Improvement of consensus convergence speed for linear multi-agent systems based on state observer

A novel consensus protocol is presented for linear multi-agent systems(MASs) in which all the agents have the same continuous-time linear dynamics with high dimension. A fast convergence speed can be obtained by using this protocol for the MASs where each agent gets its own state through the full-order observer. Every agent can also obtain the states of its neighbors׳ and its instantaneous neighbors׳ to reach consensus. Necessary and sufficient conditions of consensus for the MASs are proposed to improve the convergence speed. Finally, the corresponding convergence speeds of different protocols are showed by numerical simulation, and extended to directed topology.