Optimizing UAV Placement With Spatial Threat Exclusion for Secure and Energy-Efficient Wireless Communications

Wireless communications in military and civilian applications demand robust security and energy efficiency against evolving threats. This paper presents a novel strategy for optimizing UAV placement via spatial threat exclusion to achieve secure and energy-efficient links without relying on threat channel state information or complex signal processing techniques like beamforming or artificial noise. Our approach is motivated by the inherent alignment between covertness and interference mitigation; it is therefore equally applicable to interference scenarios, enhancing security by lowering detection risk and improving energy efficiency by reducing the communication distance between the UAV and the base station (BS). To ensure robustness, the model is developed for a conservative coverage-edge scenario, which represents the worst-case condition for both signal quality and secrecy, thereby serving as a lower-bound performance benchmark. The proposed method combines closed-form solutions for initial placement with iterative refinements, enabling real-time adaptation to dynamic threat mobility without extra hardware or computational burden. Two algorithms are developed: first for a single-threat, single-user scenario and then extended to multi-threat, multi-user settings. Algorithm 1 is a computationally efficient, quadratic difference-based method for rapid deployment, while Algorithm 2 is an inverse quadratic penalty-based technique for threat avoidance in dynamic environments. Simulation results for both scenarios confirm that the proposed algorithms consistently achieve positive secrecy rates across diverse propagation environments while yielding significant energy and time savings compared to an optimal-placement baseline. A comparative evaluation with a particle swarm optimization (PSO) baseline further demonstrates that our algorithms match optimal performance with substantially lower computational overhead.