Please use this identifier to cite or link to this item:
http://hdl.handle.net/2080/489
Full metadata record
DC Field | Value | Language |
---|---|---|
dc.contributor.author | Swain, A K | - |
dc.contributor.author | Subudhi, B | - |
dc.date.accessioned | 2007-09-18T07:06:20Z | - |
dc.date.available | 2007-09-18T07:06:20Z | - |
dc.date.issued | 1997 | - |
dc.identifier.citation | Journal of Indian Institute of Science, Vol 77, Iss 1, 63-69 | en |
dc.identifier.uri | http://journal.library.iisc.ernet.in/archives/V77-1.html | - |
dc.identifier.uri | http://hdl.handle.net/2080/489 | - |
dc.description | Copyright belongs to Indian Institute of Science | en |
dc.description.abstract | The paper presents a neural net-based scheme embodying linear prediction techniques and the SVD algorithm to estiimate the parameters of exponetially damped sinusoids satisfactorly under low SNR conditions. In the method proposed a three layer feed-forward neural network is employed at the output of the SVD block for suppressing bias in the estimated singular values due to the presence of noise. The ANN block is used to keep the singular values constant at their noiseless counterpart, even at SNR less than 0 dB. The method is considered to be the most efficient for parameter estimation at very low SNR. | en |
dc.format.extent | 1586264 bytes | - |
dc.format.mimetype | application/pdf | - |
dc.language.iso | en | - |
dc.publisher | Indian Institute of Science, Bangalore | en |
dc.subject | Singular value decomposition | en |
dc.subject | artificial neural networks | en |
dc.subject | linear predictive coding | en |
dc.subject | backward linear prediction | en |
dc.title | Artificial neural network approach for parameter estimation of exponentially damped sinusoids using linear prediction | en |
dc.type | Article | en |
Appears in Collections: | Journal Articles |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
artificial1997.pdf | 1.55 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.