Trackability-Based Cooperative Learning Control for Heterogeneous Networks

This paper investigates the cooperative learning control problem for heterogeneous networks, where the cooperative perfect tracking objective is of specific interest. The cooperative trackability is first proposed for heterogeneous networks to reveal the cooperative tracking ability of agents subject to the given desired reference, for which a distributed cooperative trackability criterion is established through the interactions among agents regardless of the heterogeneous dynamics of agents. Given the cooperatively trackable desired reference, a cooperative learning control law is developed with the insertion of the complete experiences of agents at the previous iteration. In the presence of the quasi-strong connectivity, the cooperative perfect tracking objective can be realized for heterogeneous networks with agents having irregular dynamics.