Performance Analysis and Optimization of AAV Jammer-Assisted Covert Communication Systems With Noncolluding Wardens

This work investigates covert detection and joint power-deployment design in an autonomous aerial vehicle ( AAV) jammer-assisted communication system with multiple noncolluding wardens, where a secure-zone abstraction is introduced to represent a minimum stand-off distance between the covert transmitter and potential wardens. By analyzing the optimal detection behavior under randomized-power jamming, the study demonstrates that the resulting interference uncertainty can effectively impair warden detection capability. On this basis, an analytical framework is developed to characterize covert performance under two information scenarios, depending on whether the warden locations and warden-side channel state information (CSI) are exactly known or only statistically known. Building on this framework, joint power allocation and jammer deployment problems are formulated to maximize the covert rate under covertness constraints. A sample average approximation (SAA)-based particle swarm optimization (PSO) algorithm is developed for the exact-information scenario, while an SAA-based coarse-to-fine joint optimization algorithm is proposed for the statistical-information scenario. Numerical results corroborate the analysis and show that the randomized AAV jamming strategy provides noticeable covertness gains over representative baselines. Moreover, the proposed optimization framework achieves the highest covert-rate performance in the exact-information scenario among the considered schemes, and attains essentially the same covert-rate performance as the benchmark methods with lower computational cost in the statistical-information scenario.