Joint Beamforming and Dynamic Beam Hopping Based on MAPPO for LEO Satellite Communication System
Meng Meng, Bo Hu, Shanzhi Chen, Shaoli Kang
Abstract
To meet time-varying and nonuniform service requirements, how to flexibly allocate resource-constrained to meet the traffic demands is an important topic in LEO satellite communication system. We introduce joint beamforming and dynamic beam hopping (BH) algorithm to support hybrid wide-spot beam coverage in LEO satellite communication system, which can transmit control signalling and user data simultaneous. Considering that the joint decision of beamforming and BH will lead to explosive growth of the action space dimension, a cooperative Multi-Agent Proximal Policy Optimization (MAPPO) algorithm is presented to solve this problem in hybrid wide-spot beam coverage scenario. The beamforming problem is decomposed into two subproblems, power allocation and analog beamforming problems. The analog beamforming problem is solved by ZF beamforming algorithm, and the power allocation and BH problems are solved by MAPPO algorithm. In MAPPO algorithm, each agent only undertakes the BH or power allocation decision of one beam. The agents learn to cooperate with others via shared rewards to achieve a common goal of maximize the system throughput and minimize the delay fairness (DF) while ensuring the minimum rate requirement of the control beam. Simulation results show that the MAPPO algorithm can achieve real-time BH and power allocation to match time-varying traffic requests. In addition, the proposed algorithm can achieve about 8 Mbps throughput gain compared with GABH.