A framework for robust synchronization in heterogeneous multi-agent networks

A general framework for the analysis of networks of heterogeneous agents is presented. By modeling each agent as a nominal linear time-invariant system plus a possibly nonlinear perturbation and the interconnection among the agents via a memoryless operator, a general result is offered which ensures robust synchronization of the network to a subspace of ℝ^N. The result is applied to the case when the interconnection operator is a constant normal matrix and the perturbations nonlinear. The criterion is reduced to a graphical Popov criterion for the synchronization of the network.