摘要
信号的盲源分离一直是一项重要的工作,信号源数估计是盲源分离的重要步骤。对单通道接收信号的源数估计方法进行了研究,通过将信号变换到时频域,引入卷积神经网络来进行问题求解。针对该时频图像的特点,为源数估计设计了相应的卷积神经网络,提出了一种3CPB模型,能初步达到较高的源数估计的准确度。模型可在有色噪声的干扰下、不同信噪比情况下较好地完成信号源数目估计的任务,为源数估计方法提供了一种新的技术路线。
The blind source separation of signals has always been an important task. The estimation of signal source is an important step in blind source separation. The source number estimation method of single-channel received signals is studied, and by transforming the signal into the time-frequency domain, a convolutional neural network is introduced to solve the problem. According to the characteristics of the time-frequency image, the corresponding convolutional neural network is designed for the source number estimation, and a 3 CPB model is proposed to achieve the high accuracy of the source number estimation. The model can better accomplish the task of estimating the number of signal sources under the interference of colored noise and different signal-to-noise ratio, which provides a new technical route for the source number estimation method.
作者
麻凯利
王川川
Ma Kaili;Wang Chuanchuan(State Key Laboratory of Complex Electromagnetic Environment Effects on Electronics and Information System,Luoyang 471003,Henan,China)
出处
《航天电子对抗》
2019年第3期7-11,共5页
Aerospace Electronic Warfare
关键词
盲源分离
源数估计
深度学习
卷积神经网络
blind source separation
source number estimation
deep learning
convolutional neural networks
