Akademik Veri Yönetim Sistemi
American Control Conference (ACC), Georgia, Amerika Birleşik Devletleri, 8 - 10 Haziran 2022, ss.48-53, (Tam Metin Bildiri)
This paper contributes to the studies in control of multiagent networks as dynamical systems. Specifically, we consider a multiagent system that consists of floating nodes, a driver node, and a misbehaving node. All node types (i.e., floating, misbehaving, and driver nodes) exchange local information with their neighboring nodes through a fixed, connected, and undirected graph. In addition to local information exchange, the driver node also accepts a control signal constructed using its state information and the misbehaving (driver or floating) node is subject to a disturbance. Owing to the fact that discrete-time information exchange predicated on periodic time intervals is practically more feasible in control of multiagent systems as compared to their continuous-time counterpart, all the results of this paper are given in a discrete-time setting. In particular, we propose a proportional-integral control signal applied to the driver node in order to suppress the negative effect of the misbehaving node on the multiagent system. We first present a system-theoretical approach to obtain the steady-state values of each node. We then present a graph-theoretical approach for undirected tree graphs and show the explicit steady-state values of each node, where this approach provides insights on how to select the driver node to minimize the largest steady-state error and maximize the number of nodes with zero steady-state error. Several illustrative numerical examples are also given to elucidate the efficacy of our contributions.