Research On Group Cooperative Control Of Multiple Agents

In recent years,cooperative control of multi-agent systems has gradually become a hot subject in control theory,and has found widely applications in many fields,such as UAV formation,playing-soccer robots in groups,marine fleet combat groups.With the development of informatization and modernization of industry,real problems are becoming more and more refined.It is necessary to divide tasks according to the types and requirements of tasks and design corresponding control protocols or algorithms so that agents can exchange information in a way to achieve a certain formation or track a state.At present,the theory on consensus of multiple agents is nearly perfect,but the research on cooperative control of multiple agents is relatively less.This paper mainly uses graph theory and matrix theory as analysis tools to study group formation control and group tracking control of multiple agents.The main contents are as follows:1.Research on group formation control: Based on the existing second-order consensus protocol,the scaled consensus algorithm is proposed and the convergence state and the conditions on convergence of the system are determined by analyzing the eigenvalues of the system matrix;Besides,based on the scaled consensus algorithm,we proposes a formation control algorithm which guarantees the agents reaching some formation.2.Research on group tracking control: For the case of the constant reference state with two sub-groups or q sub-groups,we obtain that multiple agents track the reference state if the information exchange topology has a directed spanning tree with the tracking state as the root node;For the case of time-varying tracking reference state with the similar tracking protocol to the constant case,it is derived that multiple agents track the reference state if the topology has a spanning tree,and the reference state and each agent are directly connected.