Font Size: a A A

Investigation On Distributed Consensus Algorithm

Posted on:2015-06-09Degree:MasterType:Thesis
Country:ChinaCandidate:Y Y XiongFull Text:PDF
GTID:2298330422991969Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
Distributed coordination of multiple agents has attracted in-depth investigations byresearchers from various scientific communities including physics, biology,mathematics, mechanics, computer science, and control theory, among others. Thereason for this broad range of research attention is the vast practical applicationsdistributed coordination of multiple agents has in such areas as formation control,distributed networks and distributed computation.The importance of this coordination strategy in practice requires betterunderstanding and refinement of key aspects of the theory behind, and one such aspectis the problem of consensus. Central to this problem is a set of suitably designedprotocol and algorithms which will allow individuals within a particular system to makedecisions based solely on information of themselves and their neighbors, thus achievingconsensus of information between individuals under circumstances such as limitedinformation, unreliable information exchanges, communication delays or input delays,and dynamic variations of communication topologies. The type of consensus problemthis thesis deals with is the consensus of second-order multi-agent systems under fixedtopologies and switched topologies, respectively.For the case of fixed topologies, based on current work in the literature, this thesisconsiders two scenarios, namely the consensus problem under directed and undirectedcommunication topologies, respectively. In both scenarios, it has been proved that theconsensus algorithms will enable the system to reach consensus under given conditions.Additionally, an analysis is provided for the relationship between parameters takingcertain values and existing results. The simulation that follows validates the correctnessof the conclusion.For the case of switched topologies, by incorporating existing consensusalgorithms and relevant theories, this thesis proposes to discretize the aforementionedcontinuous multi-agent systems through sampling, and then studies the consensusproblem in the switched communication topologies. Conditions for the system toachieve squared average consensus are then given, under both directed and undirectedswitched topologies. The correctness of the obtained results have been confirmedthrough simulations.Finally, the concluding remark summarizes this thesis.
Keywords/Search Tags:Multi-Agent System, Distributed Consensus, Switching Topology
PDF Full Text Request
Related items