Proceedings of 1994 28th Asilomar Conference on Signals, Systems and Computers (1994)
Pacific Grove, CA, USA
Oct. 31, 1994 to Nov. 2, 1994
Dongsheng Wu , Dept. of Electr. Eng., Maryland Univ., Baltimore, MD, USA
J.M. Morris , Dept. of Electr. Eng., Maryland Univ., Baltimore, MD, USA
One of the major obstacles for quadratic time-frequency representations to be used in real (finite)-time applications is the computational complexity. In this paper, a real (finite)-time implementable subclass of Cohen's (1966) class of distributions (CCD) is defined using windowing techniques. The kernel function for each member in this subclass can be characterized by a necessary and sufficient real(finite)-time implementable condition (RTIC) imposed on "global" kernels and "local" kernels in the ambiguity function domain. By choosing special "local" kernels, one can obtain the spectrogram (SPG) and the pseudo-Wigner-Ville distribution (PWVD) as special cases of this subclass. Our results show that time-frequency representations suitable for real-time implementation can be obtained that provide better resolution tradeoff between the time and frequency domains than occurs with the SPG and better frequency interference than occurs with the PWVD.<
signal representation, signal sampling, spectral analysis, Wigner distribution, time-frequency analysis, signal resolution
Dongsheng Wu and J. Morris, "A real-time implementable subclass of quadratic time-frequency representations," Proceedings of 1994 28th Asilomar Conference on Signals, Systems and Computers(ACSSC), Pacific Grove, CA, USA, 1995, pp. 64-67.