Output-Feedback-Based Consensus Control Without Continuous Interagent Communication

• Published in: IEEE transactions on systems, man, and cybernetics. Systems Vol. 51; no. 5; pp. 3240 - 3250
• Main Authors: Shih, Chung-Shan, An, Han-Hsuan, Cheng, Teng-Hu
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.05.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Closed loops; Communication; Control methods; Convergence; Event-triggered control; intermittent communication; limited communication; multiagent systems; Network topologies; Network topology; Observers; Output feedback; Stability analysis; State feedback; Switches; Switching sequences; time-varying network topologies; Time-varying systems
• ISSN: 2168-2216; 2168-2232
• DOI: 10.1109/TSMC.2019.2919738

A decentralized event-triggered approach is developed for leader-follower systems with time-invariant and time-varying network topologies in the absence of full-state feedback. Continuous interagent communication is eliminated by communicating information only during events. At the event times, an observer communicates with the neighboring agents to obtain update information. During the interevent intervals, an open-loop estimate is used as a substitute for local feedback. The advantages of our novel approach are that: 1) the amount of communication required can be reduced and determining the event times requires no interagent communication, with the event-triggered condition designed using stability analysis so that bounded consensus is guaranteed; 2) bounded state consensus can be achieved without receiving full-state information from the followers; and 3) the switching sequence of the time-varying network topologies can be arbitrary, since a switching controller is developed based on the common-Lyapunov-function approach to ensure state consensus. Intermittent feedback is present and the dynamics of the time-varying network topologies are discrete, which together prove that Zeno behavior never occurs. The stability of the closed loop is analyzed to guarantee bounded convergence for the designed control methodologies. The results of simulations are presented that demonstrate the efficacy of our novel controllers.