On sum-rate maximization in downlink UAV-aided RSMA systems

Abstract

The synergy of uncrewed aerial vehicles (UAVs) and the rate-splitting multiple access (RSMA) technique has thrived as a crucial enabler for multi-user broadband sixth-generation networks. This paper investigates a UAV-aided RSMA downlink communication system considering user mobility and the uniform rectangular array antenna design. In particular, the sum-rate maximization problem is formulated where the UAV beamforming matrix, common rate allocation, and UAV trajectory design are jointly optimized. The deep deterministic policy gradient (DDPG) approach has been proposed to address the nonconcave objective function. In addition, the safe action shaping technique is incorporated into the algorithm to satisfy the variable constraints of the problem statement. The numerical simulation results demonstrated that the proposed approach outperformed other benchmark schemes in various settings of transmit powers and fading environment levels.

Keywords

• Rate-splitting multiple access (RSMA)
• Uncrewed aerial vehicle (UAV) trajectory
• Deep reinforcement learning
• User mobility