Cooperative H∞ filtering for multi-agent systems

This article investigates the cooperative H
_∞
filtering problem of multi-agent systems. The filter input consists of relative states among neighboring agents and the individual states of a few agents. A topology graph is applied to represent the communication links among agents. The filtering error system is composed of the original system state, filter system state and the estimation error of relative performance variable among neighboring agents. Then by choosing a Lyapunov function candidate, sufficient conditions for the existence of filters in terms of linear matrix inequalities (LMIs) are derived, which guarantee the filtering error system asymptotically stable with an H
_∞
performance. By using the designed filters, signals of relative performance variables are estimated. An algorithm is proposed to calculate the estimation of the individual performance variable of each agent with the aforementioned estimated signals. Finally, the simulation results and detailed explanations are provided to illustrate the effectiveness of the designed filter and the proposed calculation algorithm.