Modem Modeling and Characterization.

A method for modem modeling and characterization is presented. The properties of Fourier transforms and the properties of correlation and power spectral density for periodic and random functions are reviewed. The modems modeled are characterized by the theoretical calculation of the power spectral density and autocorrelation function of the output signal from the modem under ideal conditions. The input signal to the modems characterized is either a random pulse train or a special periodic function. Also, power spectral density curves are constructed to determine the data rate at which a modem is operating. Though no empirical data exists, the theoretical calculations indicate that modems can be characterized by the power spectral density and the autocorrelation function. More research, however, is needed for accuratey determining the data rate at which a modem is operating. (Author)