| 摘要: |
| 非合作条件下直扩信号感知问题一直是通信对抗领域研究的热点，传统的感知方法在低信噪比条件下性能下降严重。为有效提升直扩信号感知的性能，提出了一种基于改进残差神经网络的直扩信号感知方法。首先，通过广义互相关算法，快速提取直扩信号的相关峰特性；然后，以残差神经网络为基础，融合注意力机制，建立网络模型，分析识别抽象特征；最后，在仿真的数据集中进行实验验证。结果表明，相较于时域自相关法、卷积神经网络法等，所提方法具备更好的感知效果，能够在信噪比为-16 dB的情形下有效地感知直扩信号。 |
| 关键词: 直扩信号感知 广义互相关 残差神经网络 注意力机制 |
| DOI：10.20079/j.issn.1001-893x.230814003 |
|
| 基金项目: |
|
| A DSSS Signal Sensing Method Based on Improved Residual Neural Networks |
| REN Jiangfei,XU Kui,LIU Yang,LU Rui,ZHANG Mi,YE Zilu |
| (（College of Communications Engineering,Army Engineering University of PLA,Nanjing 210007,China）) |
| Abstract: |
| The problem of direct sequence spreading spectrum(DSSS) signal sensing under non-cooperative conditions has always been a hot research topic in the field of communication countermeasures,and the performance of traditional sensing methods is seriously degraded under low signal to noise ratio(SNR) conditions.In order to effectively improve the performance of DSSS signal sensing,a DSSS signal sensing method based on an improved residual neural network(ResNet) is proposed.Firstly,the correlation peak characteristics of the DSSS signal are quickly extracted by the Generalized Cross Correlation(GCC) algorithm.Then,based on the ResNet,the attention mechanism is introduced to build a network model to analyze and identify the abstract features.Finally,experimental validation is carried out in the simulated dataset.The results show that,compared with the time-domain autocorrelation method and the convolutional neural network method,the proposed method has a better sensing effect and can effectively sense the DSSS signals with -16 dB SNR. |
| Key words: DSSS signal sensing generalized cross correlation residual neural network attention mechanism |
