基于LSTM的LEO卫星链路自适应算法

低地球轨道(low earth orbit, LEO)卫星由于其传输损耗低、地面干扰小等优点成为空天地一体化网络的重要组成部分。由于星地传输链路的时延大,现有卫星通信过程无法实时地进行信息交互,导致系统无法适应信道的变化。针对这个问题,提出了基于长短期记忆(long short-term memory, LSTM)网络的信噪比(signal to noise ratio, SNR)预测方法,并利用预测的SNR调整系统的调制与编码方案(modulation and coding scheme, MCS),使其与快速变化的信道相匹配。仿真结果表明,提出的基于LSTM网络的SNR预测方法可以达到较高的准确度,并且根据预测的SNR实时调整MCS的方案大幅度地提高了系统的总吞吐量。

Link adaptation algorithm based on LSTM network for LEO satellite

Low earth orbit (LEO) satellite becomes an important part of the space-air-ground integrated network for its low transmission loss and little terrestrial interference. Due to the large delay of satellite-to-ground transmission link, the existing satellite communication process cannot achieve real-time channel state information interaction, which leads to the system's inability to adapt to the change of channel status. To solve this problem, an signal to noise ratio (SNR) prediction method based on long short-term memory (LSTM) network is proposed. Modulation and coding scheme (MCS) can be adjusted based on the predicted SNR to match the rapidly changing channel. Simulation results show that the proposed SNR prediction method based on LSTM network can achieve high accuracy, and the adjustment of MCS based on the predicted SNR in real-time can greatly improve the overall throughput of the system.