Time frequency representation kernel
kernel = Ctfrker(N_Doppler, N_Delay, kernel_type) kernel = Ctfrker(N_Doppler, N_Delay, kernel_type, optional_parameters)
Number of doppler bins, number of rows in the kernel matrix
Number of delay bins, number of columns in the kernel matrix
Computed kernel
one of the following types:
Multiform tiltable exponential kernel 7 parameters required : [alpha beta gamma r tau_0 nu_0 lambda]
Radially gaussian kernel : an odd number of parameters required, the Fourier descriptors of the contour function [c a1 ... ap b1 ... bp]
Generalized marginals Choi-Williams kernel at least two parameters required : the kernel width and the angles of the branches in [0 pi]. [sigma angle_1 .. angle_2]
Wigner-Ville kernel : no parameters required
Spectrogram kernel The parameters are the window with EVEN length
/ Computes the Cohen's group Time Frequency Representation's (TFR) kernel, in the ambiguity plane.
note: more kernels will be implemented in future versions
H. Costa and G.F. Boudreaux-Bartels, Design of Time-Frequency Representations Using a Multiform, Tiltable Exponential Kernel IEEE Trans. on Signal Processing, October 1995, vol. 43, no 10, pp 2283-2301
X.-G. Xia and Y. Owechko and B. H. Soffer and R. M. Matic, Generalized-Marginal Time-Frequency Distributions, TFTS 1996, pp. 509-51