Matching Based Formation Control and Analysis of Large-Scale Multi-agent Systems

This paper is concerned with formation control and stability analysis for multi-agent systems. The individual sensing range is taken as limited and agents are assumed not able to communicate. The concept of extendible periodic formation is established to represent a class of large-scale systems and later used in the proof of stability. In the control strategy, the periodic formation is first partially reached with some misalignments. Then, in order to eliminate the misalignments, a multi-shape matching algorithm is designed to identify them and reach the predefined formation. The concept of group control edge is proposed, whose different types are analyzed. It is shown that under optimal sizes, the convergence rate could be maximized. Moreover, a premature stop strategy is presented that could reduce computation burden. Analysis is made on numerical examples to demonstrate the performance and merits of the proposed method.