扩谱通信抗干扰的现代信号处理应用研究
发布时间：2018-10-30 13:25
【摘要】：凭借优异的信息隐藏能力和易于组网等优点,扩谱通信技术已成为抗干扰通信中的主流方法。单纯依赖扩谱抗干扰技术难以适应未来战场复杂电磁环境下的军事通信需求。为了进一步提高现有通信装备的抗干扰能力,本文围绕扩谱通信抗干扰中的现代信号处理方法及相关扩展应用展开研究。 本文首先探讨了利用盲源分离混合矩阵逆问题求解方法进行扩谱通信抗干扰的可行性,设计了点对点和组网方式下基于盲源分离的扩谱通信抗干扰系统结构,提出了结合负熵对照函数的盲源分离直接序列扩谱抗相关干扰算法。为提高强噪声环境下的抗干扰效果,充分利用实际采样速率较高的系统特性,设计了基于均值滤波的盲源分离抗相关干扰算法。分析了非协作盲接收环境下用户异步时延的构成,从前后符号在当前接收窗口的有效能量角度推导了用户异步时延对抗干扰算法收敛性的影响,并利用三角不等式进行了论证。 考虑跳频载频的跳变对混合信号统计独立性的影响,设计了基于盲源分离的跳频通信系统抗干扰算法,并对各种类型干扰进行了算法性能仿真及硬件测试。根据跳频信号与各种干扰信号之间时频分布均不相同的特性,提出了基于时频联合的盲源分离抗干扰算法,从矩阵的联合对角化出发实现了跳频信号与干扰信号的有效分离。基于跳频图案等先验信息,设计了基于跳频图案的半盲分离跳频通信抗干扰算法。 根据扫频干扰信号的特点,设计了匹配扫频信号的非线性变换,提出实时扫频信号特征检测方法,通过自学习获得了能量聚集特性优异的扫频干扰压缩感知字典。根据多音干扰与跳频信号在时间、频率分布上不一致,分别提出了自学习多音干扰压缩感知字典和自学习跳频信号压缩感知字典构造方法,使得低信噪比时能够识别多音干扰和正常的跳频信号,为实时可靠地解调创造了有利条件。 将基于形态结构特征学习的字典构造与压缩感知方法结合,提出多形态自学习压缩感知跳频通信抗干扰算法。利用自学习压缩感知字典的稀疏表达能力,设计了单通道自学习压缩感知跳频信号抗干扰算法,在自学习过程中将检测到的强干扰分量去除以减小干扰影响。理论分析和仿真验证表明本文算法在低信噪比和强干扰条件下具有优异而高效的抗干扰能力。 最后,总结了信号自身内在形态结构特征的广泛存在性,提出匹配信号形态特征的稀疏表达域自学习构造方法,从而可以在较小的字典规模下高效地进行信号的重构,为压缩感知方法实用化进行了有益的探索,并拓展应用于图像超分辨率重构,获得了优异的重构效果。
[Abstract]:With the advantages of excellent information hiding ability and easy networking, the spread spectrum communication technology has become the mainstream method in anti-interference communication. It is difficult to meet the requirements of military communication in the complex electromagnetic environment of the future battlefield by relying solely on the spread spectrum anti-jamming technology. In order to further improve the anti-jamming capability of the existing communication equipment, this paper focuses on the modern signal processing methods and related extended applications in the spread spectrum communication. In this paper, we first discuss the feasibility of solving the inverse problem of blind source separation hybrid matrix for anti-jamming of spread spectrum communication, and design the anti-jamming system structure of spread spectrum communication based on blind source separation under point-to-point and networking mode. A blind source separation (BSS) direct sequence spread spectrum (DSS) anti-correlation interference algorithm combining negative entropy contrast function (NSE) is proposed. In order to improve the anti-interference effect in strong noise environment and make full use of the high sampling rate of the system, a blind source separation anti-correlation interference algorithm based on mean filter is designed. This paper analyzes the structure of user asynchronous delay in non-cooperative blind reception environment, deduces the influence of user asynchronous delay on the convergence of anti-jamming algorithm from the point of view of the effective energy of front and rear symbols in the current receiving window, and proves it by using triangular inequality. Considering the influence of hopping frequency hopping on the statistical independence of mixed signals, a frequency hopping communication system based on blind source separation is designed, and the algorithm performance simulation and hardware test are carried out. According to the different time-frequency distribution between frequency-hopping signals and various interference signals, an anti-jamming algorithm for blind source separation based on time-frequency joint is proposed. The effective separation of frequency-hopping signal and interference signal is realized from the diagonalization of matrix. Based on the prior information such as frequency hopping pattern, a semi-blind anti-jamming algorithm for frequency hopping communication is designed. According to the characteristics of the frequency-sweeping interference signal, the nonlinear transformation of the matched frequency-sweeping signal is designed, and the method of detecting the feature of the real-time frequency-sweeping signal is proposed. Through self-learning, a dictionary of scan interference compression perception with excellent energy aggregation characteristic is obtained. According to the difference of time and frequency distribution between multi-tone interference and frequency-hopping signal, a self-learning multi-tone interference compression perception dictionary and a self-learning frequency-hopping signal compression perception dictionary are proposed, respectively. It makes it possible to recognize multi-tone interference and normal frequency-hopping signal at low signal-to-noise ratio (SNR), which creates favorable conditions for real-time and reliable demodulation. By combining the dictionary construction based on morphological structure feature learning with the compression sensing method, a multi-morphological self-learning compression sensing frequency hopping communication anti-jamming algorithm is proposed. Using the sparse expression ability of self-learning compression perceptual dictionary, a single-channel self-learning compression sensing frequency hopping signal anti-jamming algorithm is designed. In the process of self-learning, the detected strong interference components are removed to reduce the interference effect. Theoretical analysis and simulation results show that the proposed algorithm has excellent and efficient anti-jamming capability under the condition of low SNR and strong interference. Finally, this paper summarizes the extensive existence of the inherent morphological structure features of the signal itself, and proposes a sparse expression domain self-learning construction method to match the morphological features of the signal, so that the signal can be reconstructed efficiently on a small dictionary scale. This paper makes a useful exploration for the practical application of compression sensing, and extends its application to image super-resolution reconstruction, and obtains excellent reconstruction results.
