Research On Multi-agent Formation Control Based On Consensus Theory And Artificial Potential Field

This paper mainly studies the consensus theory and formation control theory in cooperative control of multi-agent system.The multi-agent system seeks to complete the overall goal of each agent in an orderly and coordinated manner through a connected communication topology and a distributed control algorithm.The theory of consensus has evolved into a relatively independent field of research for a long time,and it is also the theoretical basis of many consensus derivative problems.The formation control problem is becoming more and more concerned by the researchers as the field which is closely related to the practical application,and is also the main subject of this paper.In this paper,the formation tracking and obstacle avoidance control in complex environment is realized by the combination of multiple formation control methods,and the flocking control of the large-scale multi-agent system is realized.The research results of this paper include the following aspects.Firstly,the linear first-order and second-order consensus algorithms are studied based on graph theory and matrix theory,and the necessary and sufficient conditions for the convergence of linear consensus algorithms are proved.On this basis,two kinds of consensus tracking algorithms are proposed,and the tracking of static signals and dynamic signals is realized respectively.In this paper,we design the first-order and second-order circuit structure from the electrical point of view and use the node potential as a state variable to simulate the linear first order and second order system consensus problem and signal tracking problem.The consensus protocol is established according to the law of current and voltage in the circuit,and the convergence condition of the consensus is analyzed from the angle of electrical law and energy.Secondly,the basic idea of artificial potential field method and the design method of artificial potential function are introduced.This paper discusses the causes and solutions of the local extreme value problem of artificial potential field method.In this paper,the artificial potential field method is used to realize the path planning of the agent under the obstacle condition.Based on the artificial potential field method,the formation tracking of multi-agents in the constraint space is realized by combining with the virtual structure method.The simulation results of the formation tracking,including formation response and the change of formation error is analyzed.We also study the flocking control of multi agent systems,based on velocity matching and artificial potential field method.The stability of the flocking is proved by choosing the proper Lyapunov function,and the flocking control and obstacle avoidance of 25 agents are realized by simulation.Finally,according to the basic concepts and related theories of fluid mechanics,a path planning method based on stream function is proposed.The method realizes the obstacle avoidance and path planning of the agent in the complex environment by simulating the flow of the fluid particle in the plane flow field.The problem of flocking of multi agents in flow field is studied by path planning method based on stream function.Compared with the artificial potential field method,the proposed method has smooth path planning,and has no local extremum problem,and has better formation simulation effect.