Many diagnosis approaches for discrete event systems are diagnoser-based. A diagnoser is a deterministic automaton that can be built directly from the underlying system model by performing ε-reduction and determinization operations. The diagnoser-based approaches allow for analyzing diagnosability, but they also support online diagnosis in a straightforward way. The procedure for investigating diagnosability on the diagnoser consists in checking the existence of indeterminate cycles, and requires to verify for every F-uncertain cycle in the diagnoser whether there exists two corresponding cycles in the system model such that one is fault-free while the other is faulty. The present work aims to improve the efficiency of the diagnoser-based approaches by establishing a diagnoser variant that offers a convenient structure to help enhance the diagnosability analysis procedure. Namely, it consists in separating the normal states from each faulty state classes in each diagnoser node. Such a distinction serves to track the faulty and fault-free sequences in the diagnoser paths more efficiently. On the basis of various features offered by the diagnoser variant, we put forward simplified necessary and sufficient conditions for two diagnosability properties: (i) diagnosis of fault occurrences and (ii) detection of fault absence, i.e., non-fault detection. Such conditions are established for both cases of a single fault class and multiple fault classes, using the notion of indicating sequences associated with the F-uncertain cycles in the diagnoser.

离散事件系统的许多诊断方法都是基于诊断器的。诊断器是一种确定性自动机，可以通过执行ε 约简和确定化操作直接从底层系统模型构建。基于诊断器的方法允许分析可诊断性，但它们也以直接的方式支持在线诊断。在诊断器上调查可诊断性的过程包括检查不确定循环的存在，并要求在诊断器中的每个 F 不确定循环中验证系统模型中是否存在两个相应的循环，使得一个无故障，另一个有故障。目前的工作旨在通过建立一个诊断程序变体来提高基于诊断程序的方法的效率，该变体提供方便的结构来帮助增强可诊断性分析程序。即，它包括将正常状态与每个诊断节点中的每个故障状态类分开。这种区别用于更有效地跟踪诊断器路径中的故障和无故障序列。在诊断器变体提供的各种特征的基础上，我们提出了两个可诊断性属性的简化充要条件：（i）故障发生的诊断和（ii）故障缺失的检测，即，非故障检测。使用以下概念，为单个故障类别和多个故障类别的情况都建立了这样的条件指示与诊断器中 F 不确定循环相关的序列。