论文部分内容阅读
本文研究空域和时域的自适应处理结构的基本的相似性和重要的差别。主要想强调这样的事实:从天线阵不同单元接收到的信号之间的去相关,要用一种级联的结构;而一个时间序列的相继样本之间的去相关,则可以采用一种很不相同的,而且更有效的格形滤波器的结构。这个差别可以通过考察Levinson-Durbin递推算法得到解释。格形滤波器的结构,实质上就是由这个算法产生的。把这种算法表示成能考虑非平稳统计性质的影响的一般化形式,并清楚地说明使用时间上的递推算法提高了它的有效性。这种算法显然不适合于天线阵列的去相关,或者一般地说,它不适合于空域的去相关。
This article studies the basic similarities and important differences between adaptive processing structures in airspace and time domains. The main emphasis is on the fact that the decorrelation between the signals received from different elements of an antenna array requires a cascaded structure and that the decorrelation between successive samples of a time series can be very, Not identical, but more efficient lattice filter structure. This difference can be explained by examining the Levinson-Durbin recursion algorithm. The structure of the trellis filter is essentially the result of this algorithm. Represent this algorithm as a generalized form that takes into account the effects of non-stationary statistical properties and clearly demonstrate the effectiveness of using recurrence algorithms over time. This algorithm is obviously not suitable for the decorrelation of antenna arrays or, generally speaking, it is not suitable for spatial decorrelation.