宽带高速信号检测与频域测量技术研究

发布时间：2018-09-01 09:01

【摘要】：电子支援是一种获取敌方军事情报的特殊军事侦察手段,在雷达对抗领域发挥着极其重要的作用。随着雷达技术的快速发展,雷达的数量持续增长。现代电子战环境中的雷达信号变得日益密集,信号形式也变得复杂多样。信号的带宽越来越宽、信号的频率越来越高,这给传统奈奎斯特采样框架下的信号检测与参数测量带来了极大压力。随着信号向着宽带与高频的发展,传统的信号处理方法受采样器件的电子瓶颈限制,将难以满足未来战场需求。为了避免对宽带高频信号的高速采样,近年来国内外学者做了相关研究,其中,压缩感知理论与微波光子学在降低系统采样率上的应用受到了广泛的关注。在压缩感知理论下利用信号的稀疏性,用远低于奈奎斯特采样率的速率对信号进行采样,通过信号少量的压缩采样数据完成信号检测与参数估计的任务；微波光子技术将光子技术和微波技术结合,可以实现宽带宽、低损耗、抗电磁干扰的微波信号处理系统。本文旨在提高压缩感知框架下复杂电磁环境中的宽带线性调频(Linear Frequency Modulated, LFM)信号检测与参数估计性能和探索光采样技术在高频微波信号瞬时测频中的应用,做了相关的分析与研究,提出了一些解决方法。本文的主要贡献与研究内容包括：1.针对压缩感知框架下,基于波形匹配字典的LFM信号检测算法中检测性能受高斯白噪声影响较大的问题,提出一种基于分数阶Fourier变换(Fractional Fourier Transform, FrFT)字典的LFM信号检测算法。该算法利用LFM信号在分数阶Fourier变换域的能量聚集特性,首先构造分数阶Fourier变换矩阵作为信号的稀疏变换字典,然后利用正交匹配追踪算法,对信号在分数阶Fourier变换字典中的系数向量进行部分重构,最后通过对该系数向量进行检测判决,达到信号检测的目的。实验表明,与基于波形匹配字典的算法相比,该算法可以在更少的压缩采样点数和更低的信噪比(Signal-to-Noise Ratio, SNR)条件下获得更高的信号检测成功概率。2.针对压缩感知框架下,基于分数阶Fourier变换字典的LFM信号检测算法性能受窄带干扰信号影响严重的问题,引入形态学成分分析方法,提出一种基于级联字典的LFM信号检测算法。该算法基于形态学成分分析原理,构造分数阶Fourier变换字典和Fourier变换字典,将两者级联构成冗余字典,通过正交匹配追踪算法,对信号在级联字典中的系数向量进行部分重构,进而将信号与干扰分离。实验表明,该算法可以有效抑制窄带干扰信号,在窄带干扰信号存在的条件下,相比于基于分数阶Fourier字典的算法,该算法可以获得更高的信号检测成功概率。3.针对压缩感知框架下,基于波形匹配字典的LFM信号参数估计算法中参数估计性能受高斯白噪声和强窄带干扰信号影响严重的问题,提出了一种基于分数阶Fourier变换字典的两次搜索策略的LFM信号参数估计算法。该算法利用LFM信号在分数阶Fourier变换域的能量聚集特性,通过对变换阶次进行粗搜索与精搜索,得到信号的最佳分数阶Fourier变换阶次,进而得到LFM信号的调频斜率和起始频率的估计。实验表明,与基于波形匹配字典的算法相比,该算法能够在更少的压缩采样点数和更低的SNR的情况下,获得更高的信号参数估计成功概率。4.针对基于分数阶Fourier变换字典的两次搜索策略的参数估计算法中粗搜索步骤运行时间较长的问题,提出一种改进型的两次搜索策略的LFM信号参数估计算法。该算法利用LFM信号在时频域和分数阶Fourier变换域是稀疏信号的特性,首先对变换阶次在时频域进行粗搜索,然后根据阶次粗搜索值对变换阶次在分数阶Fourier变换域进行精搜索,得到信号最佳分数阶Fourier变换阶次,进而得到LFM信号调频斜率和起始频率的估计。实验表明,相比于基于分数阶Fourier变换字典的算法,该算法的粗搜索过程运行时间大大减少,且相比基于波形匹配字典的算法在,在相同SNR与压缩采样点数的条件下,该算法同样能够获得更高的信号参数估计成功概率。5.研究了基于微波光子学的高频微波信号瞬时测频算法。利用光采样过程中的宽带宽、低损耗和抗干扰等优点,提出一种基于光采样的高频微波信号瞬时测频算法。该算法利用光强度调制器将高频微波信号调制到低重复频率采样光脉冲上,进而达到对高频微波信号光采样的目的。将快速Fourier变换(Fast Fourier Transform, FFT)与线性调频z变换(Chirp-z Transform, CZT)结合,对欠采样条件下产生的频率余数进行准确估计,进而通过中国余数定理对信号频率进行重构。实验表明,该算法可以对39GHz带宽内信号频率进行准确测量。
[Abstract]:Electronic support is a special military reconnaissance means to obtain enemy's military information, which plays an extremely important role in the field of radar countermeasures. With the rapid development of radar technology, the number of radar continues to grow. In modern electronic warfare environment, radar signals become increasingly dense and the forms of signals become more complex. With the development of signal to broadband and high frequency, the traditional signal processing methods are limited by the electronic bottleneck of the sampled devices, which will be difficult to meet the future battlefield requirements. In recent years, scholars at home and abroad have done some research on the high-speed sampling of signals. Compressive sensing theory and microwave photonics have attracted wide attention in reducing the sampling rate of the system. Quantity compressed sampling data completes the task of signal detection and parameter estimation; microwave photon technology combines photon technology with microwave technology to realize wideband, low loss, anti-electromagnetic interference microwave signal processing system. Odulated, LFM) signal detection and parameter estimation performance and the application of optical sampling technology in the instantaneous frequency measurement of high frequency microwave signal are analyzed and studied. Some solutions are proposed. The main contributions and research contents of this paper include: 1. LFM signal detection algorithm based on waveform matching dictionary under compressed sensing framework. In this paper, a novel LFM signal detection algorithm based on Fractional Fourier Transform (FrFT) dictionary is proposed to solve the problem that the detection performance is greatly affected by Gaussian white noise. Transform the dictionary, and then use the orthogonal matching pursuit algorithm to reconstruct the coefficient vector of the signal in the fractional Fourier transform dictionary. Finally, detect and judge the coefficient vector to achieve the purpose of signal detection. The number of points and the lower SNR (Signal-to-Noise Ratio) can improve the probability of signal detection success. 