一种适用多种空时编码方案的MIMO自编码器通信系统

A MIMO Autoencoder Communication System Suitable for Multiple Space-Time Coding Schemes

为提升现有端到端通信系统的适应性与信道估计的准确性,提出了一种适用于多种空时编码方案的多输入多输出(Multiple-Input Multiple-Output,MIMO)自编码器通信系统。该系统将基于卷积神经网络的自编码器引入到MIMO系统中,并结合信道估计网络实现信道均衡,通过端到端的学习方式实现各种空时编码方案下信号映射传输,以获得最佳接收性能。仿真实验表明,在不同信道环境、调制方式以及空时编码方案下,所提系统表现出了良好的适应性和泛化能力;并且在垂直空时分层码方案下,所提系统可在低信噪比情况匹配最大似然检测性能,在高信噪比情况获得优于传统非线性方法的信号检测性能。与传统方法相比,所提的信道估计方法在长期演进的扩展典型城市信道下可获得5.5 dB左右的性能提升。

In order to improve the adaptive capacity of existing end-to-end communication systems and the accuracy of channel estimation methods,a multiple-input multiple-output(MIMO)autoencoder(AE)communication system suitable for multiple space-time coding schemes is proposed.This AE-MIMO system introduces the autoencoder based on the convolutional neural network into the MIMO system,and combines the channel estimation network to achieve channel equalization.This system implements signal mapping and transmission under various space-time coding schemes through end-to-end learning to obtain the best receiving performance.Simulation experiments show that the proposed system has good adaptability and generalization ability under different channel environments,modulation methods and space-time coding schemes.Moreover,under the vertical space-time layered coding scheme,the proposed system can match the maximum likehood detection performance in the case of low signal-to-noise ratio(SNR),and obtain better signal detection performance than traditional nonlinear methods in the case of high SNR.Compared with the traditional method,the proposed channel estimation method obtains a performance improvement of about 5.5 dB under long term evolution extended typical urban model.