| As an important model of control theory,simple-pendulum has been paid much attention by many scholars.In the era of booming artificial intelligence,control technology,and complex systems,the scientific research value of a single simple-pendulum is limited.Therefore,multiple simple-pendulum network systems are formed by connecting lots of simple-pendulum systems and communicating with each other.Each simple-pendulum in multiple simple-pendulum network systems has a certain execution capability,and can also realize the work of distributed coordinated control of the multi-agent system in the form of proximity interaction,and promote the multi-agent system to show high practical value in the aspects of multi-unmanned aerial vehicle formation,multi-manipulator cooperative work,integrated circuits,micro-grid economic scheduling,etc.However,in practical applications,due to network bandwidth and attacks,environmental interference,equipment restrictions and other reasons,some individuals in the multi-agent system always fail,and even lose contact or completely missing from time to time disconnection and loss may occur,seriously affecting normal operation of systems,and even causes systems collapse.Therefore,it is urgent to propose a consensus control method considering the failure or absence of part agents in multi-agent systems.In this research work,the multiple simple-pendulum network systems are taken as the research object,and two tracking control strategies,state feedback and output feedback,are studied respectively for the case of individual fault or absence in the system.The main research contents include:1.The distributed multiple simple-pendulum network systems model based on multi-agent systems is introduced,and state space description we need is obtained by using modern control theory.The potential fault individual is designed,and then the multiple simple-pendulum network systems with individual fault state is obtained as research object of this work.Meanwhile,algebraic graph theory,matrix theory and Lyapunov stability theory are introduced briefly to provide a theoretical basis for this study.2.A distributed adaptive consensus tracking control method with fault detection algorithm based on multi-agent systems is proposed for multiple simple-pendulum network systems with possible fault individual.Firstly,fault detection observers are designed,and an adaptive threshold detection method is proposed to detect the fault individuals accurately and effectively and deal with them in time.Next,utilizing the fully distributed adaptive technology without depending on the global information of communication topology,an adaptive consensus tracking control protocol based on state feedback and only using local information is presented,which aims to make the following simple-pendulums achieve the objective of tracking the leader before and after the failure,the stability proof is given together.Finally,numerical simulation and result analysis are carried out to illustrate the method can provide effective control both before and after the failure or loss of partial individuals.3.In practical multi-agent systems applications,there is a situation that the full state information of adjacent agent information interaction is unavailable.This research further adopts the idea of dynamic output feedback to solve the control problem of potential faulty individuals.Distributed observers are established for each following simple-pendulum,and the estimated states are used to design the fault detection threshold and the distributed adaptive consensus tracking protocol respectively.The proposed distributed output feedback adaptive coordinated tracking control method not only does not rely on the global information of the communication topology network and the state information of the intelligent agent,but also can effectively detect faulty individuals,And achieve consensus tracking objective for healthy multiple simple-pendulum network systems. |
PDF Full Text Request