Domanda a Seth Shostak del SETI Institute:
I'm curious to know the thought of Seth Shostak on KLT in Allen Telescope Array
Dear Old Lady SETI, it's time for your lifting!
FFT Fast Fourier Transform = SETI's Stone Age
FFT = inferior possibilities to hunt ETI signals!!!
Innovative SETI using the KLT!
SETI from narrow-band to wide-band: a Revolution called KLT
Author: Dr. Claudio Maccone Ph.D.
SETI searches are, by definition, the extraction of very weak radio signals out of the cosmic background noise. When SETI was born in 1959, it was �natural� to attempt this extraction by virtue of the only detection algorithm well known at the time: the Fourier Transform (FT).
In fact:
1. SETI radioastronomers had adopted the viewpoint that a candidate ET signal would necessarily be a sinusoidal carrier, i.e. a very narrow-band signal. Over such a narrow band, the background noise is necessarily white. And so, the basic mathematical assumption behind the FT that the background noise must be white was �perfectly matched� to SETI for the next fifty years!
2. In addition, the Americans J. W. Cooley and J. W. Tukey discovered in April, 1965 that all FT computations could be speeded up by a factor of N/ln(N) (N is the number of numbers to be processed) by replacing the old FT with their own Fast Fourier Transform (FFT) algorithm.
Then, SETI radioastronomers all over the world gladly, and unquestioningly, adopted the new FFT. But in 1982, the French SETI radioastronomer Fran�ois Biraud dared to challenge this view. He argued that we only can make guesses about ET�s telecommunication systems, and that the shifting trend on Earth was from narrow-band to wide-band telecommunications.
Thus, a new transform was needed that could detect signals over both narrow and wide bands, regardless of the colored noise distribution over this finite bandwidth.
Such a transform had actually been pointed out as early as 1946 by the Finnish mathematician, Kari Karhunen and the French mathematician, Maurice Lo�ve, and is thus named KLT (Karhunen Lo�ve Transform) for them. In conclusion, Fran�ois Biraud suggested to �look for the unknown in SETI� by adopting the KLT rather than the FFT.
Independently of him, the American radio astronomer, Bob Dixon, of Ohio State University, reached the same conclusions also, but published his results only much later.
And, independently of them both, this author also reached the same KLT conclusions and started �preaching the KLT for SETI� in 1987: first at the SETI Institute itself, then at the Italian CNR SETI facilities located at Medicina, near Bologna. But while Fran�ois Biraud and Bob Dixon were both halted by the difficult computational problem of finding the eigenvalues and eigenvectors of the very large symmetric autocorrelation matrices, the story took a different path in Italy.
The director of the CNR radio telescopes, Stelio Montebugnoli, was willing to help, and, little by little, young bright students (notably Domenico Caliendo) succeeded in programming the KLT algorithm at the Medicina CNR radio telescopes. By the year 2000, the advent of programmable cards finally made the �miracle� happen: the KLT for SETI was fully implemented at the SETI-Italia facilities for the first time in SETI history, and is now able to detect weaker signals than the FFT. This paper describes what the KLT mathematical revolution is, and how it works for SETI.
"Reproduced from The SETI League Web site, http://www.setileague.org/, by permission.
Copyright � 2004 by The SETI League, Inc. All rights reserved."
I'm curious to know the thought of Seth Shostak on KLT in Allen Telescope Array.
Bruno Moretti IK2WQA SETI ITALIA Team G. Cocconi
=======================================================================
Le risposte di Seth Shostak (in Italiano)
& mia replica:
S�, per�...
quel "monello" di Seth...
ha "dribblato alla Van Basten" la mia domanda sulla KLT in ATA!!! ;-)
Gh� poc de fa':
anche senza voler essere "campanilisti"
ITALIANS DO IT BETTER!!!
Per aspera ad astra
Bruno IK2WQA
Argomento correlato:
SETI ITALIA: una rivoluzione chiamata KLT