正弦调频傅里叶变换方法及雷达目标微动特性反演技术研究

发布时间：2018-05-08 13:54

本文选题：微动特性 + 正弦调频信号　；参考：《国防科学技术大学》2014年博士论文

【摘要】：微动是目标或目标部件除了质心平动之外的一种具有小幅、往复特点的运动分量,反映了目标的精细特征,在雷达目标识别和成像技术领域占有独特的重要作用。当前,微动参数估计面临一些技术瓶颈,传统时频分析方法估计精度低,基于模型的方法对不同微动形式的适应性差,给以上述两类方法为主体的微动参数估计理论的实用性带来了严峻挑战。本文围绕雷达目标微动特性反演问题展开,提出了正弦调频傅立叶变换等一系列微动参数估计方法,大幅度提高了微动参数估计精度和微动模型的适用范围,研究成果对目标微动特性反演、雷达成像等问题具有指导意义。第一章为绪论,阐述了论文的研究背景及意义,综述了微动特性反演领域的发展概况和研究现状,并介绍了论文的主要工作。第二章从微动模型出发,将微动分为简谐微动与复杂微动、普通微动与小幅微动,基于多正弦调频信号建立了囊括上述微动形式的微动目标雷达回波模型。研究了时频分析方法与基于模型的方法的局限性,以伪魏格纳分布为例,分析了基于时频分布估计正弦调频信号瞬时频率的偏差和随机误差,从理论上揭示了时频分析方法在正弦调频信号分析中面临的技术瓶颈。最后,指出了时频分析方法在微动参数估计中的适用条件。第三章建立了正弦调频信号空间,提出了正弦调频傅立叶变换。与时频分析采用短时窗做瞬时近似不同,正弦调频傅立叶变换通过微多普勒调制信息的长时间积累,显著提高了微动参数估计精度,并从模型角度,根本上解决了多正弦调频信号参数多、估计算法复杂的难题。研究了正弦调频傅里叶变换估计精度与可估计的微动幅度范围,提出了微多普勒噪声比的概念,分析了正弦调频傅里叶变换中的相位缠绕问题与相位解缠方法,并研究了多分量多正弦调频信号正弦调频傅里叶变换的特点,得到了新方法在估计多正弦调频信号参数时的归一化均方根误差经验公式以及多分量信号变换结果的经验公式。接下来,提出了两种相应的解决具体问题的算法:基于正弦调频傅立叶变换的瞬时频率估计方法以及车辆振动谱估计方法。理论分析和仿真实验结果表明,正弦调频傅立叶变换能够精确估计调制微弱的多正弦调频信号参数,较现有方法大幅提高了估计精度和抗噪性能,降低了可估计的微动幅度下限,能够估计未知微动形式、多微动频率成分的目标微动特性。第四章提出了正弦调频稀疏恢复算法,在正弦调频信号空间建立傅立叶调频字典,并引入稀疏恢复方法,借用正弦调频信号空间分析正弦调频信号的优势,以及稀疏贝叶斯学习对非均匀欠采样的适应能力,能够分析低数据率、非均匀采样的正弦调频信号,实现微动参数估计。较现有基于稀疏恢复的正弦调频信号参数估计方法,正弦调频稀疏恢复算法将参数维数降至1维,具有更高的稳健性、估计精度和算法效率。分析了正弦调频稀疏恢复算法的微动谱模糊问题,分析了傅立叶调频字典的相干性和网格划分问题。研究了远程低频段雷达中段目标进动参数估计问题,针对远程低频段雷达估计中段目标进动参数受限于数据率低、采样非均匀、目标微多普勒效应弱等难题,基于正弦调频空间稀疏恢复算法,采用美国铺路爪雷达参数仿真,通过约5分钟的跟踪数据积累,可实现进动频率的精确估计。第五章提出了可实现大转角下不同散射中心类型的清晰成像的微动目标非理想散射中心成像方法。将正弦调频傅立叶变换与HRRP序列结合,先提取不同散射中心的微动信息,实现其位置的粗估计。建立了散射中心复合模型,进而,通过散射中心类型判别、基于微动谱的滑动型散射中心参数估计等一系列处理方法,实现散射中心位置的精估计,并反演相应的目标结构特征。目标非理想散射中心成像方法能够反映目标不同类型散射中心的空间分布,给出目标主要结构的尺寸角度信息,相比传统成像方法,更加直观、全面。第六章总结全文,并指出需要进一步研究的问题。本文从信号模型与信号处理方法角度在传统微动特性反演技术方面取得了新的突破,在正弦调频信号空间中提出了一系列微动目标参数估计、特性反演、非线性调频信号分析的理论与方法,实现了目标微动特性的精细反演,拓展了微动幅度的可估计范围。取得的成果对于微动参数估计、雷达成像、目标识别和电子对抗具有一定的参考价值。
[Abstract]:As a target or target component, a motion component with small amplitude and reciprocating features, in addition to the translational motion of the center of mass, reflects the fine characteristics of the target. It plays a unique and important role in the field of radar target recognition and imaging technology. At present, the estimation of the microdynamic parameters is faced with some technical bottlenecks, and the traditional time frequency analysis method has a low estimation accuracy. The model method has brought a severe challenge to the practicability of the theory of microdynamic parameter estimation with the two methods as the main body, which is poor in the adaptability of different microdynamic forms. In this paper, a series of parameter estimation methods such as sinusoidal FM Fu Liye transform, such as sinusoidal FM transformation, are proposed in this paper. The accuracy of the estimation of dynamic parameters and the scope of the application of the microdynamic model, the research results have the guiding significance for the inversion of the target fretting characteristics and the radar imaging. The first chapter is the introduction, expounds the research background and significance of the paper, summarizes the development and research status of the field of microdynamic inversion, and introduces the main work of the paper. Second chapters From the microdynamic model, the micromotion is divided into simple harmonic fretting and complex micromovement, ordinary micromotion and small amplitude micromotion. Based on the multi sinusoidal FM signal, the microdynamic target radar echo model including the above micromotion is established. The limitation of the time frequency analysis method and the model based method is studied. The pseudo Wegener distribution is taken as an example to analyze the time based time. The frequency distribution is used to estimate the deviation and random error of the instantaneous frequency of sinusoidal frequency modulation signal. In theory, it reveals the technical bottleneck of the time-frequency analysis method in the analysis of sinusoidal frequency modulation signal. Finally, it points out the application conditions of the time frequency analysis method in the estimation of the microdynamic parameters. The third chapter establishes the sinusoidal frequency modulation signal space and puts forward the sinusoidal frequency modulation. Fu Liye transform is different from the instantaneous approximation of time frequency analysis using short time window. Sinusoidal FM Fu Liye transform can significantly improve the precision of the estimation of microdynamic parameters by long time accumulation of micro Doppler modulation information. From the model angle, the multi sinusoidal frequency modulation signal is more complex and the algorithm is complex. The concept of