| Summary: | 碩士 === 國立臺灣科技大學 === 電子工程技術研究所 === 86 === Most of the conventional methods dealing with adaptive filters in a nonstationary environment are commonly accomplished by fine tuning an adjustable parameter to track the time-variant characteristics of the unknown systems. In this thesis, a new approach has been proposed to deal with the time-variant system estimation by adaptive filter array, which collects the necessary information from a collection of adaptive filters with multiple adjustable parameters of similar values and yields better estimate of these adjustable parameters in order to track the time- variant characteristics. The goal of this adaptive filer array is to adjust the parameters according to the time- variant characteristics of the unknown system, but in the meantime, the extra computation requirement incurred is limited. In this thesis, a LRLS adaptive filter formed by adjusting order and forgetting factor has been applied in the tracking and estimation of time-variant system.
|
|---|
