Result Details
DSP Implementation Aspects of a FLMS Frequency-Domain Adaptive Filter
Smékal Zdeněk, prof. Ing., CSc.
This article deals with one of the most widely used adaptive filters – FLMS (frequency-domain, least mean square). As the name suggests, this algorithm processes input samples in a frequency domain on an adaptive basis. The initial letter “F” in a name means “fast” and that is actually its purpose – to be an efficient real-time algorithm requiring only O(M) operations. FFT’s, fast correlations and fast convolutions are highly utilized in the kernel of this algortihm. However, computational complexity is just a single measure of an adaptive algortihm’s performance. It is just a member of a set of describing factors called together as “DSP aspects”. DSP aspect of an algortihm is any useful information relevant to DSP implementation considerations. Among others we often evaluate convergence rate, memory consumption, decomposability to simpler forms, to name a few. This paper analyzes DSP aspects of an FLMS adaptive algorithm.
@inproceedings{BUT8328,
author="Vladimír {Malenovský} and Zdeněk {Smékal}",
title="DSP Implementation Aspects of a FLMS Frequency-Domain Adaptive Filter",
booktitle="Proceedings of the 10th International Workshop on Systems, Signals and Image Processing - IWSSIP 2003",
year="2003",
pages="4",
publisher="Sdělovací technika, s.r.o.",
address="Praha, Czech Republic",
isbn="80-86645-05-3"
}