空间识别的可检测性条件

摘要 仅输出/盲识别的范围仅限于随机/统计过程，但在本研究中首次研究了一般仅输出子空间识别的可检测性条件。这有助于以非常现实的方式扩展输入源的范围，而不仅仅是随机输入。为此，分配子空间框架以在输出信号内容和 LTI 系统顺序之间建立联系。提出了一些实质性的假设和代数陈述，肯定了用于识别目的的真实输出序列的充分性。这可以看作是状态空间模型重建的基石。为了实事求是地巩固观念，