Hybrid Diagnosis with Unknown Behavioral Modes

A novel capability of discrete model-based diagnosis not impose such a strong modeling assumption. Its concept of the methods is the ability to handle unknown modes where no assumpunknown mode allows diagnosis of systems where no assumption is tion is made about the behavior of one or several components of the made about the behavior of one or several components of the syssystem. This paper incorporates this novel capability of model-based tem. In this way, it captures unspecified and unforeseen behaviors diagnosis into a hybrid estimation scheme by calculating partial filof the system under investigation. This paper provides an approach ters. The filters are based on causal and structural analysis of the to incorporate the concept of an unknown mode into our hybrid esspecified components and their interconnection within the hybrid autimation scheme[9]. As a result we obtain an estimation capability tomaton model. Incorporating unknown modes provides a robust esthat can detect unforeseen situations. Furthermore, it allows us to timation scheme that can cope, unlike other hybrid estimation and continue estimation on a degraded basis. We achieve this by causal multi-model estimation schemes, with unmodeled situations and paranalysis[ 17, 20], structural analysis[7] and decomposition of the systial information. tem. This paper starts with a brief introduction to our hybrid systems modeling and estimation scheme. Upon this foundation, we extend