Joint 3-D Localization-Enabled Detection-Aware Covert Transmission for HAP–ISAC Networks

Abstract

To address the dual threats of eavesdropping and active detection posed by an illegal autonomous aerial vehicle (AAV), this paper proposes a novel reconfigurable intelligent surface (RIS)-assisted integrated sensing and communication (ISAC) secure cooperation scheme. By leveraging the wide-area coverage of the high-altitude platform (HAP) and the channel reconstruction of RIS, the proposed scheme enables covert downlink command transmission from the HAP to the ground gateway while ensuring secure uplink data transmission for ground users. Specifically, the HAP, equipped with both communication and sensing capabilities, interacts with the ground gateway with the assistance of RIS, while continuously sensing and estimating the positions of an AAV using reflected echoes. For the mobile AAV, a factor graph optimization (FGO) method is proposed to achieve accurate AAV state estimation by exploiting the temporal correlation of continuous observation data. Based on the estimated AAV coordinates, we obtain the channel state and further analyze the detection performance of AAV in coherent and non-coherent detection scenarios, deriving closed-form solutions for the detection error probability (DEP) and the optimal detection threshold. Building on the above analysis, an optimization problem is formulated to maximize the effective covert rate (ECR), subject to constraints of sensing accuracy and covertness. For this high-dimensional and non-convex problem, a channel-aware graph neural network (CA-GNN) algorithm is proposed to jointly optimize sensing and communication beamforming as well as RIS phase shifts. Simulation results demonstrate the superiority of the proposed scheme in improving system security and covert performance. Compared to the genetic algorithm-based scheme and the deep neural network (DNN) scheme, the proposed scheme improves the ECR by 34.5% and 21.5%, respectively.