【作者】 张各各；

【导师】 王俊；

【作者基本信息】 西安电子科技大学 ， 信号与信息处理， 2014， 博士

【摘要】 外辐射源雷达又称为被动相干定位(PCL)雷达，属于新体制雷达中的一种。与传统雷达不同，外辐射源雷达自身不主动发射电磁波，而是利用各种通信、广播信号（比如调频广播、电视台、GSM基站和GPS卫星等）作为照射源，被动接收目标的反射回波信号实现目标探测和成像。从结构上来说，外辐射源雷达属于双基地雷达配置，使其不易被敌方发现，具有较好的抗干扰和抗反辐射导弹能力；并且借助于超高频和甚高频的照射源特性，外辐射源雷达具有良好的反隐身和低空探测能力；同时具有便携、低成本、无需频谱分配、无电磁污染、无探测盲区等优点。因此，为了应对复杂战场环境下的目标探测问题，外辐射源雷达目标探测已成为一种有效的解决手段，具有重要的研究价值。本论文主要针对简易配置的外辐射源雷达目标探测系统中的若干具体问题，如直达波提纯、杂波相消、参数估计以及数据融合等问题进行深入研究，所取得的主要研究成果如下：1、针对复杂战场环境下，接收机周围可能存在多个照射源，照射源基本信息未知（主要是基站方向和信号波形信息），无法区分感兴趣信号和干扰信号，无法确定目标探测中各照射源和目标参数的配对关系等问题，提出利用信号特性来抑制干扰。考虑到大部分通信信号都满足循环平稳的特性，提出利用信号的循环平稳特性估计直达波的方向信息，并抑制其它基站干扰。在目标探测阶段，推导二阶循环平稳和模糊函数的等价性，提出利用二阶共轭循环平稳代替模糊函数，实现对感兴趣信号和目标参数信息的配对。2、针对实际工程中参考信号包含多径，常规外辐射源雷达参考通道和预警通道都需要杂波相消的问题，提出在预警通道中利用盲源分离方法直接分离直达波和目标回波信号，达到减少运算量和提高目标回波信号信噪比的目的。针对常规盲源分离不适用于外辐射源雷达中病态矩阵的问题，提出利用循环平稳特性来抑制干扰，达到增强弱目标信号的目的，在天线个数少于信号源个数的欠定情况，本方法同样适用。在信号分离阶段，针对实际混合信号中还混有噪声的问题，提出鲁棒的斜投影方法来抽取目标回波信号。3、考虑到OFDM信号在同步和抗干扰方面的优点，提出基于OFDM信号特性的外辐射源雷达目标探测。针对外辐射源雷达中参考信号信息可能无法直接获取的问题，提出利用信号特性来估计OFDM信号信息。具体考虑到OFDM模糊函数中的副峰是由导频的周期性造成的，并且模糊函数等价于二阶循环平稳，提出利用二阶循环平稳估计导频的周期性，从而能够确定导频的位置信息，继而实现杂波信道脉冲响应估计和直达波重构。4、针对目前通信广播系统大都采用OFDM调制信号，并且支持SFN结构的情况，提出基于OFDM-SFN的外辐射源雷达系统提高目标探测的可靠性。但在SFN中，多个信号从不同的方向照射到目标，极有可能造成大时延扩展。并且OFDM发射正交子载波，对载频频偏很敏感。针对这两个问题，在没有参考信号信息的情况下，提出在接收通道中利用阵列天线把大时延转化为小时延，建立新的张量接收模型，并利用联合对角化方法同时估计波束、时延和频偏三个矩阵变量，解决频域盲源分离算法在大时延扩展中估计精度不高的问题。5、针对OFDM-SFN中存在的大时延，提出另一种基于循环移位和空时编码的收发方式，接收端使用单天线就能对抗大时延，达到简化接收端系统配置的目的。并针对SFN多站定位中，常规定位算法存在两步估计误差的问题，提出基于OFDM信号特性的直接目标定位方法。相比常规的定位方法，本方法能够获得更高的定位精度，尤其是在目标回波信号低信噪比情况下。更多还原

