NOMA-Enabled Covert and Fair Data Collection for Multi-UAV Wireless Network Under Imperfect CSI

초록

This paper studies NOMA-enabled covert and fair short-packet data collection in multi-UAV networks under imperfect channel state information (CSI), assisted by a reconfigurable intelligent surface (RIS) to serve Internet of Things (IoT) nodes (INs). Motivated by sensitive IoT missions, we consider a scenario in which suppressing detectability can be more restrictive than maximizing throughput, and where finite-blocklength transmission and link heterogeneity necessitate a joint treatment of covertness, reliability, and fairness. Accordingly, this paper first characterizes the feasible boundary of covert communication under imperfect CSI and constructs an optimization-compatible covert constraint formulation. Then, a joint optimization framework is proposed to maximize the minimum amount of data collected from all INs, by jointly optimizing INs’ transmit power, RIS phase shifts, and the 3D trajectories of UAVs. To address the strong coupling and high non-convexity in the short-packet reliable rate expression with respect to the power variables under imperfect CSI, a two-layer nested alternating optimization algorithm for 3D covert and fair data collection (TLNA-3DCF) is developed. In the inner layer, robust NOMA power allocation is performed under imperfect CSI; in the outer layer, the RIS phase shifts and UAV 3D trajectories are alternately optimized based on the inner-layer results, with feasibility enforcement and convex approximations used to ensure convergence and manageable complexity. Simulation results demonstrate that the proposed scheme, while strictly satisfying covert communication constraints, significantly improves the service level of edge INs, thereby enhancing fairness in data collection. © 1972-2012 IEEE.

키워드