Method and apparatus for spectral estimations adapted to non-uniformities of sampling

摘要

A method and apparatus for digital spectral estimations of wideband analog signals in a frequency range not limited by the mean sampling frequency is disclosed. To widen the frequency range by elimination of aliasing while sampling an analog input signal at mean frequency considerably lower than the upper signal frequency, signal sample values are taken non-uniformly at pseudorandom time instants. To reduce the estimation cross-interference errors due to the non-uniformities of the sample value taking process, the obtained sample values are subdivided into a number of periodic phase-shifted pseudo-randomly decimated sample value sub-streams. A special signal processor, adapted to the specific non-uniformities of the particular sample value sub-stream, is used for performing spectral estimations by processing each of these sub-streams. As these sample value sub-streams are basically periodic with some sample values pseudo-randomly skipped, well-defined aliasing takes place at this stage. However, the aliasing conditions for processing each of the sample value pseudorandom sub-sequence are different due to variant phase-shifts of the respective periodic pseudo-randomly decimated sampling processes. Aggregating of processing results from all processors provides spectral estimates not corrupted by aliasing and errors due to the sampling non-uniformities.