A distributed minimum restrictive connectivity maintenance algorithm

The dynamic character of autonomous mobile multi-agent networks makes the control of the individual agent's movement a challenging task when certain properties or conditions concerning the global network state must be fulfilled. Many distributed algorithms require that the communication topology of the regarded multi-agent network is connected at all times to execute properly. Since particularly motion control algorithms directly influence the connectivity itself, the fundamental question how connectivity of the network can be maintained is raised. As an additional difficulty, individual agents often have only access to local information, whereas connectivity is a property of the global network. In this work, a distributed hybrid connectivity maintenance algorithm is proposed, which guarantees global connectivity of the network based on local information. The algorithm is divided in two components: First, a discrete topology control algorithm determines periodically the structure of a minimum cost spanning tree on the network communication graph. Second, a continuous connectivity motion control algorithm constrains the location of each individual agent such that the determined spanning tree is preserved. This combined algorithm effectively maintains connectivity under the presence of any finite-valued primary motion control law.