Localized data-driven consensus control for continuous-time multi-agent systems

This article proposes a localized data-driven consensus framework for leader-follower multi-agent systems with unknown continuous-time agent dynamics for both noiseless and noisy data scenarios. In this setting, each follower calculates its feedback control gain based on its locally sampled data, including the states, state derivatives, and inputs. We propose novel distributed control protocols that synchronize the distinct dynamic feedback gains and achieve leader-follower consensus. Design methods are provided for the devised data-based consensus control algorithms, which rely on low-dimensional linear matrix inequalities. The validity of the developed algorithms is demonstrated via simulation examples.