Fault estimation of multi-agent systems with transient performance improvement

In this paper, the fault estimation of multi-agent systems with transient performance improvement is fully investigated. Based on the traditional Luenberger estimator, a necessary and sufficient condition for achieving a set of fault estimators considering the H_∞ index with transient performance is obtained which requiring only measurement information of the relative output among neighbouring agents. In order to isolate the estimated faults, a double-level network structure is introduced, and the faults are reconstructed by a common supervisor. The simulation example is performed to show the superiority of the proposed fault estimation approach.