【作者】 肖慧；

【导师】 郁文贤； 胡卫东；

【作者基本信息】 国防科学技术大学 ， 信息与通信工程， 2008， 博士

【摘要】 与传统应用背景不同,本文主要研究了LFMCW(linear frequency-modulatedcontinuous wave)雷达中具有大动态范围参数目标的检测和估计问题。所谓大动态范围,主要是指目标具有距离远、高速/加速运动的特点。在所建立的信号检测模型基础上,依次解决了高速运动目标跨距离单元走动校正、高速加速运动目标的参数估计、测距范围的扩展等问题,并利用模糊函数理论分析了扩展测距范围方法的特点。在传统差拍-傅立叶频谱分析-MTD(moving target detection)结构基础上,第二章中首先分析了MTD处理的信噪比改善因子,并据此定义了最优检测度量。针对高速/加速运动目标的特点,定量分析了高速运动和加速运动对二维积累性能的影响程度,并在差拍-傅立叶频谱分析-MTD结构基础上建立了跨距离单元走动情况下的检测模型。针对高速运动目标参数估计问题,第三章主要研究了跨距离单元走动的校正方法。根据差拍信号的特点,提出了一种基于速度预补偿的MTD处理方法,通过在速度动态范围内进行盲搜索的方式对差拍信号进行运动补偿;利用对称三角LFMCW信号的调频对称性,得出距离走动项对上、下调频差拍信号形成对称影响的结论,提出了基于二次混频的MTD处理方法;结合速度盲搜索补偿和二次混频方法各自的特点,提出一种多目标情况下的联合处理方法;进一步将SAR/ISAR成像处理中的Keystone变换引入到LFMCW雷达中,提出了基于Keystone变换的MTD处理方法,并拓展性地将其应用于单调频率LFMCW信号的距离-速度解耦合问题;最后对上述三种校正方法的适用性进行了分析讨论。基于第二章中给出的二维耦合差拍信号模型,第四章研究了高速加速运动目标的参数估计问题。首先推导了最大似然参数估计模型,并得到了高斯白噪声环境下参数估计的Cramer-Rao下界;根据目标二维谱的局部展宽特性,提出了基于局部补偿的速度-加速度模板匹配方法,有效降低了参数估计过程中的运算量,利用局部补偿思想实现最大似然估计,在满足一定输入信噪比要求的条件下能达到接近于CRB下界的参数估计性能;然后从二维解耦合的思想出发,分别提出了二次混频-DPT(discrete polynomial-phase transform)处理方法和直接相位差分方法,在无需进行模板匹配搜索的条件下有效解决了参数估计问题;最后在直接相位差分法的基础上,提出了DPT-CZT(Chirp-Z transform)处理方法,该方法无需借助其它速度解模糊处理即可获得无模糊的速度估计结果,从而可以实现单调频率LFMCW信号对高速加速运动目标的参数估计。针对现有差拍-傅立叶处理结构所存在的测距范围受限问题,第五章研究了扩展LFMCW信号测距范围的方法及相关问题。利用对称三角LFMCW信号的调频对称性和周期重复性的特点,提出了双差拍-傅立叶处理方法,通过两次具有互补特性的差拍处理有效地将时延参数的可测范围扩展到了整个信号重复周期;然后研究了双差拍处理的多周期积累问题,并在此基础上研究了其多目标配对方法,根据距离谱峰值频率特性提出一种距离-速度联合配对法,有效解决了双差拍处理中的多目标配对问题;最后借鉴第四章中的局部补偿模板匹配思想,解决了双差拍处理中高速加速运动目标参数估计问题。论文第六章推导给出了双差拍处理下的模糊函数,详细分析了其对称特性和分辨率特点,并与传统差拍处理下的模糊函数特性进行了比较。从模糊函数的角度得到了双差拍处理的一些定量分析结果,为其工程实用奠定了理论基础。更多还原

【Abstract】 Unlike the traditional application fields, this dissertation is devoted to the detection and parameter estimation problem of targets with high dynamic range parameters in LFMCW (linear frequency-modulated continuous wave) radar. The so-called high dynamic range refers to long range, high-speed and acceleration. Upon the signal detection model constructed, rectification of MTRC (migration through range cell) for high-speed targets, parameter estimation of high-speed accelerating targets and extension of detectable range have been studied and been solved. And the properties of the method for detectable range extension are analyzed using the theory of ambiguity function.Based on the traditional structure of beat-Fourier analysis-MTD (moving target detection) processing, SNR (signal-to-noise ratio) improving factor is obtained in Chapter2, and hereby optimal detection metric is defined. Then quantitative analysis of the effect of high-speed and acceleration on the two-dimensional integration performance is carried out for high-speed/accelerating targets. Finally detection model under MTRC is established upon traditional processing flow.Aiming to parameter estimation of high-speed targets, rectification methods for MTRC are mainly focused in Chapter3. From the characteristic of the beat signal, a velocity precompensation method is proposed which accomplishes motion compensation by searching the optimal compensation velocity in its dynamic range. For the triangular LFMCW signal, MTRC of up-sweep beat signal and down-sweep part is symmetry, based on which secondary mixing-MTD processing is proposed. And recurring to the respective advantage of the above two methods, a combination processing for multi-targets is presented. Furthermore Keystone transform in SAR (synthetic aperture radar) and ISAR (inverse SAR) is introduced into LFMCW radar and a MTD processing based on Keystone transform is propounded which is developed to solve the range-velocity coupling problem in single chirp LFMCW signals. At the end of Chapter3, the applicability of these three rectification methods is discussed.Using the two-dimensional coupling signal model acquired in Chapter2, parameter estimation for high-speed accelerating targets is stressed in Chapter4. Firstly the maximum likelihood (ML) estimation model is derived and the Cramer Rao bound (CRB) in Gaussian white noise circumstance is obtained. Considering that the widening of the two-dimensional spectrum appeared in a local spectrum region, a velocity-acceleration template matching method based on partial compensation is proposed. The concept of partial compensation can effectively alleviate the computation burden during template matching course. And when some input SNR requirement is satisfied, the performance of ML estimation based on partial compensation can approach to the CRB. Then with the intention of decoupling, secondary mixing-DPT (discrete polynomial-phase transform) processing and direct phase differentiation (DPD) method are proposed respectively, which can realize the parameter estimation without searching and template matching operations. And when some modification is made in DPD method, DPT-CZT (Chirp-Z transform) method is produced which can achieve unambiguous velocity estimation without other ambiguity resolving processing. Therefore this method can be applied to parameter estimation for high-speed accelerating targets with single chirp LFMCW signal.To solve the ranging limitation problem in the beat-Fourier analysis-MTD structure, the objective of Chapter5 is to explore some technique to extend the detectable range of LFMCW radar and to investigate the corresponding problems. Considering the modulation symmetry and waveform repetition of triangular LFMCW signal, double beat-FFT processing is proposed in which the beat-FFT processing is performed twice. And by the analysis of the two complementary results, the measurable time-delay of targets is effectively extended to the whole repetition period. Then multi-cycle integration is discussed, and joint range-velocity pairing method for multi-targets is proposed according to the characteristics of range spectrum peaks. In virtue of the partial compensation concept in Chapter4, the parameter estimation problem of high-speed accelerating targets is solved in double beat-FFT processing.At last, the ambiguity function of triangular LFMCW signal adopting the double beat-FFT processing is derived in Chapter6. Under rational approximations, the symmetry and resolution properties are analyzed in detail and compared to those of the traditional processing. Some quantitative results for the double beat-FFT processing are obtained from the point of ambiguity function, which will provide theory foundation for the engineering applications.更多还原