Adaptive D-Hop Connected Dominating Set in Highly Dynamic Flying Ad-Hoc Networks

By exploring the intelligent cooperation of unmanned aerial vehicle (UAV) swarms, the formed flying ad-hoc networks (FANETs) can support a variety of collaborative operations with real-time communications in emergency scenarios. To reduce the prohibitively high routing overhead with the connectivity guaranteed of multi-hop links, UAV swarms can construct a virtual backbone network (VBN) based on the graph-theoretical d-hop connected dominating set (d-CDS), where each UAV outside VBN can send collected data to VBN within d-hop distance. However, due to the high dynamics of FANETs in emergency scenarios, the optimal solution may not match the current status, which results in frequently intermittent connectivity. Besides, recomputing the solution from scratch will lead to significant maintance costs. Therefore, it is crucial to adapt the minimal d-CDS to topology changes. To this end, we propose an \mathcal {O}(d\log (N))-approximation algorithm (i.e., N denotes the maximal number of nodes) with expected \mathcal {\widetilde{O}}(d\Delta ^2) (i.e., \Delta denotes the maximal degree of a vertex over the sequence of updates) time per update. The simulation results demonstrate that our adaptive solution can strike a better trade-off among the routing overhead, response time, and maintance costs per topology update compared with state-of-the-art schemes in emergency scenarios.