Decentralized formation flocking for multiple non-holonomic agents

In this paper, we present a novel decentralized flocking algorithm to realize formation control of multiple non-holonomic agents travelling in two-dimensional space. The flocking mechanism consists of five essential properties: (1) collision avoidance; (2) alignment of their heading angles; (3) consensus on their thrusting speeds; (4) consensus on their rotational speeds; (5) minimization of the whole potential energy. By using the nearest neighbor rules, a distributed flocking algorithm is presented on the basis of the combination of consensus and attraction/repulsion functions. Furthermore, arbitrarily shaped formation control problem of multi-agent system can thus be solved by adding the information of the desired formation topology to the potential functions of the proposed flocking algorithm. Consider that the initial communication topology of the system is an undirected connected graph, the stability of the closed-loop system is proved by applying LaSalle-Krasovskii invariance principle. The connectivity of the communication network can be preserved throughout the evolution by means of the effect of the potential functions. Finally, the numerical simulation results validate that the proposed algorithm is effective.