Robust Power Minimization for RSMA-Aided UAV Downlink in Maritime IoT Networks

Abstract

Maritime Internet-of-Things (IoT) networks require reliable and energy-efficient wireless connectivity in open-sea environments where terrestrial infrastructure is sparse. This paper investigates robust transmit power minimization for a rate-splitting multiple access (RSMA)-aided unmanned aerial vehicle (UAV) downlink. In the considered system, a UAV equipped with a uniform planar array serves spatially distributed single-antenna maritime terminals by transmitting one common stream and multiple private streams. Under Rician fading channels with Gaussian channel state information (CSI) errors, we formulate a joint optimization problem for beamforming, common-rate allocation, and three-dimensional UAV positioning, subject to per-terminal probabilistic quality-of-service (QoS) constraints. To handle the intractable chance constraints, a Bernstein-type inequality is used to obtain deterministic conservative constraints. An alternating optimization algorithm is then developed. For a fixed UAV position, the robust beamforming and common-rate allocation are updated via semidefinite relaxation (SDR) and linear programming, respectively. For fixed beamforming directions, the UAV position is optimized through a power-scaling reformulation and successive convex approximation (SCA). Simulation results show that the proposed robust RSMA design requires lower transmit power than the considered non-robust RSMA, SDMA, and NOMA benchmarks under CSI uncertainty and stringent QoS requirements.