Consensus Of Multi-Agent Systems Under Several Different Control Strategies

Since the coordination control of multi-agent systems has been widely used in biological systems,network robots and UAV systems,it has attracted increasing number of attention.In the research of multi-agent systems,the consensus means that each agent only interacts with its neighbours and uses local information gathered from neighbouring agents to achieve a common state or a global behaviour.The objective of the consensus control is to design a distributed control protocol such that all agents reach a common state using the local information.This paper mainly studies the consensus of multi-agent systems based on low gain control,guaranteed cost control and observer-based control.The first part studies the high-order consensus of multi-agent systems with feedforward nonlinear and a time-varying communication delay in a directed network.First,a low gain control protocol is designed to eliminate the effects of the potential nonlinearity and the bounded time delay on the system.Moreover,by using Lyapunov stability theory and Newton-Leibniz formula,some sufficient conditions for the system to achieve consensus are given.It is found that,for any upper bound time-varying delay and strongly connected diagraph,the proposed controller can solve the consensus problem of multiagent systems with feedforward nonlinearity if the designed parameter ? is great than the threshold value.The second part investigates the leader-following guaranteed-performance consensus problem for second-order multi-agent systems.First,for the case without the time delay,using the state decomposition method,the influence of coupling gain and the pinning strength between agents on consensus are carefully considered,the conditions for the system to achieve consensus are given,the upper bound of the cost function is determined.Moreover,for the case with time delay,combining with Jensen inequality and NewtonLeinz formula,the upper bound of the cost function and the relevant criteria for the system to achieve consensus are obtained.The third part addresses the observer-based distributed consensus of linear multiagent systems subject to input saturation.First,the observing control protocol is designed to estimate the state of the agent,and the saturation restriction in the control input is removed based on the low-gain feedback method,The consensus condition of the multiagent systems with directed strongly connected topology is developed by using Ricatti equation.Moreover,applying the scheme of tree type transformation,the consensus of multi-agent systems with general directed spanning tree networks is studied.Finally,the method is extended to the general switching topologies.It is found that the consensus can be achieved if the communication times of the system with and without the directed spanning tree topology satisfy some conditions.