Project Description

Decentralized diagnosis of discrete event systems has received a lot of attention to deal with distributed networked systems or with systems that may be too large to be diagnosed by one centralized site. This work casts the problem of decentralized diagnosis of networked systems in a new hierarchical framework. A key feature of this framework is the exploitation of different local decisions together with appropriate rules for their fusion in the diagnostic process. This includes local diagnosis decisions that can be interpreted as ``conditional decisions". Under this new framework, a series of new decentralized architectures are being defined and studied. The properties of their corresponding notions of decentralized diagnosability are being investigated and their relationship with existing work described. Corresponding verification algorithms also need to be developed. The publications listed below summarize our results to-date on this problem.