BRB-based distributed fault diagnosis for leader-following consensus of multi-agent systems under noisy measurement

In this paper, the belief rule base (BRB) based distributed fault diagnosis problem is studied for leader-following multi-agent systems (MASs) with measurement noises, where each agent can only measure the outputs of its neighbors in a noisy environment. A novel distributed protocol based on the relative measurement outputs is developed to achieve leader-following output consensus of MASs without fault in stochastic sense. Then, on the basis of the proposed distributed output feedback controller, novel distributed BRB-based fault diagnosis models are developed for each agent by using the measurement output information of neighbor agents. It should be noted that the proposed BRB based distributed fault diagnosis models can diagnose the fault and meanwhile eliminate the influence of the unpredictable measurement noises in engineering practice. Finally, a simulation example is provided to illustrate the effectiveness of the proposed method.