基于深度学习的多STBC盲识别算法

针对空时分组码(space-time block code, STBC)识别中多种编码类型难区分的问题, 提出了一种基于卷积神经网络的STBC盲识别算法。该算法首先将接收信号采用自相关函数的频域预处理, 输入到卷积神经网络中对信号特征进行提取, 全连接层对特征进行映射, 实现对6种STBC类型的识别。仿真实验结果表明, 在无信道和噪声等先验信息的条件下, 所提算法能够有效区分3种相似度高的STBC3码, 且将STBC可识别的编码类型由目前的4种扩充到6种, 识别准确率能达到96%。该方法的复杂度较低, 不需要利用大量样本数据, 实时性高, 具有较好的工程应用价值。

Blind recognition algorithm for multi-STBC based on deep learning

To solve the problem that different coding types are difficult to distinguish in space-time block code(STBC)recognition,a blind algorithm is proposed for STBC recognition based on convolutional neural network.In this algorithm,the received signal is preprocessed in frequency domain by autocorrelation function,input into the convolutional neural network to extract the signal features,and the features are mapped at the full connection layer to realize the recognition of six STBC types.Simulation experiment results show that the proposed algorithm can effectively distinguish three STBC3 codes of high similarity in the absence of channel and noise under the condition of a priori information,and the recognizable code type of STBC can be expanded from the current four to six,identification accuracy can reach 96%.The complexity of this method is low,and does not need to use a large number of sample data,Which has high real-time performance and good engineering application value.