Abstract
The rapid expansion of the Internet of Things (IoT) is hindered by energy constraints and security vulnerabilities. This paper presents a comprehensive secrecy analysis for IoT networks where an Unmanned Aerial Vehicle (UAV) serves as both a wireless power source and a mobile edge computing (MEC) hub. By deriving closed-form expressions for the Secrecy Successful Computation Probability (SSCP) under Nakagami-$m$ fading, we demonstrate how to optimize energy harvesting and UAV positioning to protect data offloading from passive eavesdroppers. The analysis accounts for real-world imperfections, including imperfect channel state information (iCSI) and imperfect successive interference cancellation (iSIC).
Introduction
IoT devices often lack the battery life and hardware capabilities for heavy encryption or complex local computations. While Mobile Edge Computing (MEC) allows for task offloading and Non-Orthogonal Multiple Access (NOMA) enhances spectral efficiency, wireless transmissions remain vulnerable to interception. Terrestrial networks also face "dead zones" in dense urban environments. This study proposes using a UAV as a flexible, high-altitude relay providing Wireless Power Transfer (WPT) and secure computation resources to ground-based IoT clusters.
System Model and Protocol
The proposed system consists of a UAV, two clusters of IoT edge devices (EDs)—Near and Far—and a passive eavesdropper (E). The operation follows a quad-phased protocol:
- Energy Harvesting (EH): The UAV broadcasts RF signals, and EDs harvest this energy.
- Secrecy Offloading: EDs use harvested energy to offload tasks to the UAV via NOMA.
- Computation: The UAV processes the offloaded tasks.
- Downloading: The UAV transmits results back to the EDs.
Secrecy Performance Analysis
The core metric, Secrecy Successful Computation Probability (SSCP), measures the likelihood that a task is computed successfully and remains secure from the eavesdropper within a specific time limit. The analysis utilizes Nakagami-$m$ fading channels to model a wide range of propagation environments. Crucially, the model incorporates:
- iCSI: Errors in channel state estimation.
- iSIC: Residual interference from imperfect signal separation in NOMA.
Key Insights and Results
The EH Ratio Trade-off
The parameter $\eta$ represents the fraction of time allocated to energy harvesting. A critical trade-off exists: insufficient EH time leaves devices with low transmission power, while excessive EH time reduces the window for data offloading and processing.
UAV Positioning and Device Density
There is an optimal UAV altitude; too low results in obstacle-blocked non-line-of-sight (NLoS) paths, while too high increases path loss. Interestingly, increasing the number of EDs improves the SSCP, as the UAV can select "best-case" candidates with higher-quality, secure links.
Impact of Imperfections
The performance gap between perfect (pCSI-pSIC) and imperfect (iCSI-iSIC) scenarios is significant, highlighting that estimation errors and hardware limitations are major bottlenecks in practical IoT deployments.
Conclusion
This work provides a mathematical framework for optimizing UAV-assisted secure MEC. Future research will explore active eavesdropping (jamming), dynamic UAV trajectory optimization, and Multi-Input Multi-Output (MIMO) configurations to further enhance network resilience and capacity.
Summary updated