Robust Sliding Mode Consensus For Uncertain Multi-agent Systems

In recent years,due to the wide application of multi-agent systems,their related research has aroused great interest from many experts.In many research subjects on multi-agent systems,the consensus is the basis and premise for the coordinated control of multi-agent systems.It means that the agent communicates with each other through information interaction devices to achieve agreement with its neighbors.In practice,the working environment of multi-agent systems is often complex and changeable.Due to factors such as measurement accuracy,sensor accuracy,and environmental changes,there are various uncertainties in the systems.For example,the communication between agents may be affected by external disturbances.What's worse,there may be a shortage of network resources,or the agents may break down.In order to solve these problems,this thesis applies sliding mode control method to study the robust consensus problem of multi-agent systems.The main contributions of this thesis can be summarized as follows:(1)The robust sliding mode consensus problem of uncertain multi-agent systems under quantization mismatch is considered.First,the design method of the linear sliding surface is proposed,and the sliding variable parameter is solved by linear matrix inequality(LMI)technique;Second,according to the coding and decoding characteristics of the digital communication channel,the impacts of quantization parameter mismatch and external disturbances are fully considered,and a novel sliding mode reaching control law is proposed,which guarantees the designed sliding surface can be reached in finite time and the goal of quantized consensus is achieved.Finally,the effectiveness of the proposed method is verified via some simulation comparisons.(2)The robust sliding mode consensus problem of multi-agent systems driven by events is considered.The design of the guaranteed cost sliding surface is considered for a class of linear multi-agent systems with external disturbances.The parameters of the sliding surface are solved by LMI technique,which ensure the system states are asymptotically stable after the arrival of the sliding surface,and it also makes the sliding mode have certain robust performance.On this basis,the event-triggered mechanism is introduced.The event-triggered sliding mode controller is designed to reach the sliding mode band in finite time.Finally,the minimum bound of the trigger intervals is given to exclude the Zeno behavior.The numerical simulations show the effectiveness of the method.(3)The research contents of the previous chapter are further extended to the leaderfollower multi-agent systems.Besides the fault tolerance problem is considered at the same time.For a class of second-order leader-follower multi-agent systems subject to external disturbances,considering the possible partial failure of the actuator,a sliding mode control method based on event-triggered technique is proposed.First,the effect of introducing the event-triggered mechanism is analyzed,which the states cannot reach the ideal sliding mode but a sliding mode band.And the bandwidth of the sliding mode band is calculated.Second,combined with the idea of robust control and the boundary information of actuator failure rate,a distributed fault-tolerance controller is designed to make the multi-agent systems achieve the leader-follower consensus even in the case of partial failure.Finally,the simulation results verify the effectiveness of the method.