A Meta-Learning Framework for Resilient UAV Swarm Networks

Abstract

Maintaining the self-healing of communication connectivity (SCC) is a primary challenge for resilient Unmanned Sensor Networks (USNETs). This document proposes an intelligent framework utilizing Graph Convolutional Neural Networks (GCNs) and meta-learning to address unpredictable external destructions (UEDs). By integrating the Communication-Relaxed Meta Graph Convolution (CR-MGC) algorithm, the system enables autonomous topology reconstruction. Simulations demonstrate that this collaborative intelligence approach significantly reduces connectivity restoration time and computational complexity compared to traditional heuristic methods.