摘要
针对通信信号调制方式识别的特征选择和提取困难以及人工提取的特征表征能力有限等瓶颈问题,研究了一种基于短时傅里叶变换(STFT)和双向长短期记忆网络(BiLSTM)的通信信号调制方式的识别方法.该方法首先通过短时傅里叶变换获取信号时频域二维联合分布信息;然后采用双三次插值法抽样实现特征图的降维;最后利用双向长短期记忆网络学习时频联合分布谱图的深度特征,进行调制方式识别的分类.仿真结果表明,该方法的准确率优于传统方法.
In order to solve the bottleneck problems such as difficulty in feature selection and extraction of the characteristics of modulation recognition of communication signals,limited feature representation ability of manual extraction,etc.,this paper studies a modulation recognition method for communication signals based on short time Fourier transform(STFT)and bidirectional long short-term memory networks(BiLSTM).This method firstly obtains the 2D joint distribution information of time-frequency domain by STFT,and then adopts bi-cubic interpolation algorithm to realize the dimensionality reduction of feature map,and finally uses the Bi-LSTM to learn the depth features of time-frequency joint distribution spectrum,and classify the modulation pattern recognition.Simulation results show that the accuracy of the proposed method is better than that of the traditional algorithm.
作者
彭岑昕
程伟
李晓柏
张永利
PENG Cenxin;CHENGWei;LI Xiaobai;ZHANG Yongli(Air Force Early Warning Academy,Wuhan 430019,China;No.95209 Unit,the PLA,Changsha 410004,China)
出处
《空军预警学院学报》
2020年第1期39-45,共7页
Journal of Air Force Early Warning Academy
关键词
调制方式识别
短时傅里叶变换
双向长短期记忆网络
modulation recognition
short time Fourier transform(STFT)
bidirectional long short-term memory network(BiLSTM)