Distributed Constrained Consensus of Multiagent Systems With Uncertainties and Disturbances Under Switching Directed Graphs

This article provides a distributed leaderless consensus control framework for nonlinear multiagent systems (MASs) with time-varying asymmetric state constraints, uncertainties, and disturbances under switching directed graphs. In such a framework, original constrained states of agents are first transformed into free states in a transformed state space. To deal with switching directed graphs, we drive agents toward consensus in the transformed space by leveraging a model reference control scheme, and it is sufficient that the original states reach consensus strictly subject to the time-varying constraints under mild assumptions. A single-layer neural network with weights adapted online is leveraged to approximate the uncertainties in agent dynamics. For external disturbances and reconstruction errors in the approximation, we introduce a robust term with an adaptive gain for compensation. Distributed consensus algorithms are proposed, respectively, for MASs with first- or second-order dynamics. We prove convergence to consensus via Lyapunov analysis and study the proposed algorithms' performance using numerical simulations.