Research On The Impulsive Consensus Of Switched Multi-agent Systems

With the rapid development of information technology nowadays,the model of multiagent systems has attracted the attention of scholars in various fields for its high controllability and strong coordination ability,and its research results have been continuously applied to more and more extensive projects,for example,coordinated control of actions between smart robots,energy optimization of smart grid,and drone formation.On the other hand,due to the different needs of tasks or the influence of various unfavorable factors,such as the change among system operation modes,the failures of systems components,the interference of external environment,etc.,the switching phenomenon can be seen almost everywhere in a variety of practical systems.For this reason,the introduction of various complex properties such as switching dynamics into multi-agent systems has attracted great attention from researchers and has gradually become a hotspot in the field of control.As a result,based on the framework of multi-agent systems,focusing on the effects of system evolution caused by the interaction of discrete events that switches and impulses,this dissertation try to relax some restrictive constraints in the existing results,and to explore some factors that affect the consensus of switched multi-agent systems with the knowledge of stochastic process,Lyapunov stability theory,the constraints on dwell time and impulsive interval.The main research results and contributions of this dissertation are presented as follows:1.A class of nonlinear stochastic switched leader-following multi-agent systems,in which the switching signal satisfies the continuous-time semi-Markov process with its transition rate all or partly known,are investigated separately.the concept of the continuous-time semiMarkov kernel is introduced,which realizes that the dwell time can subject to different types of distributions.Then,some sufficient conditions are derived to achieve the consensus of the considered system with the proposed impulsive protocol.Based on the derived results,the transition rate with partly and even all unknown is further considered.By means of two different kinds of Lyapunov functions,sufficient conditions dependent and independent of the transition rate to achieve consensus are derived respectively.2.The average consensus problem of the hybrid stochastic switched multi-agent systems is studied.Taking the diversity of practical systems and the interference of stochastic noise,the existing preliminary hybrid multi-agent system model is improved.The novel model contains a class of hybrid subsystems that subject to distinct dynamics are proposed,and it realizes the possibility of arbitrary switching between subsystems with different characteristics.In the meanwhile,some sufficient conditions for achieving the average consensus of the considered system with asynchronous impulses and switches are given,where the stochastic analysis technique,and two general restrictive constraints combing average dwell time and average impulsive interval are applied.3.Taking both external interference and stochastic noise into consideration,the quasiconsensus problem of a class of nonlinear time-varying stochastic switched multi-agent systems is investigated.The stochastic noise and external inputs maybe coexist in the practical proj ects due to the complex environment.Under such an assumption,a reasonable hypothesis on the time-varying matrices of systems is introduced,and by means of the stability analysis technique of time-varying systems,the restrictive constraint is removed compared with the existing results.At the same time,two different impulsive protocols are proposed,acting on some distinct agents and a single agent at each impulsive instant,respectively.Finally,with the help of the proposed protocols,knowledge of probability theory,and the reasonable hypothesis,some sufficient conditions for the considered systems to achieve the quasi-consensus are given.