基于深度学习的MIMO系统信道估计算法

精确的信道估计对于保证无线通信系统性能至关重要。针对多输入多输出(multiple input multiple output, MIMO)系统传统信道估计算法需已知信道统计信息以及性能与复杂度折中等问题，提出一种基于深度学习的多网络级联MIMO系统信道估计方案。基于卷积神经网络构建信道信息重建网络，初步重构出信道信息，进而基于深度残差网络构建信道估计网络进行级联得出估计结果，并利用多个损失函数对网络进行优化。仿真结果表明，在牺牲一定时间复杂度的情况下，所提方案的均方误差随信噪比增加逐渐优于线性最小均方误差(linear minimum mean squared error, LMMSE)估计算法，且不受信道统计信息的约束。

Channel estimation algorithm for MIMO systems based on deep learning

Accurate channel estimation is very important to guarantee the performance of wireless communication system. To solve the problem that traditional channel estimation algorithms for m ultiple input multiple output (MIMO) systems need to know channel statistics and compromise performance complexity, we propose a channel estimation scheme for multi-network cascaded MIMO systems based on deep learning. Firstly, the channel information is constructed based on the convolutional neural network to reconstruct the channel information, and then the channel estimation network is constructed based on the deep residual network to cascade and obtain the estimation results, and multiple loss functions are used to optimize the network. Simulation results show that the mean square error (MSE) of the proposed scheme is better than that of LMMSE with the increase of SNR at the expense of certain time complexity, and is not constrained by channel statistics.