The Research On Distributed Tracking Strategy Of The Descriptor Multi-agent Systems

Descriptor multi-agent systems,as the extension of general linear systems,are concerned by more and more scholars for the reason that the systems are more suitable for modeling and analyzing systems such as power,machinery and mobile intelligent vehicles.As the most basic and important issue,Consensus problem emphasizes the useing of local information obtained by individuals to design corresponding control strategies,so that each individual's state or output variables will be consistent as time went by.This problem is divided into two categories: with leaders and without leaders.The former is also known as tracking problem.By designing control strategy,the followers are consistent with the leader's state or output,which means the follower can follow the leader's track.The latter mainly designs control protocol that the state(or output)of all individuals come to be finally consistent.This paper mainly focuses on the tracking problem of descriptor multi-agent systems under the directed network,and the specific work and innovation are mainly included in several aspects as follows:First,the fault tolerant tracking problem of descriptor multi-agent systems is studied.Influenced by the industrial technology level,in the real environment,the actuator can not perform the task 100 percent according to the requirement,which leads to the lack of system control precision or even error.In this paper,the control algorithm of fault tolerance mechanism is considered,which is different from the existing work.This paper assumes that the error of the actuator is a variable of time and the upper bound of the error is unknown.A fully distributed control protocol is designed based on individual local information,and the coupling gain weight related to network structure can be self adjusted by adaptive law.Based on the Lyapunov theory,the stability of the closed loop system is proved strictly,that is,it is good to eliminate the influence of the actuator error.In addition,as a special case,the control protocol proposed in this part can be used to solve the fault tolerance tracking problem of the general linear multi individual system as well.Second,different from the general cooperative network structure,we study the tracking problem of continuous linear descriptor multiagent systems based on signed digraph(network weight can be negative).First,it is proved that the bipartite consensus of the descriptor linear system is equivalent to the conventional consensus.Based on the generalized Riccati equation,a distributed adaptive control protocol is designed to achieve the bipartite consensus of the systems.The consensus of the descriptor multi-agent systems is extended to the bipartite consensus,which extends the application scope of the descriptor system.Secondly,the special case of the system with nonlinear disturbance is considered,which shows that the proposed control protocol has good robustness.