论文部分内容阅读
MTD 系统原理上由抑制杂波的白化滤波器与用 FFT 实现的相干积累级联组成。当信号长度为 P 时,为得到有效的相干积累,应对整个信号一次处理。但考虑到相干积累取幅后,通常还要经单元平均恒虚警率(CFAR)电路进行检测,作一次处理常会带来大的恒虚警损失,特别是在参考单元不太多时,更为严重。为此,提出了将 P 长度的信号分成几段,分别经相干积累取幅后,再作非相干积累,然后进行单元平均恒虚警率检测。这样做,显然会带来非相干积累损失,但可以减少恒虚警损失。如果后者的得益大于前者的损失,则分段处理会带来好处。本文对此进行了详细分析,并求得在一定条件下的最佳分段数。
The principle of MTD system is composed of a whitening filter for suppressing clutter and a coherent accumulation concatenation realized by FFT. When the signal length is P, in order to obtain effective coherent accumulation, the entire signal should be processed once. However, taking into account the coherent accumulation of framing, usually through the unitary average constant false alarm rate (CFAR) circuit to detect, for a treatment often bring a large constant false alarm loss, especially in the reference unit is not too much, more serious. Therefore, it is proposed to divide the signal of P length into several segments, accumulate them by coherent accumulation, then accumulate non-coherently, and then detect the unit average CFAR. In doing so, it is clear that non-coherent accumulation losses will result, but the loss of constant false alarm can be reduced. If the latter’s benefit is greater than the former’s loss, then the sub-processing benefits. In this paper, a detailed analysis of this and get the best number of segments under certain conditions.