Deep Reinforcement Learning-Based Power Allocation for Rate-Splitting Multiple Access in 6G LEO Satellite Communication System

• DOI: 10.1109/lwc.2022.3196408
• Publication Date: 2022-08-04T19:20:55Z
• Authors: Jingfei Huang; Yang Yang; Longfei Yin; Dazhong He; Qing Yan
• Keywords: Transmitter power output; Performance metric; Communications satellite; Q-learning

Rate-splitting multiple access (RSMA) softly reconciles and decodes the extreme interference by non-orthogonal transmission, which can remarkably solve spectrum scarcity for future six-generation (6G) low earth orbits (LEO) satellite communication system. In this letter, we investigate power allocation problem in LEO networks with RSMA mechanism based on deep reinforcement learning (DRL) technique. Specifically, order to achieve better performance, base station (SBS) has effectively allocate transmit common private streams, is very challenging due uncertain limited information of channel distribution. To problem, a highly-effective proximal policy optimization (PPO) scheme further proposed, enables SBS learn an optimal strategy maximize sum rate system without knowing any prior information. Simulation results prove that proposed significantly outperforms other three baseline schemes terms metric computation complexity.