DocumentCode
1198624
Title
Limiting Conditions on the Correlation Properties of Random Signals
Author
Kraus, G. ; Potzl, H.
Volume
3
Issue
4
fYear
1956
fDate
12/1/1956 12:00:00 AM
Firstpage
282
Lastpage
285
Abstract
A system of 
signals that has random character is completely characterized by its self-power spectra and cross-power spectra which, when normalized, can be combined in a correlation matrix. It is shown that the correlation matrix must fulfill the fundamental condition that none of its principal minors is negative. It is shown that no further restriction exists, that is, that all correlation matrices fulfilling this condition are realizable.
signals that has random character is completely characterized by its self-power spectra and cross-power spectra which, when normalized, can be combined in a correlation matrix. It is shown that the correlation matrix must fulfill the fundamental condition that none of its principal minors is negative. It is shown that no further restriction exists, that is, that all correlation matrices fulfilling this condition are realizable.Keywords
Fourier transforms; Frequency; Impedance; Interpolation; Noise level; Power generation; Signal generators; Smoothing methods; Thermal resistance; Voltage;
fLanguage
English
Journal_Title
Circuit Theory, IRE Transactions on
Publisher
ieee
ISSN
0096-2007
Type
jour
DOI
10.1109/TCT.1956.1086329
Filename
1086329
Link To Document