【作者】 王万林；

【导师】 廖桂生；

【作者基本信息】 西安电子科技大学 ， 信号与信息处理， 2004， 博士

【摘要】 在非均匀环境下，由于缺乏足够的与待检测样本中干扰独立同分布(IID)的训练样本，常规统计空时自适应处理(STAP)方法的性能往往会急剧下降，因此能够适应非均匀环境的新型STAP算法逐渐成为目前STAP技术的研究热点之一。本论文着重研究的就是各种非均匀情况下的相控阵机载雷达的STAP杂波抑制算法及目标检测问题。第二章至第六章主要研究功率非均匀、干扰目标和孤立干扰等非均匀问题的解决方法，而第七章和第八章则主要研究非正侧视阵情况下近程杂波的非均匀现象及解决方法。 1．第二章主要针对辅助通道法降维STAP处理来讨论。首先给出了如何选取辅助通道构造变换矩阵的统一准则，该准则弥补了传统辅助通道法无法用于非正侧视阵列结构的不足，属于自适应结构的降维算法，而传统方法属于固定结构。然后提出了一种抗干扰目标的新辅助通道法，即在有限次样本条件下，传统方法中主通道的目标信号会泄漏到辅助通道引起目标相消现象，新方法由于已将目标滤除不存在这种现象，因而特别适合于非均匀环境下训练样本比较有限的情况。 2．第三章首先提出了先直接数据域(DDD)处理后统计处理的级联方法，该方法有效弥补了DDD算法空时孔径损失较大的缺点。然后提出了用实测导向矢量校正接收数据的误差校正方法，保证了DDD算法只能用于等距阵列这一先决条件。最后还提出了一种新的信号滤除方法，该方法可以用于实际相控阵机载雷达系统。 3．第四章提出了一种新的非均匀环境下两级空时自适应处理的级联算法，即首先根据均匀杂波的分布特性选取辅助波束用辅助通道法滤除均匀杂波，然后用直接数据域算法抑制孤立干扰。本方法克服了Adve等的级联方法中杂波和孤立干扰相互影响的缺点，具有性能优良、鲁棒性好和计算量低等优点。 4．第五章提出了强孤立干扰的抑制新方法，具体给出了两种处理措施。当样本数目较少时用斜投影技术抑制含有孤立干扰样本中的孤立干扰，抑制后的样本继续参与杂波抑制自适应权的训练。当样本数目较多时用干扰置零技术同时抑制含有孤立干扰样本中的孤立干扰和均匀杂波，抑制后的样本直接进行目标检测，不再参与杂波抑制自适应权的训练。上述方法避免了DDD算法逐距离单元矩阵求逆运算，有效降低了计算量。 5．第六章提出了一种基于最大似然估计的动目标检测的非均匀方法，首先利用自回归(AR)方法对雷达接收数据进行时域白化处理，然后利用最大似然方法逐距非均匀环境下的相控阵机载雷达STAP研究 离单元估计目标回波幅度，最后由估计结果与自适应门限比较得到检测结果。 该方法具有主杂波区性能好、孤立干扰抑制能力强、计算量低等优点，比较适 合于功率非均匀情况。6.第七章针对非正侧视相控阵机载雷达，由于杂波在距离上呈现出严重的非均匀 性，尤其是在近程这种依赖距离的杂波谱呈现的非均匀现象更为严重，使得用 于计算自适应权的协方差矩阵不可能直接由距离单元数据统计平均来估计得 到，否则STAP处理及其降维技术难以获得理想的杂波抑制性能和目标检测性 能，甚至恶化。本章从杂波谱频移的角度提出了一种基于空间角频移的距离相 关性补偿方法，即对不同距离单元分多普勒通道施加不同的补偿变换，使各通 道内所有单元的杂波能量在二维平面内完全集中在参考单元的杂波位置上，从 而消除了杂波的非均匀性，减小了目标检测盲区。此外，还提出一种基于正交 投影的预处理技术，目的在于降低各单元中模糊杂波的强度，改善补偿效果。7.第八章则从基于导数更新但B功的角度来研究非正侧视相控阵机载雷达的距离 依赖的杂波谱校正，即:首先分析了杂波谱随距离单元斜距和仰角余弦的变化 规律，然后提出了一种基于仰角余弦导数更新伍DBU)的距离依赖杂波谱校正新 方法。EDBU方法由于同时利用了距离单元斜距、载机高度和地球曲率半径的 信息，因而比DBU方法更能准确描述STAP杂波抑制自适应权矢量的变化规 律，校正性能更好。与DBU方法一样，EDBU方法也可以用于双基地机载雷 达的杂波谱校正。关键词:辅助通道处理，干扰目标，直接数据域，误差补偿，孤立干扰，斜投影，功率非均匀，最大似然估计，距离相关性，基于空间角频移补偿，基于仰角余弦导数更新补偿西安电子科技大学博士论文更多还原

