DocumentCode :
2607391
Title :
Improved and quantified accuracy for linear spectral estimates
Author :
Ninness, Brett ; Hjalmarsson, Håkan
Author_Institution :
Dept. of Electr. & Comput. Eng., Newcastle Univ., NSW, Australia
fYear :
2000
fDate :
2000
Firstpage :
251
Lastpage :
254
Abstract :
In the context of spectral estimation, this paper examines the use of an `orthonormal basis´ wherein prior knowledge (in the form of fixed poles) may be incorporated in the solution and FIR structures are then seen as a special case of implicitly involving prior knowledge of all poles at the origin. The main technical results are ones that quantify the accuracy of the resulting spectral estimates in a way that clearly exposes how it is affected by the fixed pole choice
Keywords :
filtering theory; poles and zeros; spectral analysis; transient response; FIR structures; fixed poles; linear spectral estimates; orthonormal basis; quantified accuracy; Australia Council; Channel estimation; Computational complexity; Estimation error; Finite impulse response filter; Modems; Signal processing; Solid modeling; Spectral analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Adaptive Systems for Signal Processing, Communications, and Control Symposium 2000. AS-SPCC. The IEEE 2000
Conference_Location :
Lake Louise, Alta.
Print_ISBN :
0-7803-5800-7
Type :
conf
DOI :
10.1109/ASSPCC.2000.882480
Filename :
882480
Link To Document :
بازگشت