Energy- and Latency-Efficient Offloading for Secure Cooperative UAV-NOMA-Enabled Mobile Edge Computing Networks

Abstract This work investigates energy- and latency-efficient computation offloading for secure cooperative UAV-NOMA-enabled MEC networks in the presence of an eavesdropper. A subset of UAVs serves as jammers to mitigate eavesdropping, while the rest perform partial offloading. Given the passive eavesdropper assumption, secrecy outage probability is adopted as the secrecy metric, and its closed-form expression is derived by analyzing SINR statistics at the ground access point and the eavesdropper. Three non-convex optimization problems are formulated: minimizing individual UAV energy consumption, minimizing total energy of offloading and jamming UAVs, and minimizing task latency under an energy budget. Suboptimal solutions are obtained via constraint simplification, multi-to-single-variable transformation, and numerical methods. Results demonstrate that partial offloading balances energy and latency, with optimal offloading bits and jamming power; larger latency deadlines, secrecy outage probability thresholds, or energy budgets enhance performance.