Advanced Approaches to Multiagent Coordination
Magdi S. Mahmoud,Yuanqing Xia
DOI: https://doi.org/10.1016/b978-0-12-823698-7.00015-1
2021-01-01
Abstract:In this chapter, we deal with four problems. First, we consider the synchronization analysis and control problems for a class of nonlinear stochastic complex dynamical networks (NSCDNs) consisting of identical nodes and account for asymmetric coupling configurations, nonlinear inner coupling structures, and nonidentical exogenous disturbances. A notion of quasisynchronization in probability is introduced to assess the synchronization performance of the addressed NSCDN. An event-triggered pinning feedback control strategy is adopted to control a small fraction of the network nodes to reduce the frequency of updating and communication in the control process while preserving the desired dynamical behaviors of the controlled networks. A general framework using Lyapunov analysis is established within which the problems of dynamics analysis and controller synthesis are solved for the closed-loop stochastic dynamical networks. Second, the consensus of multiagent systems with linearized dynamics is considered based on an observer-type protocol using the relative outputs of neighboring agents on a directed communication topology. The consensus problem is cast into the stability of a set of matrices with the same low dimension as that of a single agent. For neurally stable agents, it is shown that there exists an observer-type protocol having a bounded consensus region in the form of an open unit disk, provided that each agent is stabilizable and detectable. For unstable agents, an algorithm is proposed to construct a protocol having an origin-centered disk as its consensus region and radius to further satisfy a constraint related to the unstable eigenvalues of a single agent for the case where each agent has a least one eigenvalue outside the unit circle. Third, the tracking problem for multiagent systems with a time-varying reference state is investigated. The tracking control is designed based on the event-triggered scheme, which can lead to a significant reduction of the information communication burden in the network. To realize the tracking control, only the communication between the agent and its local neighbors is needed; therefore, the designed control is essentially distributed. Fourth, the robust output regulation of linear multiagent systems under a directed interaction topology is addressed. The digraph is assumed to contain a spanning tree. Every agent or subsystem is identical and uncertain, but subsystems have different external disturbances. Based on the internal model and the general discrete-time algebraic Riccati equation, a distributed consensus protocol is proposed to solve the regulator problem.