关键词: 蜂群无人机, 一致性, 渗流, MOLLY-REED准则, 鲁棒性, 灰预测

Abstract: Aiming at the characteristics that the short battle cycle and link quality’s quick change of the drone swarm system, a fast link evaluation algorithm under the condition of small samples is proposed by introducing the gray prediction method. On this basis, the quick estimation of nodes’ degree distribution is achieved, which provides data basis for network robustness control of the drone swarm system. Then the average consistency method is combined with the MOLLY-REED criterion in the percolation theory to overcome the disadvantage that the MOLLY-REED criterion needs the whole network’s flooding and is not suitable for the application of the drone swarm system. By configuring the degree distribution of nodes, a network topology with high robustness to node loss is established. The simulation results show that the proposed method can effectively improve the robustness of the drone swarm system’s communication network to node loss.

Key words: drone swarm, consensus, percolation, MOLLY-REED criterion, robustness, gray prediction