机载自适应阵中的多普勒—特征波束形成

Doppler-Eigen Beamforming in Airborne Adaptive Arrays

在机载相控阵雷达下视情况下，多普勒展宽的时变杂波可以用自适应时空二维处理器得到最优的抑制。然而对于大型阵列的实时应用来说，这种最优时空二维处理却很难实现，这是由于它要求很高的计算量（以典型的采样协方差矩阵求逆算法￣［２］为例，其运算量大约为每次迭代Ｏ｛（ＮＭ）￣２｝，其中Ｎ为阵元数，Ｍ为时间采样数）。在本文中，首先通过杂波子空间分析证明了杂波自由度小于Ｎ＋Ｍ，而且阵列信号在经过多普勒滤波后，单个多普勒滤波器输出的杂波自由度很小，从而可以用一个仅有Ｎ＋Ｍ个自由度的低阶（部分自适应）处理器来实现准最优的杂波抑制；随后，提出了用Ｎ＋Ｍ个多普勒─特征波束（多普勒滤波器后跟空间特征波束）来近似杂波子空间的办法，并据此方法构造了一个使用这种多普勒一特征波束的广义旁瓣对消器并用它实现了准最优的杂波抑制，所需计算量仅为每次迭代Ｏ｛Ｎ＋Ｍ）Ｎ＋（Ｎ＋Ｍ）￣２｝。由于空间特征波束所具有的误差补偿能力，该处理器在实际应用中对于处理器自由度不足是不敏感的。

The Doppler broadened and time varing clutter can be optimumly suppressed by an adaptive spaee-time processor in an airborne phased array PD radar,but real-time application is difficult for large arrays because of its very high computatonal cost(about O{(NM)￣2}per iteration step,where N is the number of sensors and M is the number of pulses).By clutter subspace analysis the degree of freedom of clutter is proved to be less than N+M and the degree of freedom ofclutter in the output of a single doppler filter is very small,so that optimum clutter suppression can be per-formed by a low-order partial-adaptive processor with degree of freedom of only N+M.The clutter subspace can be well approximated by some doppler-eigen beams,which are doppler filers followed by spatial eigen beams,and a generalized sidelobe canceller using these beams efficiently realizes near-optimum clutter suppression at a computational cost of only O{(N+M)N+(N+M)￣2}periteration step.As a result of the error compensation ability of the spatial eigen beams,in realistic situations it is not sensitive to the lack of degree of freedom of the processor.