Analysis and Control of Event-Based Decentralized Networked Coupled Systems Driven by Incomplete Information

Project: Research

View graph of relations


A decentralized networked coupled system (DNCS) consists of collaborating agentswhose behavior is coordinated through wired or wireless communication links. It providesa powerful tool to describe the dynamical behavior of many real-world engineeringsystems such as sensor networks and cyber-physical systems. Due to limited bandwidthof networks communication channels, it is necessary to save resources among agents inDNCSs. In addition, the demands for sharing resources efficiently in large DNCSs arecontinuously increasing. In many situations, there is a challenging conflict between theamount of transmitted data and the response time. Hence, efficient communicationamong agents is one of the key factors for DNCSs to achieve desired performance in acollaborative fashion.It is known that event-triggered control strategies have great potential in many cases tobe more efficient than conventional time triggered (or sampled) control. To optimize theuse of resource, it is desirable to design an effective event based sampling for DNCSssuch that the signals are transmitted only when it is necessary. In addition, the networksystems always suffers different types of networked signal transmission constraints.These constraints include limited communication capacity, network induced time delayand random packet loss. When these kinds of constraints occur, the distributed behaviorof the event based DNCSs often exhibits highly non-trivial dynamical characteristics.Hence, this will certainly lead to great reliability difficulties in controlling the networkperformance, and even cause instability of the network system.Our main aims in this proposal are(i) to study the coordination control problem in DNCSs;(ii) to develop a set of distributed event based algorithms that can be used by a group ofagents to carry out prescribed tasks efficiently under limited information transmissionconstraints. The approach will use a set of logical functions described the event-basedmodel and semi-tensor product matrices techniques for performing the analysis andcontrol;(iii) the proposed event-based coordination framework will also be applied to thesynchronization and performance analysis of power grid described by Kuramoto-likemodel.This proposal has set up three new, challenging yet achievable objectives to thetheoretical investigation. The results would have both theoretical and practicalsignificance in communication-based autonomous systems, in which actions must betransmitted to ensure a coordination behavior. In particular, the applications in thepower grid model will greatly benefit from the improvement of reliability of theproposed distributed event-based scheme.?


Project number9042223
Grant typeGRF
Effective start/end date1/12/1510/06/19

    Research areas

  • Control,stability,distributed ntworks,multi-agents,