Consensus of Higher Order Agents: Robustness and Heterogeneity

This paper explores the use of Kharitonov's Theorem on a class of linear multiagent systems. First, we study a network of the mth order (m ≥ 2) linear uncertain interval plants and provide conditions for achieving fullstate consensus, which relate the stability margins of each agent to the spectrum of the graph Laplacian. Then, a robustness analysis for such systems is presented when an edge weight in the underlying graph is perturbed. The same Kharitonov-based analysis proves useful in a related problem, where heterogeneous higher order linear models of agents are considered in a setup similar to pinning control, and conditions for consensus among the follower agents are derived. Numerous simulation examples validate the results.