Unmanned Aerial Vehicles (UAV) have several uses in civilians and military applications, such as search and rescue missions, cartography and terrain exploration, industrial plant control, surveillance, public security, firefight, and others. Swarms of UAVs may further increase the effectiveness of these tasks, since they enable larger coverage, more accurate or redundant sensed data, fault tolerance, etc. Swarms of aerial robots require real-time coordination, which is just a specific case of M2Mcollaboration. But one of the biggest challenges of UAV swarming is that this real-time coordination has to happen in a wide-area setting where it is expensive, or even impossible, to set up a dedicated wireless infrastructure for this purpose.
Instead, one has to resort to conventional 3G/4G wireless networks, where communication latencies are in the range of 50-150 ms. In this paper we tackle the problem of UAV swarm formation and maintenance in areas covered by such mobile network, and propose a bandwidth-efficient multi-robot coordination algorithm for these settings. The coordination algorithm was implemented on the top of our mobile middle ware SDDL, uses its group-cast communication capability, and was tested with simulated UAVs.