Joint Secure and Energy-Efficient Computation Offloading: Trajectory and IRS Reflective Design for UAV-MEC Networks

Abstract

Airborne edge computing built on unmanned aerial vehicles delivers viable solutions for heavy computational workloads in crowded metropolitan areas. Reconfigurable Intelligent Surfaces (RIS/IRS) eliminate radio signal obstructions from buildings and mitigate risks of unauthorized data eavesdropping. Two critical bottlenecks restrict low-power confidential workload uploading within RIS-assisted aerial edge systems: physical barriers and hidden wiretappers compromise transmission privacy, while UAV flight paths, RIS phase configurations, communication and computational resources exhibit tight coupling that raises joint optimization complexity. To address these two challenges, this work constructs an integrated co-design framework covering user workload assignment, UAV path planning, RIS phase tuning and multi-resource coordination. Numerical experiments validate that the proposed method sustains stable secure data delivery with suppressed power overhead, and outperforms all benchmark schemes by a considerable margin.