通信信号自动调制识别的研 - 图文

摘要

摘 要

通信信号调制类型的分类识别是一种典型的模式识别问题，它涉及到很多复杂的特殊因素。随着通信技术的飞速发展，通信信号的体制和调制样式变的更加复杂多样，信号环境日益密集，使得常规的识别方法和理论很难适应实际需要，无法有效的对通信信号进行识别，这也给通信信号的识别研究提出了更高的要求。

近几十年来人们在通信信号的识别方面作了大量有益的探索，提出了很多新思路和新方法，但是这些方法都是基于固定的信噪比，没有涉及信噪比变化时的信号识别问题。实际上，通信信号经过无线信道的传输，信噪比变化范围较大，通常在几分贝到几十分贝的范围内变化，这将导致从同一类信号的不同信噪比样本中提取的同一种特征有可能产生严重的畸变，相当于成倍增加了待识别信号的类别，使分类器的识别率降低。

本文主要工作体现在瞬时特征参数的提取、模糊特征选择和分类器设计这三个方面。其创新之处在于：

1、研究瞬时参数提取的目的就是减少噪声对瞬时参数的影响，使基于瞬时参数提取的特征参数对噪声不敏感。本文主要研究基于小波脊、短时傅里叶脊的信号瞬时参数提取方法、基于小波变换的瞬时参数提取方法和基于自适应时频分析的瞬时参数提取方法。

2、为了简化分类器的设计，提高分类能力和效率，本文给出了一种新的特征选择准则，即基于模糊特征估计准则。用于寻找最优特征的算法采用模糊遗传算法。

3、为了改善分类器的推广能力，本文着重研究了模糊神经网络分类器和模糊支持向量机分类器设计方法。为了实现大信噪比变化范围内通信信号的有效识别，本文提出了基于模糊积分和神经网络的组合分类器设计方法，基于模糊积分和支持向量机组合分类器设计方法和基于类间距离和模糊神经网络组合分类器设计方法。

关键词：自动调制识别、瞬时特征参数提取、模糊特征选择、分类器设计

I

ABSTRACT

ABSTRACT

The classification of modulation types of communication signals is a problem of typical pattern recognition. It involves many perplexing and special factors. With rapidly developing of communication technology, the system and modulation manner of communication signals became more and more complicated and various, and circumstance of signals became increasing denseness. It results in that the routine methods and theory of recognition can hardly satisfy practical requirement and can`t effectively recognize for communication signals. So the strict demand has been presented for study on recognition of communication signals.

For the last several decade years, the people have helpfully explored many methods of solving the question of recognition of communication signals. However these methods were presented in condition of fixed signal noise ratio(SNR),and did not involve the problem of signal recognition when SNR was changed. In practice, when communication signals are transmitted by wireless channel, the variation range of SNR is very large, and it is generally between several dB and several ten dB. The result is that the serious distortion of the same sort feature extracted from the different SNR samples for same type signals is caused. It is equal to increase multiply types of the recognized signals, and the recognition probability of classification is reduced.

The main contribution of this dissertation includes three aspects. They are instantaneous feature parameters extraction, fuzzy feature selection and classifier design.

Several valuable and important results which bring forth new ideas are achieved and listed as the following:

1. The goal of studying instantaneous parameters extraction is to reduce the noise influence of instantaneous parameters and make feature parameters extracted from instantaneous parameters insensitive to noise. This dissertation mainly study the extraction methods that are built on wavelet ridge, short time Fourier ridge, wavelet

II

ABSTRACT

transform and adaptive time-frequency analysis respectively.

2. In order to simplify the classifier design and improve the ability and efficiency of classifier, a new feature selection method is studied in this dissertation. It is fuzzy feature evaluation method. The algorithm of finding optimal feature is fuzzy genetic algorithm.

3. In order to improve the generalized ability of classifier, fuzzy network classifier and fuzzy support vector machine classifier are proposed in the dissertation. For efficient recognition of communication signals in large variation range of SNR, three new methods which are based on fuzzy integral and neural network, fuzzy integral and support vector machine, interclass distance and fuzzy neural network respectively, are used to design combined classifier.

Keywords: Automatic modulation identification, Instantaneous feature parameters extraction, Fuzzy feature selection, Classifier design

III

目 录

第一章 绪 论 ......................................................1 1.1 通信信号识别概述 ..............................................1 1.2 发展简史和研究概述（国内外研究状况） ..........................2 1.3 本论文的主要工作 ..............................................4 1.4 本文的章节安排 ................................................4 第二章 通信信号的特征提取与选择 ...................................6 2.1 几种常用的通信信号 ............................................6 2.2 常用通信信号的特征提取 .......................................11 2.3 基于模糊遗传算法的特征选择 ...................................16

2.3.1 基于模糊特征估计的特征选择准则 ..........................16 2.3.2 模糊遗传算法 ............................................18 2.4 小结 .........................................................20 第三章 通信信号的瞬时参数提取 ....................................22 3.1 基于短时傅里叶脊的瞬时参数提取 ...............................22

3.1.1 短时傅里叶变换 ..........................................22 3.1.2 短时傅里叶脊与瞬时参数的关系 ............................23 3.1.3 基于奇异值分解的短时傅里叶脊的确定 ......................24 3.1.4 仿真实验 ................................................25 3.2 基于小波变换的瞬时参数提取 ...................................28

3.2.1 小波变换 ................................................28 3.2.2 基于小波变换的瞬时参数提取 ..............................29 3.2.3 基于小波脊的瞬时参数提取 ................................29

IV

3.2.4 仿真实验 ................................................31 3.3 基于高分辨自适应时频分析的瞬时参数提取 .......................36

3.2.1 Capon波束形成算法 ......................................36 3.2.2 高分辨自适应时频分析 ....................................38

3.2.2.1 基于Capon的短时傅里叶表示 ........................39 3.2.2.2 基于Capon的wigner时频表示 .......................40 3.3.3 仿真实验 ................................................41 3.4 小结 .........................................................47 第四章 单个分类器设计 ............................................48 4.1 基于神经网络的分类器设计 .....................................48

4.1.1 MLP神经网络 ...........................................49 4.1.2 仿真实验 ................................................51 4.2 基于模糊神经网络的分类器设计 .................................53

4.2.1 模糊逻辑系统 ............................................53 4.2.2 模糊神经网络 ............................................55 4.2.3 仿真实验 ................................................57 4.3 基于支持向量机的分类器设计 ...................................59

4.3.1 统计学习理论 ............................................60 4.3.2 支持向量机 ..............................................62 4.3.3 一对一多类模糊支持向量机分类器 ..........................64 4.3.4 一对多多类模糊支持向量机分类器 ..........................65 4.3.5 仿真实验 ................................................66 4.4 小结 .........................................................68 第五章 组合分类器设计 ............................................69 5.1 基于神经网络的组合分类器设计 .................................69

5.1.1 模糊积分理论 ............................................69 5.1.2 模糊积分组合分类器设计 ..................................71 5.1.3 仿真实验 ................................................73

V

百度搜索“77cn”或“免费范文网”即可找到本站免费阅读全部范文。收藏本站方便下次阅读，免费范文网，提供经典小说综合文库通信信号自动调制识别的研 - 图文在线全文阅读。