Font Size: a A A

Research On Formation Control Methods For Multi-agent

Posted on:2013-07-07Degree:MasterType:Thesis
Country:ChinaCandidate:Y WangFull Text:PDF
GTID:2268330392969348Subject:Control Science and Engineering
Abstract/Summary:PDF Full Text Request
With the rapid development of artificial intelligence, information andcommunications technology and so on, it provide technical support for the multi-agentsystem, multi-agent formation have become a hot topic in current academia. In Multi-agent formation problem, a particularly important problem is how to make a Multi-agent formation form and maintain to the desire formation structure from any initialposition. Based on this problem, in this paper, we consider only use the information ofthe relative position between agents to keep the relative distance between the agents tocomplete the formation. The main works are as follows:Firstly, consider the formation control of the speed of the global leader converges tozero exponentially fast, in which the followers can only measure the relative positionbetween leader and follower. There are more than one formation structures meet thesame distance constraint under the premise of the minimally rigid, but in which thereonly one formation structure is our desired. We design the formation control law forevery follower(contains the first follower), under the control law action, the formationconvergence to the desired formation structure if the initial positions of global leaderand first follower do not coincide, and other agent’s position random. Simulation resultsillustrate the effectiveness of the proposed formation control approach.Secondly, consider the formation control of the acceleration of the global leaderconverges to zero exponentially fast, in which assume the first follower can measure thespeed of global leader and the relative position between the global leader and itself, theother followers measure the relative position between leader and follower. We designthe formation control law, under the control law action, the formation convergence tothe desired formation structure if the initial positions of global leader and first followerdo not coincide, and other agent’s position random. Simulation results illustrate theeffectiveness of the proposed formation control approach.Finally, We analyze the asymmetric formation and symmetric formation via practicalpoint of view, and propose the asymmetric formation is more suitable for the formationcontrol. This paper also discusses the necessary and sufficient conditions for minimallyrigid formation stably rigid, which is base on only use the information of the relativedistance between agents to keep the relative distance between the agents.
Keywords/Search Tags:Multi-agent system, formation control, minimally rigid formation, leader, follower
PDF Full Text Request
Related items