Fault detection for networked systems under periodic communication

This paper is concerned with the fault detection (FD) problem for a class of networked systems. We introduce a novel communication protocol which is called the Round-Robin protocol (RRP) to overcome the limited communication capacity. Under the RRP protocol, sensor nodes are prearranged in a particular sequence, and communication between sensor and the receiving end is thus periodic. Such kind of scheduling approach can effectively reduce the instantaneous communication burden. A set of zero-order holders (ZOHs) are used to store the received measurements. To generate the residual signal, a residual generator is constructed. The residual evaluation function and the threshold are designed. However, the periodical delay is inevitable in the RRP scheduling framework. Fortunately, complexity induced by the periodical delays can be greatly reduced by using the matrix blocking technique. Sufficient criteria ensuring the exponentially mean-square stability of the dynamics are established in the help of the Lyapunov stability theory. Design of the desired fault detector is reformulated as a convex optimization problem with linear matrix inequality (LMI) constraints. In the numerical example, the occurred fault is detected timely, which verifies the rationality of the investigated scheme.