ABSTRACT
This chapter considers the consensus problem for multi-agent systems with each node being a general continuous-time linear system. A distributed static consensus protocol based on the relative states of neighboring agents is proposed, under which the condition for reaching consensus depends on three correlated factors: the agent dynamics, the design parameters of the consensus protocol, and the nonzero eigenvalues of the Laplacian matrix of the communication graph. In light of the consensus region methodology, the design of the consensus protocol can be decoupled into two steps: determine the feedback gain matrix in order to yield a desirable consensus region; adjust the coupling gain to deal with the effect of the communication graph on consensus. When the relative outputs, rather than the relative states, of neighboring agents are accessible, distributed full-order and reduced-order observer-type consensus protocols are proposed.
