GNSS-spoofing-resilient drone-assisted 5G backhaul via cooperative detection and adaptive link reconfiguration

Drone-assisted backhaul can extend 5G and beyond-5G connectivity in dense, temporary, and infrastructure-constrained environments, but its service continuity depends on reliable aerial positioning. Gradual Global Navigation Satellite System (GNSS) carry-off spoofing can bias the reported position of an unmanned aerial vehicle (UAV) relay, degrading access association, backhaul selection, latency, and outage behaviour without immediate network failure. Existing UAV communication studies largely assume benign localisation, while GNSS anti-spoofing research usually prioritises receiver integrity rather than backhaul service continuity. To address this gap, this paper introduces Cooperative Spoofing Detection and Adaptive Link Reconfiguration (CSD-ALR), a spoofing-aware scheme for drone-assisted 5G backhaul. Its novelty is the coupling of cooperative spoofing-risk evidence with adaptive access and backhaul selection. CSD-ALR fuses local kinematic plausibility, peer-consistency evidence, and temporal drift accumulation into a bounded relay-risk score used for service-aware reconfiguration. A MATLAB system-level evaluation with 30 paired Monte Carlo trials assesses four deployment regimes across 8, 15, 120, and 200 UAV relays. In the 200-UAV constrained-backhaul case with the largest observed S4–S1 gain, CSD-ALR improves throughput by 459.07 Mbps, reduces latency by 58.53 ms, improves demand satisfaction ratio (DSR) by 0.2617, and reduces outage by 0.4420 relative to spoofing without defence. The results indicate that cooperative risk evidence can improve aerial backhaul resilience under GNSS carry-off spoofing, while exposing fairness, tail-latency, and coordination-overhead limits at large swarm scale.