【Abstract】 Passive bistatic radar, also known as passive coherent location (PCL) radar,belongs to a new radar system. Passive bistatic radar, different from the conventionalradar, doesn’t emit electromagnetic waves, but to use a variety of communications,broadcast signals (such as FM radio, television stations, GSM base stations and GPSsatellites, etc.) as illuminators, to passively receive the reflected echo signals fromtarget for target detection and imaging. In structure, passive bistatic radar belongs to thebistatic radar configurations, it can’t easily be found by the enemy, and thus it has betteranti-jamming and anti-radiation missiles capabilities; And by means of UHF and VHFsignal characteristics, passive bistatic radar also has good anti-stealth and low-altitudedetection capabilities; Except that, it also has the advantages of portable, low-cost, nospectrum allocation, no electromagnetic pollution and no blind spot detection etc.Therefore, in order to cope with the complexity of modern battlefield environment,passive bistatic radar has become an effective way of target detection and has importantresearch significance.This thesis focuses on several specific issues in the passive bistatic radar targetdetection system, such as the direct wave reconstruction, clutter cancellation, parameterestimation, data fusion and so on; the main research results obtained are as follows:1. For a complex battlefield environment, there may be multiple illuminationsources around the receiver, the basic information of illuminator is unknown (mainlythe direction of base station and the signal waveform), can not distinguish between theinterested signals and interference signals, and can not determine the mathing betweenthe illuminator and target parameter in the target detection. Considering that most of thecommunication signal is cyclostationary, we propose a blind method to suppressinterference. Cyclostationary information of signal is made to estimate the direct signaldirection while suppressing interference from other base stations, which finally obtainrelatively accurate direct signal. In the target detection stage, we derive equivalencebetween ambiguity function and second-order cyclostationarity, and propose secondorder conjugate cyclostationary instead of ambiguity functions to achieve matchingbetween interested base station and target parameter.2. Considering reference signal containing the multipath, and then both thereference channel and surveillance channel require clutter cancellation in the actualprojects, we propose the blind source separation to extract the direct signal and the target echo signal directly from the sole surveillance channel, which achieve the goal ofdecreasing the computational complexity and enhancing the weak echo signal. For theconventional blind source separation not suitable for the ill-posed problem in thepassive bistatic radar, we propose to enhance the weak target by using cyclostationarity,and apply to the underdetermined case in which the number of antennas is less than thenumber of signal sources. In the signal separation stage, a robust oblique projectionmethod is given to extract the target echo signal from the mixed signals containingnoise.3. Taking into account the advantages of OFDM signal in synchronization andanti-jamming, we propose passive radar target detection method based on OFDM signalcharacteristics. For the problem of reference signal information in passive radar may notbe directly acquired, we propose a blind method to estimate the OFDM parameters.Taking into account the side peaks in the ambiguity function of OFDM is caused by theperiodicity of pilot signals, but also because of the equivalence between ambiguityfunction and the second-order cyclostationarity, we propose to estimate the periodicityof pilot signals, which obtain the pilot position information, then to achieve the clutterchannel impulse response estimation and reconstruct the direct signal.4. For the most current communication and broadcasting systems using OFDMmodulated signal, and supporting single frequency network structure, we adviseOFDM-SFN passive bistatic radar to improve target detection reliability. But in the SFN,the multiple signals from different directions to detect target, it is most likely to causethe large delay spread. Furthermore, transmitting the orthogonal subcarriers, OFDM isvery sensitive to the carrier frequency offset. For these two problems, in the case ofhaving no illuminator information, we propose to covert the large time delay into asmall delay by using the array antenna, model a new tensor receivers, and simultaneousestimate delay and Doppler frequency offset by using joint diagonalization method,which solve the poor estimation accuracy of frequency-domain blind source separationalgorithm in the large delay spread.5. For the large delay in OFDM-SFN, suggests another way to send and receivebased on cyclic shift and space-time coding, the receiver will be able to use a singleantenna to combat the large delay spread, which achieve the goal of simplifying thesystem configuration at the receiver. And for two-step positioning error in theconventional location algorithm for the SFN multi-station positioning, we adopt to usethe direct target location based on OFDM signal characteristics. Compared with the conventional positioning method, the proposed can achieve higher estimation accuracy,especially in the low SNR.更多还原