Evolutionary spectrum estimation by positivity constrained deconvolution

We present a deconvolution technique to obtain the evolutionary spectrum (E S) of nonstationary signals by deconvolving the blurring effects of the tim e-frequency distribution (TFD) kernel from bilinear TFD's. The resulting sp ectrum is non-negative and has desirable properties such as higher resoluti on and higher concentration in time frequency. The new technique is computa tionally more efficient compared with the recently proposed entropy-based d econvolution technique, and, unlike the entropy method, it is not restricte d to deconvolution of spectrograms with Gaussian windows. This makes the me thod applicable to deconvolving many of the bilinear time-frequency distrib utions.