摘要
针对典型的水下数字通信信号检测与调制识别问题,本文提出了基于特征提取的神经网络算法。通过对接收信号进行离散傅里叶变换与奇异谱分析,及利用提取的特征训练分类模型,成功实现了对MFSK,MPSK与对称α稳态分布三种情况的正确分类。最后,本文仿真了具有随机多径参数的水声信道,结果显示模型仍具有良好的分类效果。
Aiming at the problem of typical underwater digital communication signal detection and modulation recognition,this paper proposes a neural network algorithm based on feature extraction.Through the discrete Fourier transform and singular spectrum analysis of the received signal,and using the extracted feature to train classification model,the correct classification of MFSK,MPSK and symmetricαsteady state distribution is successfully realized.Finally,this paper simulates the underwater acoustic channel with random multipath parameters,and the result shows that the model still has good classification effect.
作者
郑飞洋
张铁军
ZHENG Feiyang;ZHANG Tiejun(Key Laboratory of Information Technology for Autonomous Underwater Vehicles,Institute of Acoustics,Chinese Academy of Sciences,Beijing,100190,China;University of Chinese Academy of Sciences,Beijing,100049,China)
出处
《网络新媒体技术》
2019年第2期46-52,共7页
Network New Media Technology
关键词
水下数字通信
对称a稳态分布
奇异谱分析
神经网络
Underwater digital communication
Symmetric a steady-state distribution
Singular spectrum analysis
Neural network
