Quantitative Analysis of UAV Intrusion Mitigation for Border Security in 5G with LEO Backhaul Impairments

TL;DR

The paper addresses the problem of detecting and mitigating uncooperative UAV intrusions in border corridors over a 5G network with LEO satellite backhaul, where outages can threaten timely lockdowns. It introduces a minimal end-to-end MATLAB simulator that couples terrestrial gNBs with LEO backhaul, a detection logic based on UAV mobility/RSS patterns, and a lockdown mitigation that can rely on local fallback to bound delays. Through Monte Carlo sweeps across UAV speed/altitude and backhaul outage scenarios, the study shows that satellite outages can cause unbounded delays unless mitigated by a 2 s local fallback deadline, while collateral impact on patrol UEs remains small. The findings support hybrid terrestrial-NTN architectures with edge-based fallback as a robust design principle for timely and safe UAV interdiction in border security applications.

Abstract

Uncooperative unmanned aerial vehicles (UAVs) pose emerging threats to critical infrastructure and border protection by operating as rogue user equipment (UE) within cellular networks, consuming resources, creating interference, and potentially violating restricted airspaces. This paper presents minimal features of the operating space, yet an end-to-end simulation framework to analyze detect-to-mitigate latency of such intrusions in a hybrid terrestrial-non-terrestrial (LEO satellite) 5G system. The system model includes terrestrial gNBs, satellite backhaul (with stochastic outages), and a detection logic (triggered by handover instability and signal quality variance). A lockdown mechanism is invoked upon detection, with optional local fallback to cap mitigation delays. Monte Carlo sweeps across UAV altitudes, speeds, and satellite outage rates yield several insights. First, satellite backhaul outages can cause arbitrarily long mitigation delays, yet, to meet fallback deadlines, they need to be effectively bounded. Second, while handover instability was hypothesized, our results show that extra handovers have a negligible effect within the range of parameters we considered. The main benefit of resilience from fallback comes from the delay in limiting mitigation. Third, patrol UEs experience negligible collateral impact, with handover rates close to terrestrial baselines. Stress scenarios further highlight that fallback is indispensable in preventing extreme control-plane and physical security vulnerabilities: Without fallback, prolonged outages in the satellite backhaul delay lockdown commands, allowing rogue UAVs to linger inside restricted corridors for several seconds longer. These results underscore the importance of complementing non-terrestrial links with local control to ensure robust and timely response against uncooperative UAV intrusions.

Figures (5)

• Figure 1: System model for UAV intrusion detection and mitigation in a border-security scenario
• Figure 2: Detect $\rightarrow$ Mitigate delay. Top: median heatmaps across the $(speed, altitude)$ grid. Bottom: pooled CDFs with bootstrap confidence bands.
• Figure 3: Extra handovers between detection and mitigation. Top: median heatmaps. Bottom: pooled CDFs with confidence bands.
• Figure 4: patrol UE handover rate (per minute). Left: median heatmaps. Right: pooled CDFs.
• Figure 5: Stress case: long outage (10 s) versus fallback with 2 s deadline. Fallback enforces a hard cap on mitigation delay.