自校正最优去卷平滑器

A Self-Tuning Optimal Deconvolution Smoother

本文从时间序列分析的新观点阐述自适应信号去卷问题.对于通过已知多变量线性系统被观测的未知的多变量平稳ARMA信号,基于ARMA新息模型,应用时间序列分析方法和射影响理论,本文提出了一种新的最优去卷平滑器和相应的自校正去卷平滑器,它不同于基于状态空间模型的Goodwin和Sin的结果.本文结果包括了文献中的许多问题和结果作为特例,且可应用于地震信号去卷,信号处理,通讯系统,语音识别等领域.

Adaptive signal deconvolution problem is reviewed by the new view point of the time series analysis. For unknown signal described by unknown multivariable sta-tionary ARMA model, and observed through a known multivariable linear system, based on ARMA innovation model, using the time series analysis method and the pro-jection theory, this paper presents a new optimal deconvolution smoother and the cor-responding self-tuning deconvolution smoother, that is different from Goodwin and Sin's result which is based on the state -speace model. A simulation example is given to show the usefulness of results proposed. The results of this paper involve many problem and results in references as special cases, and can be applied to seismic deconvolution, signal processing, communication, speech identification and other fields.