基于盲压缩感知的非合作跳频信号处理技术研究

发布时间：2019-02-26 20:23

【摘要】：进入21世纪以来,战争形态已由原来的机械化逐步向信息化转变,双方的对抗也不再局限于战略装备的比拼,高效精准的信息处理手段成为决定战争走向的至关重要因素。非合作跳频信号因具良好的抗远近效应及干扰鲁棒性,在信息化战争中得以推广,研究非合作跳频信号相关特性,寻找有效的拦截,识别及恢复算法,是当前军事通信领域备受瞩目的课题之一。随着硬件设施的成熟完善,非合作跳频信号逐渐向“高跳速,多频点,超带宽”方向发展,信号波形复杂度急剧上升,所包含的先验信息量却微乎其微,这些都对现代信号处理手段提出空前挑战。盲压缩感知理论是一种基于投影测量,稀疏编码及变换基构造过程提出的采样重构框架。一方面,该理论保留了传统压缩感知理论压缩采样的优势,利用投影测量对信号中的有用信息进行提取,在不破坏目标信号结构的基础上,最大限度的压缩了数据量;另一方面,该理论融合字典学习过程,通过交替迭代找到一对乘积最优的稀疏矩阵及变换基,用来恢复原始信号结构。盲压缩感知理论真正实现了无先验信息重构理念,对非合作通信的发展提供了重要理论支撑。本文进行了基于盲压缩感知的非合作跳频信号处理技术的研究,内容包括信号模型建立,盲压缩重构及跳变参数估计三方面内容。本文首先介绍了跳频技术提出及推广过程,总结了现有的跳频信号处理方法。介绍了盲压缩感知的提出过程以及相应的公式定理,总结了使其满足唯一性的三个限制条件及所对应的最优算法,并分析比较了三种算法各自的特性。然后介绍了非合作跳频通信系统组成部分、重要参数指标及调制方式。根据非合作跳频信号的定义式,选取恰当的参数值,得到信号的仿真结果。引入盲压缩感知理论来还原信号结构,根据实验数据,分析各项参数对还原质量影响。最后,在精确重构的基础上,对非合作跳频信号进行频域研究,提出针对跳变频率及跳变周期的估计方法,根据仿真数据,分析了影响估计质量的因素。
[Abstract]:Since the beginning of the 21st century, the form of war has gradually changed from the original mechanization to the informatization. The confrontation between the two sides is no longer limited to the competition of strategic equipment, and the efficient and accurate means of information processing has become a crucial factor in determining the direction of the war. Non-cooperative frequency hopping signals have been extended in information warfare because of their good robustness against far and near effects and interference. The correlation characteristics of non-cooperative frequency hopping signals are studied, and effective interception, identification and recovery algorithms are found. It is one of the most attractive topics in the field of military communication. With the maturity of hardware facilities, non-cooperative frequency hopping signals gradually develop to the direction of "high speed hopping, multi-frequency points, super-bandwidth", and the complexity of signal waveforms increases sharply, but the prior information contained is very little. All of these present unprecedented challenges to modern signal processing methods. Blind compression sensing theory is a sampling reconstruction framework based on projection measurement, sparse coding and transform base construction process. On the one hand, the theory retains the advantages of the traditional compression sensing theory of compression sampling, using projection measurement to extract useful information from the signal, on the basis of not destroying the structure of the target signal, maximum compression of the amount of data; On the other hand, the theory combines dictionary learning process and finds a pair of optimal product sparse matrix and transform basis by alternating iteration, which can be used to restore the original signal structure. Blind compression perception theory really realizes the idea of non-prior information reconstruction, which provides an important theoretical support for the development of non-cooperative communication. In this paper, the non-cooperative frequency hopping signal processing technology based on blind compression sensing is studied, including the establishment of signal model, the reconstruction of blind compression and the estimation of jump parameters. This paper first introduces the introduction and popularization of frequency hopping technology, and summarizes the existing frequency hopping signal processing methods. This paper introduces the process of the blind compression perception and the corresponding formula theorem, summarizes the three restriction conditions and the corresponding optimal algorithm to make it satisfy the uniqueness, and analyzes and compares the respective characteristics of the three algorithms. Then, the components, important parameters and modulation methods of non-cooperative frequency hopping communication system are introduced. According to the definition of non-cooperative frequency hopping signal, the appropriate parameter values are selected and the simulation results are obtained. The blind compression sensing theory is introduced to restore the signal structure. According to the experimental data, the influence of the parameters on the reduction quality is analyzed. Finally, on the basis of accurate reconstruction, the non-cooperative frequency hopping signal is studied in frequency domain, and the estimation method of hopping frequency and period is proposed. According to the simulation data, the factors affecting the estimation quality are analyzed.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN911.7;E96

【相似文献】