Spectral estimation of segmented signal using deconvolution technique

In many instances received data have random missing parts (segmented time signals) or they appear at regular intervals, as with the collection of data only at a certain time of observation. The spiking filter deconvolution method is proposed to improve the spectral estimates of segmented time signals. Based on the theoretical approach that the spectrum of segmented signals is the convolution of the spectrum of the original signal and the spectrum of the missing pattern, the spiking filter deconvolution method is transplanted into the frequency domain. In this case, the missing pattern has to be known and the maximum entropy method should be performed to find the spectral estimate of the missing pattern and the spectral estimate of the segmented signals in order to deconvolve the spectrum of segmented signals, and thus to recover the spectrum of the original signals