跳频信号参数估计与跳频序列预测方法研究

发布时间：2018-01-21 19:23

本文关键词： 跳频通信跳频序列参数估计序列预测　出处：《电子科技大学》2014年硕士论文　论文类型：学位论文

【摘要】：跳频通信因具有抗干扰能力强、截获率低等特点而被广泛应用,尤其是在军事领域。与传统的抗干扰方式不同,它通过伪随机码控制载波的跳变,从而有效躲避干扰信号,克服了定频通信的缺陷。然而随着军事电子对抗的升级,如何有效干扰、截获乃至利用对方跳频信号已成为目前研究的一个热点问题。目前研究该问题的方法通常是:首先,对接收到的跳频信号进行参数估计,得到跳频周期、跳频时刻、跳频频率等相关参数;其次,利用估计出来的跳频频率序列进行建模预测,估计出未来时刻的频率;最后,借助同步跟踪算法实现信号的同步。本文主要研究前两个环节,即跳频信号参数估计和跳频序列预测。跳频信号参数估计作为首要环节,其估计性能直接影响整个系统的性能。在研究基本时频分析方法的基础上,重点介绍了常用参数估计方法,综合分析了基于STFT和SPWVD参数估计的性能。在此基础上,针对跳频信号这种特殊的非平稳信号,改进了一种基于STFT的参数估计方法,该方法直接从STFT窗函数提取跳频信号参数,避开了STFT时频分辨率低的限制,提高了参数估计的精度,仿真结果表明该方法可以实现参数的有效估计。跳频序列预测作为中间环节,其预测性能尤为关键。文中主要研究了基于移位寄存器产生的跳频序列和基于混沌时间序列生成的跳频序列的预测方法。针对前者,在研究基于其产生机理预测方法的基础上,改进了一种基于Berlekamp-Massey算法的移位寄存器序列预测方法,实现线性移位寄存器连续与非连续抽头序列的有效预测;至于后者,混沌跳频序列因保密性强、数量巨大而具有巨大优势,目前对该类序列的预测基本上是在相空间进行的。因此在研究相空间重构的基础上,综合分析局域预测法性能,重点研究了基于RBF神经网络、Bernstein多项式、Volterra自适应滤波和支持向量机的混沌跳频序列预测方法,通过理论分析和仿真实验比较其预测性能。同时考虑到工程实时性的需要,探讨了直接和间接多步预测法,并在仿真基础上分析其优劣。最后,讨论了非混沌同步跳对预测的影响,在具体应用预测模型时应通盘考虑。
[Abstract]:Frequency hopping communication is widely used because of its strong anti-jamming ability and low interception rate, especially in the military field. Different from the traditional anti-jamming mode, it controls the carrier jump by pseudo-random code. In order to effectively avoid interference signals, overcome the defects of fixed-frequency communication. However, with the upgrading of military electronic countermeasures, how to effectively interfere. Interception and even utilization of frequency hopping signals has become a hot issue. At present, the methods to study this problem are as follows: firstly, the parameters of the received frequency hopping signals are estimated and the frequency hopping period is obtained. Frequency hopping time, frequency hopping frequency and other related parameters; Secondly, the estimated frequency hopping sequence is used to model and predict the frequency of the future time. Finally, the synchronization algorithm is used to realize signal synchronization. In this paper, the first two links, namely frequency hopping signal parameter estimation and frequency hopping sequence prediction, are studied. The frequency hopping signal parameter estimation is the first step. The estimation performance directly affects the performance of the whole system. Based on the study of the basic time-frequency analysis methods, the commonly used parameter estimation methods are emphatically introduced. The performance of parameter estimation based on STFT and SPWVD is analyzed synthetically. Based on this, a parameter estimation method based on STFT is improved for frequency hopping signal, which is a special non-stationary signal. This method extracts the parameters of frequency hopping signal directly from the STFT window function, avoids the limitation of low time-frequency resolution of STFT, and improves the precision of parameter estimation. The simulation results show that this method can effectively estimate the parameters, and the frequency hopping sequence prediction is the intermediate link. In this paper, the prediction methods of frequency hopping sequence based on shift register and frequency hopping sequence based on chaotic time series are studied. A shift register sequence prediction method based on Berlekamp-Massey algorithm is improved on the basis of studying the prediction method based on its generation mechanism. The effective prediction of continuous and discontinuous tap sequences of linear shift registers is realized. As for the latter, chaotic frequency hopping sequences have great advantages because of their strong confidentiality and huge quantity. At present, the prediction of chaotic frequency hopping sequences is basically carried out in the phase space, so it is based on the study of phase space reconstruction. The performance of local prediction method is analyzed synthetically, and the Bernstein polynomial based on RBF neural network is studied emphatically. The prediction method of chaotic frequency hopping sequence based on Volterra adaptive filter and support vector machine is compared by theoretical analysis and simulation. The direct and indirect multistep prediction methods are discussed, and their advantages and disadvantages are analyzed on the basis of simulation. Finally, the influence of non-chaotic synchronous hopping on prediction is discussed, which should be taken into account in the application of the prediction model.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN914.41

【参考文献】