| 利用卷积-循环神经网络的串行序列空时分组码识别方法 |
| |
| 引用本文： | 张聿远,闫文君,林冲,姚成柱.利用卷积-循环神经网络的串行序列空时分组码识别方法[J].信号处理,2021,37(1):19-27. |
| |
| 作者姓名： | 张聿远 闫文君 林冲 姚成柱 |
| |
| 作者单位： | 海军航空大学信息融合研究所 |
| |
| 基金项目： | 泰山学者工程专项基金（ts201511020） |
| |
| 摘 要： | 针对多输入多输出(Multiple Input Multiple Output, MIMO)系统中的空时分组码识别(Space-Time Block Code, STBC)问题,本文提出了一种利用卷积-循环神经网络的串行序列空时分组码识别方法。将一维接收信号的实部和虚部分离后输入网络,利用卷积神经网络(CNN)提取其空间特征,结合循环神经网络(RNN)提取其深层时序特征,提高网络的特征表达能力;网络训练过程采用反向传播方法,通过计算输出与目标值的误差,将误差反向传回网络中并更新权值,完成网络的训练过程;将测试集数据输入训练好的网络中,实现对空时分组码的识别和区分。该方法将深度学习算法运用到串行序列空时分组码识别当中,训练完的网络可直接对单接收天线下的空时分组码进行识别,不需要重复计算信号的统计特征,避免了人为设计特征参数和检测阈值。该方法不需要知道信道和噪声的先验信息,适用于电子侦查等非协作通信情况。仿真实验表明,该算法能够有效地对串行序列空时分组码进行识别,并且在低信噪比下有较好的识别性能。
|
| 关 键 词： | 空时分组码 信号处理 卷积神经网络 循环神经网络 深度学习 |
| 收稿时间： | 2020-06-19 |
Serial Sequence Space-Time Block Code Recognition Method By Using Convolutional-Recurrent Neural Networks#br# |
| |
| Affiliation: | Institute of Information Fusion, Naval Aeroautical University |
| |
| Abstract: | Aiming at the space-time block code recognition problem of multiple input multiple output system, a method of space-time block code recognition for serial sequences by using convolutional cyclic neural network is proposed. The real and imaginary parts of the one-dimensional received signal are separated into the network, and the spatial characteristics are extracted by the convolutional neural network, and the deep-seated temporal characteristics are extracted by the recurrent neural network, so as to improve the characteristic expression ability of the network. The network training process adopts the back propagation method to calculate the error between the output and the target value, send the error back to the network and update the weight to complete the network training. The test data is input into the trained network to realize the recognition of STBC code. This method applies the deep learning algorithm to the recognition of space-time block codes of serial sequences for the first time. The trained network can directly recognize space-time block codes under a single receiving antenna without the need to double calculate the statistical characteristics of the signal and avoid the artificial design of feature parameters and detection threshold. Simulation results show that the algorithm can recognize the space-time block codes of serial sequences and has good recognition performance under low SNR. |
| |
| Keywords: | |
|
| 点击此处可从《信号处理》浏览原始摘要信息 |
|
点击此处可从《信号处理》下载全文 |
|
