A cooperative inverse optimal problem for multi-agent systems with output feedback

This paper investigates an inverse optimal problem for multi-agent systems with general linear dynamics. It is demonstrated that, given a stabilizing control input which only uses output information, a performance index function can be found such that the control input is optimal with respect to it. The communication graph of the agents is assumed to be undirected and fixed. A structured quadratic performance index is derived that captures the topology of the graph, which allows for global optimal control that is implemented using a distributed observed based protocols.