摘要：AbstractRecently, the problem of robust diagnosis against intermittent loss of observations (RDILO) has been proposed in the literature, where a model for the plant subject to intermittent loss of event observations is presented, and fault diagnosability verification methods are proposed based on the new plant model. In this method it is assumed that some sensors of the system are reliable and are always capable of communicating their readings to the diagnoser, while the other sensors, or communication channels between sensors and diagnoser, are subject to intermittent failures. The case of unreliable communication of all observable events cannot be addressed using the RDILO since the model of the plant subject to intermittent loss of event observations also represents their permanent losses. In this paper, we formulate a different problem of robust diagnosis that considers only intermittent loss of observations, which allows considering the case that all communication channels between the plant and diagnoser are not reliable. The new formulation leads to a different notion of robust diagnosability, called K-loss robust diagnosability.
