| Summary: | 碩士 === 國立交通大學 === 電信研究所 === 82 === Sinusoidal estimation in additive white noise is one of
important problems in digital signal processing. In this
thesis, techniques for estimating sinusoidal signals at low SNR
are presented. Our approach is to apply conventional estimators
in the subband domain. For block estimation, we choose the
iterative filtering algorithm as the frequency estimator. For
adaptive estimation, we use the prediction error filtering.
Besides, we also propose a new type of adaptive LMS algorithm
to reduce the influence of white noise. Simulation results show
that sinusoidal estimation in the subband domain gives better
performance than that in the full-band, and our new adaptive
algorithm outperform the conventional ones.
|
|---|
