【作者】 曹杨；

【导师】 水鹏朗； 冯大政；

【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】 近些年来，伴随着阵列天线在机载预警平台的使用，杂波在运动平台呈现为强烈的空时耦合特性。为了有效抑制运动平台杂波，空时自适应处理方法（Space-TimeAdaptive Processing，STAP）得到空前发展。但是，随着阵列规模越来越大、信号维数越来越高，由此便产生了计算复杂度过高和训练样本需求较多等问题。由于以上原因，促进了各种降维自适应方法（Suboptimal Dimension-Reduced STAPAlgorithms）的发展。但是，机载雷达非自适应杂波抑制方法（Non-Adaptive ClutterSuppression Algorithms）的相关研究相对而言还比较少。非自适应方法因其计算复杂度偏低、受样本影响小等特点，更利于实时处理数据。因此，本文主要研究机载相控阵平台和机载MIMO（Multiple-Input-Multiple-Output）平台的非自适应杂波抑制方法，主要工作包括以下几个方面：1.建立了机载MIMO雷达杂波空时二维数据模型，在充分利用雷达工作参数和载机速度等先验信息的基础上，提出一种机载MIMO雷达非自适应空时二维脉冲相消器（Multiple-Input-Multiple-Output Two-Dimensional Pulse-to-pulse Canceller，MIMO TDPC），并给出了关于MIMO TDPC权系数的最小二乘代价函数，从而优化得到MIMO TDPC的权系数。由于MIMO TDPC权系数仅利用了雷达工作参数和载机速度等先验信息计算得到，属于非自适应处理器，因而具有运算量小、无收敛过程等优点，并且可以作为机载MIMO雷达的杂波预滤波器，进一步改善常规MTI（Moving Target Indication）处理和降维STAP算法的性能。此外，在杂波模型中考虑了偏航角，因此MIMO TDPC不仅适用于正侧视雷达，也适用于非正侧视雷达。2.针对机载非正侧视雷达的近程非均匀杂波抑制问题，提出一种快速实现、便于使用的近程杂波抑制方法（Short-range Clutter SuppressionApproach，SCSA）。对于正侧视雷达（SideLookingAirborne Radar, SLAR）,由于不同距离单元的杂波在角度-多普勒平面分布轨迹基本重合，杂波表现为相对均匀，因此空时自适应处理方法有很好的性能。但是，对于非正侧视雷达（non-SLAR），由于杂波的分布轨迹不重合，特别是近场杂波，杂波的非均匀性更强烈，自适应方法的性能会严重下降。在机载非正侧视雷达的杂波模型基础上，利用杂波空间几何结构、雷达参数和载机速度等先验信息，构造了SCSA权系数。SCSA方法由空-时滤波和空-时匹配两部分级联构成。作为不依赖于样本的非自适应方法，与传统自适应方法不同（自适应方法性能受非均匀样本的影响很大，且求逆运算计算复杂度高），本方法不受非均匀样本的影响，计算复杂度很低，方便实时处理使用，且较自适应方法有更好的目标检测性能。3.考虑由于载机速度变化引起杂波起伏，并由此导致空-时二维平面杂波沿分布轨迹在多普勒域扩散的问题，在机载雷达地杂波模型的基础上，利用雷达参数、载机速度等先验信息，提出了一种机载雷达二维多脉冲相消器（Two-DimensionalMulti-pulse Canceller, TDMC），并给出了关于TDMC权系数的代价函数。由于使用了比两脉冲相消器TDPC更多地自由度来设计滤波器，因而TDMC比TDPC有更好的滤波器通频带性能，通过仿真数据和实测数据验证，TDMC比TDPC能更有效地抑制杂波，并且对慢速目标的检测性能有明显提升。更多还原

【Abstract】 In recent years, the phased-array is widely applied in the airborne radar. In order tosuppress the spatially-temporally coupled clutter which is induced by the movement ofthe airplane, the well-known space-time adaptive processing (STAP) algorithms arevigorously researched. As the number of the array antennas and the dimension of thesignals become larger and larger, the problems of the computational complexity and thetraining samples required become more and more serious. Thus, the suboptimaldimension-reduced STAP algorithms have been rapidly developed, too. However,compared with the above adaptive algorithms, the research of non-adaptive algorithmsis relatively insufficient. Since the non-adaptive algorithms have low computationalcomplexity and less demand for the training samples, they are more suitable for thereal-time processing. Hence, this paper will mainly focus on the non-adaptive cluttersuppression approaches in both the airborne phased-array radar and the airbornemultiple-input-multiple-output (MIMO) radar. The main contributions of the thesis aresummarized as follows:1. The spatial-temporal model of the clutter data for the airborne MIMO radar isestablished. Taking full advantage of the prior information such as the radar parametersand the platform velocity, we proposed a multiple-input-multiple-outputtwo-dimensional pulse-to-pulse canceller (MIMO TDPC). A least-square cost functionassociated with the coefficient matrix of the MIMO TDPC is organized, and thecoefficient matrix is obtained by solving the optimization problem. Since the MIMOTDPC coefficient matrix can be calculated by using the prior information, our method isa non-adaptive method which owns low computational complexity andnon-convergence process. As an efficient and convenient ground clutter pre-filteringtool before the conventional moving target indication (MTI) method and thewell-known suboptimal dimension-reduced STAP algorithms, MIMO TDPC caneffectively enhance the target detection performance. Furthermore, the drift angle isadopted in the design of our method. Thus, MIMO TDPC can be utilized in bothsidelooking radar and non-sidelooking radar.2. In order to suppress the short-range heterogeneous clutter for the non-sidelookingairborne radar (non-SLAR), we propose a short-range clutter suppression approach(SCSA). Space-time adaptive processing (STAP) methods which have been developed for suppressing the spatially-temporally coupled ground clutter in the airborne radarhave achieved good performance when applied to the sidelooking airborne radar (SLAR)where the clutter is relatively stationary. However, due to the range dependence (or thegeometry-induced heterogeneous clutter) the performance is degraded in the non-SLARespecially for the short-range clutter suppression where the range dependence is moresevere. Thus, the SCSA is established based on the geometry knowledge of the clutter,the radar parameters and the platform velocity for suppressing the short-range clutter inthe non-SLAR. The SCSA is composed of two parts the spatial-temporal pre-filteringand the spatial-temporal matching. Since the target detection performance of theclassical adaptive algorithms will be degraded by the range-dependent secondarysamples, the SCSA which is a non-adaptive algorithm can gain a relatively goodperformance. Moreover, compared with the adaptive algorithms the computationalburden of which is high due to the covariance matrix inverse operation, the SCSA whichcan be pre-calculated and made into a look-up table is more suitable for the real-timeprocessing.3. The variation of the airborne radar platform velocity will cause the clutterfluctuation, and thus lead to the clutter power spectrum widening along the cluttertrajectories in the angle-Doppler domain. On the basis of the clutter model, a noveltwo-dimensional multi-pulse canceller (TDMC) which employs more pulses of theclutter echo is established. Compared with the existing two-dimensional pulse-to-pulsecanceller (TDPC) which adopts only two pulses of the clutter echo, the TDMC utilizesmore degrees of freedom (DOFs) to organize a spatial-temporal two-dimensionalband-rejection filter to suppress the ground clutter more efficiently. Experiments of bothsimulated data and measured data show that the proposed TDMC can gain a bettertarget detection performance than the TDPC, especially for the slow-moving targetdetection.更多还原