| Cooperative control of multi-agent systems(MASs)means to achieve some control objective,such as consensus,through communication and data transmission among subsystems.And since communication and computation resources are often limited in practice,continuous information of neighbors should be avoided in control.One effective method to resolve this problem is event-triggered data transmission and control.Once an event-triggering condition is satisfied,communication and data transmission among agents will be triggered.And thus,only discrete-time information from neighbors at triggering instants are needed in control.In this paper,event-triggered consensus is studied for MASs with two different kinds of agent dynamics respectively.First,consensus of second-order linear MASs is studied via sampled-position-data-based event-triggered impulsive control under both directed and undirected communication conditions.Two kinds of impulsive controllers are designed,where only position data at triggering instants are used.Also,an event-triggering condition with a totally new structure is proposed,where only periodically sampled position data are monitored.Thus,velocity information can be avoided.Under this event-triggering scheme,the minimum length of triggering intervals is the sampling period and therefore,Zeno behavior is avoided naturally.And the maximum length of triggering intervals is set effectively and consequently,the phenomenon that a subsystem is not triggered for a long time is avoided.Then,sufficient conditions on control gains and lengths of triggering intervals are derived to achieve asymptotic consensus of both position and velocity states,where lengths of triggering intervals should be neither too large nor too small.Also,a practical method for selecting parameters is given to ensure that the sampled-data-based event-triggered scheme is truly realized.Further,some simulation examples are given to demonstrate the effectiveness of theoretical results.Next,leaderless and leader-following consensus of heterogeneous high-order MASs with nonlinear functions and uncertain parameters is studied via continuously-monitored event-triggered control.The consensus scheme consists of observers for consensus states,adaptive estimation laws for global information,event-triggering conditions,local controllers,and adaptive estimation laws for uncertain parameters.Relative information at triggering instants is utilized in observers and global information estimation laws.And only local information is employed in controllers and uncertain parameter estimation laws.Also,by using global information estimates in observers and triggering conditions,fully distributed control can be realized.And continuous relative information is detected in triggering conditions.To solve this problem,two kinds of self-triggered schemes are designed.The design idea of the first scheme is traditional,of which computation is complex;but that of the second scheme is totally new,of which computation stress can be relieved effectively.Then,asymptotic consensus is achieved for all states and Zeno behavior is excluded.Also,simulation examples are given to verify the effectiveness of theoretical results. |