micro Doppler noise ratio is proposed and the phase entanglement and phase unwrapping method in the sinusoidal FM Fu Liye transform is analyzed. The characteristics of the sinusoidal FM Fu Liye transform of multicomponent multi sinusoidal FM signal are studied, and the new method is obtained in the estimation of the multiple Fu Liye transform. The normalized mean square root error empirical formula and the empirical formula for the result of multicomponent signal transformation are given in the parameters of chirp signal. Then, two corresponding algorithms are proposed: the instantaneous frequency estimation method based on sinusoidal FM Fu Liye transform and the method of vehicle vibration spectrum estimation. Theoretical analysis and simulation experiment junction The results show that the sinusoidal FM Fu Liye transform can accurately estimate the parameters of the weak modulated multi sinusoidal frequency modulation signal. Compared with the existing methods, the estimation precision and the anti noise performance are greatly improved. The lower limit of the estimated fretting amplitude is reduced and the target fretting characteristics of the unknown fretting form and the multi fretting frequency are estimated. The fourth chapter puts forward the sinusoidal modulation. The frequency sparse recovery algorithm is used to establish the Fu Liye frequency modulation dictionary in the sinusoidal frequency modulation signal space, and the sparse recovery method is introduced, the sinusoidal FM signal is used to analyze the advantages of the sinusoidal frequency modulation signal and the sparse Bayesian learning is adapted to the non-uniform undersampling, and the sine frequency modulation signal with low data rate and non-uniform sampling can be analyzed. Compared with the existing parameter estimation method of the sinusoidal frequency modulation signal based on sparse recovery, the sinusoidal FM sparse recovery algorithm reduces the parameter dimension to 1 dimension, has higher robustness, the estimation precision and the algorithm efficiency. This paper analyzes the micro motion spectrum fuzzy problem of the sinusoidal FM sparse recovery algorithm, and analyzes the phase of the Fu Liye frequency modulation dictionary. The parameter estimation of middle target precession in long range low frequency radar is studied. Aiming at the problem of low rate of data, inhomogeneous sampling and weak Doppler effect in the middle segment of the long range low frequency radar, the US paving claw radar is used on the basis of the sine FM space sparse recovery algorithm. Parameter simulation, through the accumulation of about 5 minutes of tracking data, can achieve accurate estimation of the precession frequency. The fifth chapter proposes a non ideal scattering center imaging method which can realize clear imaging of different scattering center types at large rotation angle. The sinusoidal FM Fu Liye transform is combined with the HRRP sequence to extract the different scattering centers. The scattering center complex model is established by moving information. Then a series of processing methods, such as the type discrimination of the scattering center type and the parameter estimation of the sliding scattering center based on the microdynamic spectrum, can be used to realize the precise estimation of the position of the scattering center and inversion of the corresponding target structure characteristics. The imaging method of the target non ideal scattering center is used. It can reflect the spatial distribution of different types of scattering centers of the target, and give the dimension and angle information of the main structure of the target. Compared with the traditional imaging method, it is more intuitive and comprehensive. The sixth chapter summarizes the full text, and points out the problems that need further study. This paper from the angle of signal model and signal processing method in the traditional microdynamic characteristics inversion technology A new breakthrough has been made. In the sinusoidal frequency modulation signal space, a series of theory and methods of parameter estimation, characteristic inversion and Nonlinear FM signal analysis are proposed. The fine inversion of the microdynamic characteristics of the target is realized and the estimated range of the microdynamic amplitude is extended. The results obtained are for the estimation of the microdynamic parameters, the radar imaging, and the target recognition. Do not have a certain reference value with electronic countermeasures.

【学位授予单位】：国防科学技术大学
【学位级别】：博士
【学位授予年份】：2014
【分类号】：TN957.51

【参考文献】