Two-stage digital filtering of a signal having an unknown frequency

Abstract

Two principal methods (see [1]) are currently in use for the solution of a number of practical problems in the measurement of the unknown carrier frequency of a packet of radio-frequency pulses which appear at a constant repetition frequency F_r. The first method consists in constructing a multichannel system containing narrow-band filters for each of the comparatively narrow regions Δf_i, i = 1...N of the spectrum S(f) of the expected frequencies (Fig. 1a), while the ensemble of all the channels covers the spectrum completely ("parallel" processing). The second method consists in searching for the expected signal according to its frequency and is implemented by means of a one-channel tunable system ("sequential” processing). The resolving power of the systems is determined by the passband of each of the "elementary" filters Δf and their frequency diversity δf which, in turn, depends on the statistics of the expected signal (see [1]). An excess of the voltage above the threshold level at the output of the i-th "elementary" filter is evidence of the fact that the frequency of the processed signal f_s = iF_r/N, where N is the order of the processing system and is equal to the number of channels in the system.