摘要
摘要
通信信号的分类与识别是截获信号处理领域的重要研究课题，它需要在有噪声干扰的复杂环境中判断出信号的调制模式，为下一步的分析处理环节提供坚实根据。战术电台网络是美国军方现役的战场通信网络系统，它将交换路由设备、信息终端和各种独立的战术电台等互相联结，形成了一个战役战术一体化的通信系统，因此在第三方进行侦查时，战术电台网络中通信信号的分类与识别在电子对抗中非常重要。
课题依托国家级项目，针对战术电台网络通信信号进行分类与识别，主要工作可概括如下：
1.在没有提供其他先验信息的情况下，对战术电台网络信号进行了参数估计，包括跳频估计、码元速率估计，并利用获得的瞬时参数分析、循环谱分析与小波包分解重构相融合的方法，提取出了多种特征参数以进行调制分类。仿真结果表明，该组参数能够有效实现对战术电台网络信号的调制分类。
2.基于MATLAB仿真平台，针对上一步骤获得的特征参数组部分参数，利用支持向量机模型进行仿真，通过参数优化解决了过学习问题；利用获得的特征参数组，分别采用了支持向量机模型以及径向基神经网络模型对调制信号进行分类。仿真结果表明，支持向量机的性能要优于径向基神经网络的性能：在高信噪比下，二者性能都很好；但在低信噪比下，前者依然具有较高的准确率，后者的性能下滑严重。
关键词：战术电台网络，调制识别，特征提取，支持向量机，神经网络
ABSTRACT
Classification of the military communication signal’s modulation mode is an important research hot in the signal processing field. The modulation mode needs to be recognized in the complex circumstance with noise and interference, if that is done, the next step of signal processing could carry on. Now, Tactical Radio Network is the battlefield communication network of America, it is based on the wireless communication, interconnects the tactical radio, information terminal and routing device, and it is the integrated battle and tactical co mmunication system for US Army’s digital battlefield. As the third party to investigate, the classification and recognition of the communication signal in the tactical radio network is very important in the electronic countermeasure. This thesis relies on the national project, mainly focusing on the classification and recognition of the communication signals in the Tactical Radio Network.
The main work of the thesis can be summarized as follows:
1. The parameters of the received Tactical Radio Network signal without prior knowledge have been estimated, including estimation of frequency hopping, symbol rate estimation. And the modulation cla
ssification of the Tactical Radio Network signals can be realized with this thesis’method that combined the instantaneous parameters, the parameters of the cyclic spectrum analysis and the wavelet packet decomposition and reconstruction parameters. The simulation results show that the support vector machine used as the classification model has effectively realized the modulation classification of Tactical Radio Network signals.
2. Based on the MATLAB simulation platform, some parameters in the characteristic parameter set mentioned above have been utilized to solve the problem of over learning which is existed in the support vector machine model. The support vector machine algorithm and radial basis function neural network algorithm have been respectively used to classify the modulation signals. Simulation results have showed that it can effectively realize the modulation identification of tactical radio network communication signals with the proposed method in this thesis. The performance of support vector machine model is better than radial basis function neural network model. Under high signal to noise ratio, the performances of them are
both very good. But under low signal to noise ratio, the former still has brilliant recognition accuracy, the performance of the latter declined significantly.
Keywords: Tactical Radio Network，modulation identification，features extraction，support vector machine，radial basis function neural network
目录
第一章绪论 (1)
1.1 研究背景及意义 (1)
1.2 研究现状 (2)
1.2.1 最大似然比假设检验 (3)
未识别的网络1.2.2 统计模式识别 (5)
1.3 本文的研究内容 (8)
第二章战术电台网络 (9)
2.1 战术电台网络概述 (9)
2.2 战术电台网络结构 (10)
2.2.1 上层战术电台网络 (10)
2.2.2 下层战术电台网络 (10)
2.3 战术电台网络三大重要组成 (13)
2.3.1 SINCGARS电台 (13)
2.3.2 EPLRS电台 (14)
2.3.3 NTDR电台 (14)
2.4 美军部分现役SINCGARS电台 (16)
2.5 战术电台网络信号调制方式 (16)
2.6 本章小结 (17)
第三章战术电台网络信号的参数估计 (18)
3.1 概述 (18)
3.2 基于STFT的跳频参数估计 (18)
3.2.1 跳频通信的基本原理和模型 (18)
3.2.2 跳频系统的特征参数 (19)
3.2.3 短时付里叶变换(STFT) (20)
3.2.4 仿真分析 (21)
3.3 基于小波变换的码元速率估计 (22)
3.3.1 理论分析 (22)
3.3.2 仿真分析 (24)
3.4 基于信号功率谱的信噪比估计 (26)
3.5 本章小结 (27)
第四章基于机器学习的战术电台网络信号的调制分类与识别 (28)
4.1 战术电台网络通信信号特征参数提取 (28)
4.1.1 特征参数提取常用方法 (28)
4.1.2 基于瞬时参数的特征提取方法 (29)
4.1.3 基于循环谱相关参数的特征提取方法 (31)
4.1.4 基于小波变换的特征提取方法 (37)
4.2 基于支持向量机的信号分类识别 (39)
4.2.1 支持向量机原理 (39)
4.2.2 仿真分析 (43)
4.3 基于RBF神经网络的信号分类识别 (48)
4.3.1 神经网络原理 (48)
4.3.2 仿真分析 (51)
4.4 算法对比分析 (52)
4.5 本章小结 (54)
第五章总结与展望 (55)
致谢 (56)
参考文献 (57)