【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN975
本文编号：2300127
[Abstract]:With the advantages of excellent information hiding ability and easy networking, the spread spectrum communication technology has become the mainstream method in anti-interference communication. It is difficult to meet the requirements of military communication in the complex electromagnetic environment of the future battlefield by relying solely on the spread spectrum anti-jamming technology. In order to further improve the anti-jamming capability of the existing communication equipment, this paper focuses on the modern signal processing methods and related extended applications in the spread spectrum communication. In this paper, we first discuss the feasibility of solving the inverse problem of blind source separation hybrid matrix for anti-jamming of spread spectrum communication, and design the anti-jamming system structure of spread spectrum communication based on blind source separation under point-to-point and networking mode. A blind source separation (BSS) direct sequence spread spectrum (DSS) anti-correlation interference algorithm combining negative entropy contrast function (NSE) is proposed. In order to improve the anti-interference effect in strong noise environment and make full use of the high sampling rate of the system, a blind source separation anti-correlation interference algorithm based on mean filter is designed. This paper analyzes the structure of user asynchronous delay in non-cooperative blind reception environment, deduces the influence of user asynchronous delay on the convergence of anti-jamming algorithm from the point of view of the effective energy of front and rear symbols in the current receiving window, and proves it by using triangular inequality. Considering the influence of hopping frequency hopping on the statistical independence of mixed signals, a frequency hopping communication system based on blind source separation is designed, and the algorithm performance simulation and hardware test are carried out. According to the different time-frequency distribution between frequency-hopping signals and various interference signals, an anti-jamming algorithm for blind source separation based on time-frequency joint is proposed. The effective separation of frequency-hopping signal and interference signal is realized from the diagonalization of matrix. Based on the prior information such as frequency hopping pattern, a semi-blind anti-jamming algorithm for frequency hopping communication is designed. According to the characteristics of the frequency-sweeping interference signal, the nonlinear transformation of the matched frequency-sweeping signal is designed, and the method of detecting the feature of the real-time frequency-sweeping signal is proposed. Through self-learning, a dictionary of scan interference compression perception with excellent energy aggregation characteristic is obtained. According to the difference of time and frequency distribution between multi-tone interference and frequency-hopping signal, a self-learning multi-tone interference compression perception dictionary and a self-learning frequency-hopping signal compression perception dictionary are proposed, respectively. It makes it possible to recognize multi-tone interference and normal frequency-hopping signal at low signal-to-noise ratio (SNR), which creates favorable conditions for real-time and reliable demodulation. By combining the dictionary construction based on morphological structure feature learning with the compression sensing method, a multi-morphological self-learning compression sensing frequency hopping communication anti-jamming algorithm is proposed. Using the sparse expression ability of self-learning compression perceptual dictionary, a single-channel self-learning compression sensing frequency hopping signal anti-jamming algorithm is designed. In the process of self-learning, the detected strong interference components are removed to reduce the interference effect. Theoretical analysis and simulation results show that the proposed algorithm has excellent and efficient anti-jamming capability under the condition of low SNR and strong interference. Finally, this paper summarizes the extensive existence of the inherent morphological structure features of the signal itself, and proposes a sparse expression domain self-learning construction method to match the morphological features of the signal, so that the signal can be reconstructed efficiently on a small dictionary scale. This paper makes a useful exploration for the practical application of compression sensing, and extends its application to image super-resolution reconstruction, and obtains excellent reconstruction results.
【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN975
【参考文献】
相关期刊论文 前3条
1 石光明;刘丹华;高大化;刘哲;林杰;王良君;;压缩感知理论及其研究进展[J];电子学报;2009年05期
2 彭耿;黄知涛;姜文利;周一宇;;单通道盲信号分离研究进展与展望[J];中国电子科学研究院学报;2009年03期
3 刘良云,李英才,相里斌;超分辨率图象重构技术的仿真实验研究[J];中国图象图形学报;2001年07期
，本文编号：2300127
本文链接：https://www.wllwen.com/kejilunwen/wltx/2300127.html
