Distributed Recursive Grouping Based Fault Self-Healing of UAV Swarm With Individuals Failure

This paper addresses the issue of random individuals failure within unmanned aerial vehicle (UAV) swarms configuration and consistency, this study focuses on exploring fault self-healing strategies for ensuring the reliability of UAV swarms. Due to the cascade effect, failed individuals will cause local or global instability of the entire system, such as, the partial loss of configuration, swarm dispersed, destruction of consistency, and even the failure of the entire system tasks. A distributed recursive grouping structure is proposed for the UAV swarm self-organised configuration and consistency control, and a hierarchical hypergraph is modeled for the scale-free network structure. Afterwards, the influence of system configuration and consistency with two different cases of nodes failure is analyzed, and a fault recoverability condition is established to analysis the network recoverability after the nodes failure. Then, with the configuration phase transition adjustment and dynamic cluster maintenance mechanisms, the self-healing control approaches are presented to restore the original configuration and consistency of the swarm at the network level and the physical level respectively. Finally, a sequence of simulations is conducted to showcase the efficacy of the proposed methods.