Consensus of Multiple Lur’e Systems for Directed Communication Graphs with Distributed Adaptive Relative Output Feedback Protocol

In this paper, the consensus problem of multiple Lur'e systems over directed graphs is considered, where only relative output information between neighboring agents is used to generate the distributed protocols. The main difficulty lies in the coupling between the design of distributed observer and adaptive protocol. To overcome such difficulty, the idea of unknown input observer is borrowed to propose distributed observers, based on which a fully distributed adaptive protocol is presented. The effectiveness of the proposed adaptive protocol is verified by numerical simulations. The adaptive protocol is designed in a fully distributed way, depending only on relative output information between neighboring agents, without using the information of the global communication graph.