Analysis and design of distributed protocols for multi-agent systems subject to input saturation and time-varying delays

• Autor(a) principal: Thales Costa Silva
• Data de Publicação: 2021
• Tipo de documento: Tese
• Idioma: eng
• Título da fonte: Repositório Institucional da UFMG
• Texto Completo: http://hdl.handle.net/1843/37865
• topic: Multi-agent systems; Lyapunov-Krasovskii; Saturation; Delays; Euler-Lagrange systems; Consensus; Engenharia elétrica; Liapunov, Funções de; Lagrange, Equações de

This work deals with the problem of coordination of multi-agent systems, with interconnected systems subject to time-varying delays and input saturation. The proposed approach consists in deriving sufficient conditions for the asymptotic convergence of the network, based on the Lyapunov-Krasovskii framework. Two tasks are tackled in this manuscript: the consensus and the formation-containment. The consensus is investigated in the discrete- and continuous-time domains, considering directed networks composed of agents modeled as linear systems of arbitrary order, subject to input saturation and time-varying delays. The main results are sufficient conditions for consensus analysis and the design of feedback gains for distributed consensus protocols for multi-agent systems. Besides, in this context, it is also proposed a strategy to compute a region from which the consensus is always guaranteed. On the formation-containment problem, the node dynamics of the network are extended to distinct systems modeled by the Euler-Lagrange equation, also subject to bounded inputs and time-varying delays. The main results on the formation-containment are the design of new control algorithms and sufficient conditions to ensure the convergence of the network. The control algorithms are designed as distributed dynamic controllers, in such a way that the number of neighbors of each agent is decoupled from the bound of the control inputs. Finally, the utilization of the proposed methods is illustrated through numerical examples, and it is shown that they can improve related recent works from the literature that deal with similar issues.