Software-defined resilient networking architecture for UAV swarms

To address the dynamic survivability and mission sustainability challenges of UAV (unmanned aerial vehicle) swarms in composite soft-kill and hard-kill threat environments, focusing on reducing topology reconfiguration latency caused by terminal node failures and large-scale attrition.

A software-defined resilient networking architecture was designed, incorporating a dynamic control authority migration mechanism within a centralized-distributed hybrid control framework. Intent-driven paradigms were integrated with SDN technology for dynamic intelligent resource configuration. Efficacy was validated through a dual-mode reconfiguration simulation model: a distributed real-time perception-based autonomous substitution mechanism for local node failures, and a dynamic resource pool scheduling mechanism for large-scale node attrition. Topology reconfiguration efficiency was quantitatively evaluated across scenarios via comparative analysis.

Under identical node loss conditions, the distributed real-time perception-based neighbor substitution mechanism achieves a 38.5% average reconfiguration latency reduction compared to the resource pool scheduling mode. In sustained electromagnetic jamming scenarios, the hybrid strategy further compresses latency by 21.7% relative to pure distributed approaches. As failed nodes increase, the centralized-distributed controller combination exhibits the shortest reconstruction time in ideal environments, demonstrating significant advantages in network robustness and mission continuity.

The proposed architecture enables efficient topology reconfiguration under both local node failures and large-scale attrition scenarios, particularly in complex electromagnetic environments. It delivers critical theoretical foundations and technical underpinnings for enhancing UAV swarm survivability and resilience.