In this article we propose a hierarchical control structure for multi-agent systems. The main objective is to perform formation change manoeuvres, with guaranteed safe distance between each two vehicles throughout the whole mission, even if motion is restricted to a limited space and static obstacles are present. The key components that ensure safety are a robust control algorithm that is capable of stabilising the group of vehicles in a desired formation and a higher level path generation method that provides all the vehicles with safe paths, based on graph theoretic considerations. The method can efficiently handle a large group of any type of vehicles. As an illustration, the results are applied to a group of 18 quadrotor UAVs, where 2 of the UAVs cannot receive information from the others.

In this article we propose a hierarchical control structure for multi-agent systems. The main objective is to perform formation change manoeuvres, with guaranteed safe distance between each two vehicles throughout the whole mission. The key components that ensure safety are a robust control algorithm that is capable of stabilising the group of vehicles in a desired formation and a higher level path generation method that provides safe paths for all the vehicles, based on graph theoretic considerations. The method can effciently handle a large group of any type of vehicles. In the article we focus on the control of quadrotor UAVs, thus the results are illustrated in 4D on a group of such vehicles.