基于SE-CsiNet的大规模MIMO信道状态信息反馈与重建算法

Channel state information feedback and reconstruction algorithmfor massive MIMO based on SE-CsiNet

本文研究了大规模多输入多输出(MIMO)系统FDD模式下行链路信道状态信息(CSI)反馈与重建问题,针对传统基于压缩感知的CSI反馈与重建算法存在的计算复杂度高、对信道稀疏性要求严格等问题,以及现有的基于深度学习的CSI反馈与重建算法存在的性能不足问题,提出了一种基于联合注意力机制神经网络的大规模MIMO信道状态信息反馈与重建算法,基于自编码器神经网络架构提出了一种全新的SE-CsiNet神经网络模型,在译码器网络当中引入联合压缩-激活(SE)注意力机制算法,有效提高神经网络的特征提取能力。结果表明,所提出的算法相较于传统CSI反馈与重建算法及现有的深度学习算法性能更为优异。

The channel state information(CSI)of FDD downlink in massive multiple input multiple output(MIMO)systems is studied for feedback and reconstruction.The traditional CSI feedback and reconstruction algorithm based on compressed sensing has the problems of high computational complexity and strict requirement for channel sparsity,and the existing CSI feedback and reconstruction algorithms based on deep learning are insufficient in performance.For these reasons,a massive MIMO channel state information feedback and reconstruction algorithm based on joint attention mechanism neural network is proposed,and a new SE-CsiNet neural network model based on auto-encoder neural network architecture is proposed.The squeeze-and-excitation attention mechanism algorithm is introduced into the decoder network to improve the feature extraction ability of the neural network effectively.Experimental results show that the proposed algorithm has better performance than traditional CSI feedback and reconstruction algorithms and existing deep learning algorithms.