Title :
Some simple parameter determination rules for the generalized Choi-Williams and Butterworth distributions
Author_Institution :
Lab. d´´Autom., Nantes Univ., France
Abstract :
Papandreou and Boudreaux-Bartels (1992, 1993) proposed two new elements of the Cohen´s class. The interesting point of these time-frequency distributions is their characteristic kernels, which differ from the product type kernels in an included dissymmetry between the Doppler and lag variables. Together with these definitions, a parameter determination scheme was proposed, which requires one to compute the ambiguity function of the analyzed signal and to discriminate there the cross-terms from the signal components. The aim of the present paper is to deduce from the cross-term reduction mechanism of these distributions a simple parameter determination rule, which requires less knowledge on the signal structure. The author also defines and studies a new distribution which is the asymptotic limit of both these distributions as their order increases.<>
Keywords :
parameter estimation; signal processing; statistical analysis; time-frequency analysis; BUD; Butterworth distribution; Cohen´s class; Doppler variables; GED; ambiguity function; analyzed signal; characteristic kernels; cross-term reduction mechanism; dissymmetry; generalized Choi-Williams distribution; generalized exponential distribution; lag variable; order; signal structure; simple parameter determination rules; time-frequency distributions; Exponential distribution; Fourier transforms; Kernel; Low pass filters; Narrowband; Passband; Signal analysis; Smoothing methods; Technological innovation; Time frequency analysis;
Journal_Title :
Signal Processing Letters, IEEE