Abstract
This paper addresses the critical security challenges in Unmanned Aerial Vehicle (UAV)-enabled Cognitive Radio Networks (CRNs) facing malicious, full-duplex, and colluding eavesdroppers. We propose a joint trajectory and power optimization framework to maximize the worst-case average secrecy rate (WASR). Utilizing Block Coordinate Descent (BCD) and Successive Convex Approximation (SCA), the algorithm ensures efficient spectrum sharing with primary users while mitigating sophisticated eavesdropping threats. Simulation results demonstrate that the proposed joint optimization significantly outperforms benchmark schemes, nearly doubling the secrecy rate in high-threat scenarios.
Introduction
UAVs are increasingly deployed as "flying base stations" to provide flexible wireless coverage. However, their high-altitude, Line-of-Sight (LoS) links make them vulnerable to eavesdropping. This research tackles a complex security scenario involving Cognitive Radio constraints, where a UAV (secondary user) must share spectrum with primary users without causing excessive interference. The threat model includes full-duplex eavesdroppers that can simultaneously listen and transmit jamming signals, and collusion, where multiple attackers share intercepted data to jointly decode messages.
System Model and Security Challenges
The system consists of a UAV source ($S$), multiple primary users ($D_k$), legitimate secondary users ($U_n$), and several colluding eavesdroppers ($E_m$). The challenge lies in designing a flight path and power control strategy that balances conflicting requirements: maintaining proximity to legitimate users, avoiding eavesdroppers, and protecting primary users from interference.
Key challenges addressed include:
- Location Uncertainty: Accounting for bounded errors in the known positions of attackers.
- Interference Temperature (IT): Adhering to strict limits on the interference caused to primary spectrum users.
- Collusion Dynamics: Treating eavesdroppers as a single entity to ensure a conservative and robust security posture.
Optimization Methodology
The problem is formulated as a maximization of the Worst-case Average Secrecy Rate (WASR). Due to the non-convex nature of the coupled variables, a two-step iterative approach is employed:
- Power Optimization: For a fixed trajectory, the UAV's transmit power is optimized using first-order Taylor expansion to approximate the non-convex secrecy rate into a convex form.
- Trajectory Design: For a fixed power profile, the UAV's 2D flight path is optimized. Slack variables and Successive Convex Approximation (SCA) are used to handle complex distance-based constraints.
- Block Coordinate Descent (BCD): The algorithm toggles between these subproblems until the secrecy rate converges, typically within 8 iterations.
Performance Evaluation and Key Insights
Numerical simulations were conducted across various eavesdropper distributions (Uniform, Clustered, and Interweaved). The results highlight several critical insights:
- The "Avoidance" Maneuver: The optimized trajectory is not a straight line; the UAV actively "dodges" high-density eavesdropper zones. When proximity to an eavesdropper is unavoidable, the UAV dynamically throttles its transmit power to minimize information leakage.
- Superiority over Benchmarks: The joint optimization approach significantly outperforms fixed-power (FPOT) or fixed-trajectory (OPFT) baselines. In uniform distribution scenarios, the proposed scheme achieved secrecy rates nearly double those of the OPFT benchmark.
- IT Threshold Trade-off: The Interference Temperature threshold acts as a performance ceiling. Increasing this threshold allows for higher transmit power, boosting the secrecy rate until hardware limits are reached.
Conclusion
This study demonstrates that joint trajectory and power optimization is essential for securing aerial CRNs against colluding threats. By intelligently navigating the spatial environment and modulating power, UAVs can maintain high secrecy rates even in the presence of sophisticated attackers. Future work will explore 3D maneuvering and probabilistic LoS/Non-LoS channel models to better reflect complex urban environments.
Summary updated