Research On Event-Triggered Consensus Control Of Multi-Agent Systems

With the continuous breakthroughs in the fields of computer and communication technology,the multi-agent systems and its control problems supported by these two fields and modern control theory have received tremendous attention,and the relevant theories and applications have developed rapidly.Consensus,as the most fundamental issue in the control of multi-agent systems,aims to achieve the agreement of states/outputs of all agents through local information interaction.In the traditional consensus control,all instruction executions,including system state sampling,signal transmission,controller adjustment,etc.,are accomplished by the microprocessors equipped by the agents.Such embedded highperformance microprocessor has the disadvantages of high cost and large energy consumption.On the other hand,the bandwidth limitation of the actual communication network cannot achieve high frequent information interaction.In order to decrease information interaction and reduce the requirements on microprocessors,the event-triggered control has been introduced into the research of consensus issues,which can significantly reduce the information transmission,decrease the update of the controller,and achieve the system performance while saving communication/computation resources.Moreover,in practical applications,the system is vulnerable to external interference,communication failure and other factors,which pose new challenges to the construction of distributed event-triggering mechanisms,the design of consistency protocol design,and closed-loop stability analysis in event-triggered control.To tackle these issues,this paper follows the route of system model construction→control algorithm design→closed-loop performance analysis→simulation example verification.For the multi-agent system under the influence of objective factors such as stochastic disturbance,state constraints,network attacks,etc.,the control strategies based on static/dynamic eventtriggering mechanism are constructed,and the underlying consensus control problems are deeply researched by taking advantage of tools such as stochastic analysis,stability analysis,adaptive theory and so on.The main research contents are summarized as follows:(1)For the event-triggered consensus problem of multi-agent systems with stochastic disturbance,the system is modeled as Ito-type stochastic differential equation,and a new static event-triggered control strategy is constructed based on the relative measurement information between neighboring agents,in which the event-triggering mechanism completely excludes the infinite fast execution behavior by enforcing a strictly positive lower bound of the interexecution times.Based on this,a general stochastic convergence analysis technique is developed to achieve the almost surely consensus and mean square consensus.Such analysis technique provides a new route to analyze the event-triggered convergence of linear/nonlinear stochastic multi-agent systems in addition to Lyapunov theorem.Finally,simulation examples and comparative experiments are given to illustrate the effectiveness of the proposed eventtriggered control strategy and its advantages in reducing controller update frequency.(2)For the event-triggered consensus problem of multi-agent systems with positive state constraints and switching topology,the switching parameters of topologies are obtained according to the average dwell time approach,which ensures that number of topology switching is limited in a finite time.Based on this,a static event-triggered control protocol is designed to reduce the frequency of controller updates.Then,by combining the positive system theory and the multiple Lyapunov functions,the closed-loop performance analysis is accomplished.It is proved that the proposed piecewise constant controller has enough feedback ability,which can not only ensure that the states of all the agents stay in the non-negative quadrant and can reach consensus finally,but also clarify the quantitative relationship between the convergence rate and the switching parameters.Finally,a simulation example is provided to illustrate the validity of the proposed control strategy.(3)For the event-triggered guaranteed cost consensus problem of multi-agent systems with positive state constraints and multiple delays,a more general topology switching sequence is constructed based on the mode-dependent average dwell time approach,which ensures that each topology has its own average dwell time.Based on this,a static event-triggered control protocol is designed.By utilizing the multiple Lyapunov-Krasovskii functional method,the time-varying delay and distributed delay are handled,and the delay-dependent sufficient conditions are derived to achieve positive consensus.Furthermore,based on the quadratic cost index function,an event-triggered guaranteed cost controller is proposed.It realizes a trade-off between the consensus regulation performance and the control energy consumption,and reveals the relationship between the upper bound of the cost value,the initial condition of the system and the topology switching signal.Finally,the advantages of the proposed control algorithm in designing switching signal and reducing control cost are verified by simulations.(4)For the guaranteed cost event-triggered secure consensus problem of multi-agent systems with multi-channel DoS attacks,a classification strategy is proposed,which categorized the DoS attacks as three attack modes according to the attack strength,then the corresponding system is modeled as a switched system by means of the switching thought.Based on this,a less conservative event-triggered secure protocol is designed to ensure the system performance,which can switch the corresponding feedback gain according to various attack modes,and has the characteristics of saving communication resources and resisting attacks.Furthermore,a guaranteed cost control strategy is introduced to establish the guaranteed cost secure consensus criterion,and the relationship between the control cost,the convergence rate and the attack parameters is obtained.Finally,the advantages of the proposed algorithm in saving communication resources and reducing control costs are illustrated by comparison and simulation.(5)For the dynamic event-triggered finite-time H∞ consensus problem of multi-agent systems with multiple cyberattacks,the system model is transformed into a networked switching system to appropriately characterizes the impact of DoS attacks and deception attacks on the system.Then,based on the dynamic event-triggering mechanism and security defense method,a finite-time secure consensus protocol is designed,which enjoys the advantages of alleviating the computation/transmission burdens and resisting cyberattacks by adaptively adjusting the trigger threshold,and finally achieve consensus in a finite time interval.Additionally,to suppress the influence of external disturbance on consensus regulation performance,an event-triggered H∞ consensus criterion is established,which achieves secure consensus in finite-time interval while guaranteeing a prescribed H∞ performance index.Finally,the advantages of the proposed algorithm in saving control resources and resisting external interference are confirmed by simulations.