Distributed leader selection

In this paper, we address the problem of distributed leader selection in a network of agents. We assume that each agent´s state is defined by a scalar that evolves according to a linear dynamics involving the states of its neighbors and its own state. We propose a distributed algorithm, assuming a time-invariant communication graph, to determine: (i) those agents that should behave as leaders (i.e., agents whose state serves as a reference (or input) to the remaining agents - the followers); and (ii) a communication protocol to ensure some control-theoretical specifications and/or performance guarantees. Finally, an illustrative example using the main results of the paper is provided.