Abstract:Low-Earth orbit (LEO) satellites are recognized as one of the most promising infrastructures for realizing global Internet of Things (IoT) services. With the explosive growth of user terminals (UTs) and data traffic, the integration of massive multiple-input multiple-output (mMIMO) techniques and LEO satellite communication systems has been regarded as a novel idea to enhance system capacity and realize global seamless high-speed interconnection. However, obtaining effective downlink channel state information (CSI) and establishing a simple and efficient hybrid beamforming mechanism are challenging tasks due to the limitations of objective factors, such as high dynamic, long delay, and low payload in LEO satellite scenarios. It is embodied in three aspects: 1) the untenable channel reciprocity in time division duplex (TDD) systems; 2) the training feedback costs and feedback delay in frequency division duplex (FDD) systems; and 3) the complex nonconvex optimization process faced by hybrid beamforming design. Driven by the performance advantages of the deep learning (DL) technology to deal with various problems in the field of physical layer communications, this article proposes to use a deep neural network (DNN) to solve the above challenges, and constructs SatCP and SatHB schemes for realizing downlink CSI acquirement and hybrid beamforming design, respectively. By deeply mining the potential correlation of the uplink-downlink channels between LEO satellites and UTs and exploring the mapping relationship between CSI and beamformers, the SatCP can assist LEO satellites to directly predict the future downlink CSI based on the observed uplink CSI with no need for downlink channel estimation, while the SatHB can easily generate the corresponding beamformers based on the downlink CSI predicted by the SatCP without requiring complex optimization. Numerical results demonstrate that the proposed SatCP and SatHB can play an effective auxiliary role - n LEO satellite mMIMO communication systems.

Reinforcement Learning-Based Downlink Transmit Precoding for Mitigating the Impact of Delayed CSI in Satellite Systems

Wireless Channel Prediction for Multi-user Physical Layer with Deep Reinforcement Learning

Hybrid IRS-Assisted Secure Satellite Downlink Communications: A Fast Deep Reinforcement Learning Approach

Collaborative Computing in Non-Terrestrial Networks: A Multi-Time-Scale Deep Reinforcement Learning Approach

Flexible Robust Beamforming for Multibeam Satellite Downlink using Reinforcement Learning

Integrating LEO Satellite and UAV Relaying via Reinforcement Learning for Non-Terrestrial Networks

Toward Intelligent Non-Terrestrial Networks Through Symbiotic Radio: A Collaborative Deep Reinforcement Learning Scheme

Collaborative Deep Reinforcement Learning for Resource Optimization in Non-Terrestrial Networks

Tackling Delayed CSI in a Distributed Multi-Satellite MIMO Communication System

Deep Learning-based Joint Channel Prediction and Multibeam Precoding for LEO Satellite Internet of Things

Deep Learning (DL)-Based Channel Prediction and Hybrid Beamforming for LEO Satellite Massive MIMO System

Dynamic Resource Allocation With Deep Reinforcement Learning in Multibeam Satellite Communication

Reinforcement Learning for Cognitive Delay/Disruption Tolerant Network Node Management in an LEO-based Satellite Constellation

Adaptive anti-jamming beamforming based on the preprocessing deep reinforcement learning for downlink navigation communication

A Satellite Adaptive Modulation Coding Method Based on Deep Reinforcement Learning

Dual Network Computation Offloading Based on DRL for Satellite-Terrestrial Integrated Networks

Deep Reinforcement Learning based Downlink Beamforming and Phase Optimization for RIS-Aided Communication System

Intelligent Channel Prediction and Power Adaptation in LEO Constellation for 6G

Deep Reinforcement Learning Based Dynamic Power and Beamforming Design for Time-Varying Wireless Downlink Interference Channel