引用本文:何一帆, 贾沛璋.非高斯线性过程三重谱估计及其在反褶积方法中的应用[J].控制理论与应用,1989,6(S2):26~34.[点击复制]
He Yifan, Jia Peizhang.The Trispectral Estimation of Nongaussian Linear Process and Its Application in Deconvolution[J].Control Theory and Technology,1989,6(S2):26~34.[点击复制]
非高斯线性过程三重谱估计及其在反褶积方法中的应用
The Trispectral Estimation of Nongaussian Linear Process and Its Application in Deconvolution
摘要点击 852  全文点击 385  投稿时间:1986-08-01  修订日期:1987-11-17
查看全文  查看/发表评论  下载PDF阅读器
DOI编号  
  1989,6(S2):26-34
中文关键词  
英文关键词  
基金项目  
作者单位
何一帆, 贾沛璋 中国科学院系统科学研究所 
中文摘要
      近来的一些文献表明,地震数据处理中的反射系数近似为零对称分布,反射系统的3阶累积矩等于零。对此情形,应用三重谱技术作高阶谱反褶积效果更好。本文研究具有非高斯白噪声激励的有限阶滤波器输出的三重谱估计,利用三重谱的有限和表示形式构造了一种递推估计算法,该算法的计算量与滤波器阶的一次项同量级。我们还证明了用该算法所得估计值的渐近无偏性和一致性。将此算法应用于非最小相位反卷积仿真,几个仿真实例的滤波器参数和反射系数估计值的偏差都明显低于[1]中的估计结果。
英文摘要
      Some recent literatures show that the reflective series are almost symmetric in signal processing and its third cumulant equal to zero. In such case an approach which is based on the trispectrum suggests itself. The purpose of this paper is to develop the techniques to estimate the trispectrums of output of the finite impulse response filters with Non-Gaussian white noise input, we have constructed an efficient recursive algorithm which is based on the finite summation form of the trispectrum for MA (q), the computational cost largely is decreased. Both asymptotic unbiaseness and efficiency of the algorithm have also been proved. Specifically, by using the algorithm to some deconvolution problems which were treated in [1], the simulations have shown a great deal of improvement compare to [1].