一个时域离散反卷积的后向迭代算法

A Back Substitution Algorithm of Deconvolution in the Time Domain

本文所提出的时域离散反卷积的后向迭代算法是一个直接在时域实现离散反卷积的算法.该方法不仅能避免传统的Fourier变换法当卷积核的谱存在零点时的困难。且当输入(或输出)信号的首元存在测量误差时,仍能给出准确的解,克服了传统Z变换法所存在的误差积累.最后,论论了在输入、输出信号受随机噪声干扰的情况下,用后向迭代法和前向迭代法的平均解类减少估值h(n)的方法。

A back substitution algorithm is presented for deconvolutionne time domain. This algorithm can overcome the difficulties of traditionalmethods,i.e.,Fourier transform method and Z transform method,of deconvolution.One of the objectives of this paper is to obtain the impluse response in the timedomain when arbitrary input and output signals are given.