Automatic Recognition of Communication Signal
Modulation based on Neural Network
Xiaolei Zhu, Yun Lin, Zheng Dou
College of Information and Communication Engineering
Harbin Engineering University
Harbin, China
linyun@hrbeu.edu.cn
Abstract—In order to solve the problem of low modulation
recognition rate of digital communication signals and the difficulty
of selecting the appropriate decision threshold, the paper features a
recognition method for communication signal modulation. The
paper constructs characteristic parameters for recognizing signals
in the cyclic frequency domain, and uses a 3-layer neural network
as a classifier to identify the modulation mode. The experiment
indicates that it can recognize 2FSK, 4FSK, 8FSK, BPSK, QPSK,
MSK and 2ASK. When signal to noise ratio (SNR) is higher than 0
dB, the recognition rate achieves 95%. The results suggest that
recognition of communication signal modulation based on neural
network is accurate and feasible.
Keywords- modulation recognition; cyclic feature; neural
network
I. INTRODUCTION
Communication environment has been more and more
complex with the developing of communication, and the signal
modulation mode becomes multitudinous to satisfy a large
number of user’s requirement. The automatic recognition of
signal modulation refers to identification of a modulated signal
with noise, and to ensure demodulation and feature extraction,
which plays an important role in the military intelligence
intercepted, electronic warfare, electronic reconnaissance, and
other areas [1]. In the military field, we must judge the enemy’s
signal modulation type and estimate some critical parameters to
disturb and intercept enemy’s information. There are also many
significant use in civilian areas such as detect the illegal radio
and surveillance whether the parameter configuration of legal
radio follows standard. For the more, automatic recognition of
signal modulation is vital in cognitive ratio.
The modulation recognition is mainly divided into three
processes: data pre-processing, feature extraction and
classificatory decision. Data pre-processing is the estimation of
carrier and symbol rate after signal has been down-conversion,
providing several appropriate data for follow-up operate. The use
of feature extraction is to transform original data to extract some
features which could be classified more easily. Classificatory
decision is to judge the modulation mode according to the
features extracted. In this paper, features in cyclic frequency
domain are regarded as characteristic parameters, and the
classifier is designed based on neural network.
The rest of this paper is organized as follows. In section 2,
we briefly introduce the cyclic spectrum and characteristic
parameters that we would use in the next process. Section 3
describes a BP network as a classifier. The simulation and
analysis are presented in section 4. Section 5 concludes this
paper.
II. CHARACTERISTIC PARAMETERS
A. The definition of cyclic spectrum
Cyclostationarity is a kind of stochastic process in which the
statistical characteristics periodic change with time [2]. We know
by the definition of a cyclostationary signal that the mean of a
signal, M
x
, and the autocorrelation of a signal, R
x
, are periodic [3].
Let’s define the period by being T and the lag being τ, then:
)()( nTtMtM
xx
);();(
TtRtR
xx
Since
);(
tR
x
is periodic, it can be expressed as a sum of a
series of Fourier series, and Fourier series coefficients could be
written as is shown in the following formula [4].
tj
xx
eRtR
2
)();(
2
2
2
);(
1
)(
T
T
tj
xx
dtetR
T
R
where α=m/T, and Fourier series coefficients
)(
x
R
are called
cyclic autocorrelation function. We can get following formula by
expanding the autocorrelation function:
2
2
2*
0
)2/()2/(
1
lim)(
T
T
tj
T
x
dtetXtX
T
R
下载后可阅读完整内容，剩余3页未读，立即下载
weixin_38545923
- 粉丝: 4
- 资源: 933
我的内容管理
展开
最新资源
- 李兴华Java基础教程：从入门到精通
- U盘与硬盘启动安装教程：从菜鸟到专家
- C++面试宝典：动态内存管理与继承解析
- C++ STL源码深度解析：专家级剖析与关键技术
- C/C++调用DOS命令实战指南
- 神经网络补偿的多传感器航迹融合技术
- GIS中的大地坐标系与椭球体解析
- 海思Hi3515 H.264编解码处理器用户手册
- Oracle基础练习题与解答
- 谷歌地球3D建筑筛选新流程详解
- CFO与CIO携手：数据管理与企业增值的战略
- Eclipse IDE基础教程：从入门到精通
- Shell脚本专家宝典：全面学习与资源指南
- Tomcat安装指南：附带JDK配置步骤
- NA3003A电子水准仪数据格式解析与转换研究
- 自动化专业英语词汇精华：必备术语集锦
资源上传下载、课程学习等过程中有任何疑问或建议，欢迎提出宝贵意见哦~我们会及时处理！
点击此处反馈
