A new protocol for distributed diagnosis

In this paper, we propose a new protocol for distributed diagnosis, where distributed diagnosers share their diagnosis information through bounded-delay channels. In our previous work, distributed diagnosis was studied based on an immediate observation passing (iop) protocol, where each local site transmits its observations to other sites immediately after each observation through bounded-delay channels. It was shown that the complexity of constructing local diagnosers is exponential in the number of local sites, which limits the applicability of iop-protocol based distributed diagnosis to distributed networks with a large number of local sites. To circumvent this problem, we propose in this paper a new distributed diagnosis protocol with online diagnosis complexity at each local site being linear in the number of local sites. A synchronized diagnosis estimate truncation strategy is proposed to further reduce the space requirement. It is shown that the new diagnosis protocol has the same diagnosis capability as the iop-protocol. An example is provided to illustrate the distributed diagnosis procedure under the new protocol