The frequency of an unmodulated (CW) carrier frequency transmission may be estimated by identifying the location of the peak value of the spectrum of a windowed sample of T-seconds duration. This estimate will, in general, be somewhat in error, and the error will depend on the length and shape of the window and the signal-to-noise ratio at the observing receiver. This paper quantitatively evaluates the effect of window shape and length (leakage) as well as the effect of additive broad-band noise on the accuracy of this method of estimation. In addition, it compares the method with an optimum “least mean-squared error” technique. An interesting result is that neither elaborate “shaping" of the window, nor the use of the “least mean-squared" error method will have much advantage over the use of the spectral peak in a rectangularly windowed sample, unless the signal-to-in-band-noise ratio is of the order of 50 dB or better.
|Original language||English (US)|
|Number of pages||11|
|Journal||IEEE Transactions on Instrumentation and Measurement|
|State||Published - Dec 1982|
ASJC Scopus subject areas
- Electrical and Electronic Engineering