Leader-formation control with dynamic interaction topologies

Coordinate control of a multi-robot system is a hot research topic in recent years, since a tough task is usually performed by cooperation of multiple robots. A typical issue of coordination is formation, which requires a leader robot and its follower robots to construct certain types of patterns. A practical issue is addressed if the sensor and communication abilities are limited such that robots can only handle local information, how a group of robots form a formation. For an individual robot control, the local control strategy is applied based on dynamic model of robots with parameter uncertainty and adaptive NN methodology. The macroscopic performance of formation or stability of formation is studied, which includes invariant and variant interaction topologies. Using the local control strategy, it has been proved that the robots can form steady formation structure even if interaction topologies change over time as long as adjacency graph keeps connected.