Research On Event-Based Control And Its Applications

In network control systems,controller should be built on the digital platform due to the limitation of the communication network,online resources and the communication ability of the agent.Event-based control,which can reduce the controller update and the communication resources,has been proposed for the network control systems.On the other hand,because of the wide application of impulsive control in practical problems,it has aroused the focus of researchers.Most of the existing works focused on the case that the impulsive instants are predetermined.The frequency of impulsive control has to be designed fast enough by taking the worst scenario into consideration and thus tends to be conservative.In recent years,the event-based consensus of multi-agent systems has received widely concerned from various fields.However,in order to determine the parameters in the controllers and the triggering functions,the Laplacian matrix associated with the communication system,which is a global information,is required in most of the existing results.This means that these results are not fully distributed.Adaptive control can solve this problem.Based on the discussion above,the main points of this thesis are summarized as follows:Firstly,the exponential stabilization of continuous-time dynamic systems(CDSs)is investigated via event-based impulsive control(EIC)approach.The global exponential stability criteria via EIC are derived for nonlinear and linear CDSs,respectively.The impulsive instants are determined by certain events.Besides,the developed event-based impulsive scheme is applied to the synchronization problem of master and slave memristive neural networks.Furthermore,a self-triggered impulsive control scheme is developed to avoid continuous communication between the master system and slave system.Secondly,the leader-following consensus of multi-agent systems with linear models is investigated.The event-based impulsive controller and the triggering function based on the states of the agents are designed.The impulsive instants of agents are determined by the triggering condition and the triggering function,that is,the impulsive effect performs only at the instants when the events occur,and so as the controllers update.The sufficient conditions on leader-following consensus are proposed,and the Zeno behavior is also excluded.Thirdly,adaptive event-based consensus of multi-agent systems with general linear dynamics is considered.A novel adaptive event-based controller and a state-dependent triggering function are proposed for each agent.The sufficient conditions on consensus are proposed.Furthermore,the Zeno behavior of the concerned closed-loop system is also excluded under certain conditions.Finally,the adaptive event-based control approach is applied to study the leader-following consensus of multi-agent systems with linear dynamic models.Adaptive event-based controller and triggering function are designed for each agent,where the adaptive function is only dependent on its own event time instants.A sufficient condition on consensus is proposed.Furthermore,the Zeno behavior of the concerned closed-loop system is excluded.