Output Regulation Of Multi-agent Systems Under Communication Constrains

In recent years,the field of distributed artificial intelligence has developed rapidly,and the theory of multi-agent systems（MASs）as an important branch has gradually become a hot topic,and received extensive attention from researchers in related fields.In the actual application process,MASs often need to consider tracking external signals or restraining the influence of the disturbances.This problem is also called the output adjustment problem.The output regulation problem is not only a tracking problem of the external source system,but also a problem of restraining disturbances.The external signal that needs to be tracked is generated by an external system and is widely used in practical work.Therefore,it is of great significance to study the output regulation problem.In the actual application of MASs,since it often works in complex and harsh external environments and each agent’s dynamical is also different,it is often caused by various factors.Inter-communication causes delay or data packet loss sent by the agent,which we call here communication constraints.Communication constraints may eventually lead to system instability,so in the process of theoretical research,the influence of communication constraints cannot be ignored.For this reason,it is necessary to design a suitable distributed control protocol to deal with this problem.First of all,for the heterogeneous discrete MASs,it is assumed that all agents can estimate the exogenous system and only a part of the agents can estimate the exogenous system,and the restriction conditions of system communication are given.Using the networked predictive control model,the compensation eliminates the influence of communication constraints and obtains predictive output information.The distributed predictive control protocol is designed and sufficient conditions are given to solve the output regulation problem.Finally,through simulation,it is verified that each agent can track external signals or restraining external disturbance.Secondly,with regard to how to solve the problem of packet coordinated output adjustment,through the establishment of a networked predictive control model,the communication time lag is actively compensated and the data packet loss problem is solved.Considering intra-group and inter-group communication and using predictive information to design a distributed predictive control protocol,it provides sufficient conditions to solve the problem of group cooperative output adjustment,so that each group of agents can track different external systems and suppress this.In the end,the effectiveness of the theory is verified by simulation,and the conclusion obtained by using the predictive model is compared with the output of the system in the unpredicted state,which proves that the predictive model can effectively reduce the influence of communication constraints.