No abstract

The problem of designing a Wiener filter for the optimum processing of amplitude-modulated {AM) signals transmitted through randomly dispersive media is discussed. Results are obtained that provide effective least-mean-squared-error estimation of the input message when errors which do not represent loss of information are neglected. These results, which differ appreciably from the results that obtain from a deterministic analysis, indicate that there are certain values of the propagation parameters which narrow the filter passband to such an extent that estimation is rendered ineffective.