Finite-time Formation Control For Multi-agent Systems

Along with the development of Artificial Intelligence(AI)technology,especially with the widely applied technologies of communication,sensing,positioning and image identification,the unmanned autonomous robot comes to our sight more and more.In practice,because we usually need multiple robots to perform a task synergistically.Therefore,each autonomous robot is generally regarded as an agent,and the multi-autonomous robot system is simplified into multi-agent system.The traditional multi-agent formation control is generally based on the precise information of the agent to design an asymptotically stable controller.In practical applications,it is difficult to obtain accurate information.Therefore,the third chapter of this paper studies the problem of finite-time formation control based on coarse distance information,so that the agent can still complete the formation task even if it can only perceive coarse information.Compared with traditional asymptotically stable controllers,finite-time controllers have better robustness and can be applied to high-precision control systems.In addition,the application of the controller to the minimum persistent formation can realize the control of large-scale formation.Finally,inspired by the behavior of bird formation switching leaders in nature,we study the minimum persistent formation generation algorithm that can switch leaders,and improve the robustness of the formation by switching topologies.Chapter 4 studies the formation control problem based on local coordinate information,and uses the rooted graph as the topological structure,and then designs a finite-time formation controller.In this control scheme,multi-agents do not need to share the coordinate system,nor do they need to have the same direction.This paper gives the design of two finite-time controllers in Chapters 3 and 4.Afterwards,an extension is given in Chapter 5,that is,considering both convergence time and energy consumption,a controller based on linear quadratic method is designed.In such a controller,the convergence time and energy consumption can be adjusted by adjusting the parameters in controller to meet the needs in actual engineering.Thus,a more practical and flexible control scheme is provided for cooperative tasks of multiple autonomous robots.