【Abstract】 In the nonhomogeneous environments, the performance of space time adaptive processing (STAP) proposed by Brennan often declines sharply due to the fact that it is lack of independent identically distributed train samples. Thus, the new STAP method that can adapt to the real nonhomogeneous environments is becoming one of the main research directions of the current STAP techniques. This paper put its emphasis on the investigation of the algorithm of clutter suppressing and moving target detecting for phased array airborne radar in nonhomogeneous environment. The solutions to the nonhomogeneity about power fluctuation, interfering target and isolated interference were studied from chapter 2 to 6, the ones about range dependence of non-sidelooking array were investigated from chapter 7 to 8.1. In chapter 2, the research was focused on auxiliary channel processing (ACP) STAP Firstly, the unified rule how to select auxiliary channels was given, which can overcome the drawback the conventional ACP can’t be used in non-sidelooking array, thus is the reduced dimension method with adaptive configuration, however the conventional ACP is that with fixed configuration. Latterly, a new ACP to resist interfering target was proposed, that is, under the condition of samples being limited, the target self-canceling is resulted in the conventional ACP because the target from main channel leak to auxiliary channel, however, the new ACP has not this phenomenon because the target in main channel is also filtered, thus can adapt to the nonhomogeneous environment of the samples being limited.2. In chapter 3, a new STAP method to suppress clutter and interferer in non-homogeneous environment combined direct data domain (DDD) with statistic STAP was proposed firstly, which can effectively compensate the performance loss incurred by the spatial and temporal aperture loss of the DDD approach. An error correction technique by correcting radar receive data with real steer vectors was put forward latterly, which can guarantee the predeterminative condition of DDD algorithm that the array is uniform. A new method of target signal filter was given lastly, which can be used in real phased array airborne radar system.3. In chapter 4, a new method of two stage hybrid space time adaptive based on non-homogeneous environment was presented, in which the homogeneous clutter is firstly suppressed based on auxiliary channel processing (ACP) approach by selecting auxiliary channel according to clutter’s distributing characters in the beamDoppler space and the isolated interferers is latterly mitigated based on direct data domain algorithm. This method can overcome the drawback of the method Adve proposed that is tampering with each other between the clutter and the isolated interferer, has advantages of excellent target detection performance, good robust to whether small yaw exists or not, and low computation load.4. In chapter 5, a new method to suppress strong isolated interferers was put forward, in which two solutions were given. The strong isolated interferers are mitigated firstly by oblique projection, and the processed cells are also used as train samples of adaptive clutter suppression’s weight when the train samples is lacking. The cells with strong isolated interferers are directly detected after the interferers and clutter are both suppressed by using null technique, and the processed cells are no longer used as train samples when the train samples is abundant. Unlike DDD algorithm, the proposed methods needn’t to compute the inversion of the covariance matrix for every cell, and have low computational complexity.5. In chapter 6, a new method to moving target detection in nonhomogeneous environment based on maximum likelihood was presented, in which, the autoregressive model is used to whiten the data radar received in time domain firstly, the maximum likelihood method is used to estimate the target return amplitude in every cell latterly, and the target detection is handled by comparison of the estimated result更多还原