Consensus of generalized integrators: Convergence rate and disturbance attenuation property.

We study a generalized integrator consensus network, each node of which is a single-integrator cascaded with some heterogeneous non-integrator internal dynamics. Under the assumption that the interconnection graph is undirected and connected, we first investigate the convergence property of the consensus network and show that the interconnection is beneficial for the enhancement of the convergence rate if the non-integrator internal dynamics of each node is strictly passive. We then show that the interconnection is also advantageous for the disturbance attenuation in the consensus network.