2. In order to solve the problem that the performance of LFM signal detection algorithm based on fractional Fourier transform dictionary is seriously affected by narrowband interference signals, a morphological component analysis method is introduced and a Cascade-based LFM signal detection algorithm is proposed. Based on the principle of morphological component analysis, the algorithm constructs a fractional Fourier transform dictionary and a Fourier transform dictionary, cascades them into a redundant dictionary, and reconstructs the coefficient vector of the signal in the cascaded dictionary by orthogonal matching pursuit algorithm, and then separates the signal from the interference. It is shown that the algorithm can effectively suppress narrowband interference signals. Compared with the algorithm based on fractional Fourier dictionary, the algorithm can obtain higher probability of signal detection success in the presence of narrowband interference signals. In order to estimate the parameters of LFM signals seriously affected by Gaussian white noise and strong narrowband interference signals, a two-search strategy based on fractional Fourier transform dictionary is proposed. The optimal fractional Fourier transform order of the LFM signal is obtained to estimate the frequency modulation slope and the initial frequency of the LFM signal. The experimental results show that the proposed algorithm can obtain a higher probability of success in signal parameter estimation with fewer compression sampling points and lower SNR than the algorithm based on the waveform matching dictionary. This paper presents an improved LFM signal parameter estimation algorithm based on two-search strategy of order Fourier transform dictionary. The algorithm takes advantage of the fact that LFM signal is sparse in time-frequency domain and fractional Fourier transform domain. First, the order of transformation is in time-frequency domain. The optimal fractional Fourier transform order of LFM signal is obtained by coarse search in frequency domain, and then the estimation of FM slope and initial frequency of LFM signal is obtained. Compared with the algorithm based on waveform matching dictionary, the algorithm can also obtain higher probability of signal parameter estimation success under the same SNR and compressed sampling points. 5. An instantaneous frequency measurement algorithm of high frequency microwave signal based on microwave photonics is studied. An instantaneous frequency measurement algorithm for high frequency microwave signals based on optical sampling is proposed, which uses optical intensity modulator to modulate high frequency microwave signals to low repetition rate sampled optical pulses to achieve optical sampling of high frequency microwave signals. Combining FFT with Chirp-z Transform (CZT), the frequency remainder generated under the condition of under-sampling is estimated accurately, and then the signal frequency is reconstructed by using the Chinese remainder theorem. The experimental results show that the algorithm can accurately measure the signal frequency in 39 GHz bandwidth.
【学位授予单位】：西安电子科技大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TN957.51

【相似文献】