Abstract:
Methods for statistically characterizing the mobile radio multipath fading channel are presented. Commonly-applied statistical models are critically examined by quantifying their underlying assumptions. Improved characterizations are proposed for narrowband channels, and effective characterizations are developed for wideband channels. A new statistical model of wideband fading channels is proposed for examining empirical and theoretical justifications of existing statistical characterizations. This model is used to establish that only in limiting statistics are shadowing Iognormally distributed and fast fading Rayleigh distributed. A little known probability density function is proposed for characterizing the narrowband channel envelope which is subject to combined fast fading and shadowing. It is shown that the proposed distribution fits measured data better than both the Suzuki distribution and the more commonly applied fading distributions. Further, the proposed distribution is shown to be more convenient than the Suzuki distribution for use in performance analysis techniques such as bit error rate prediction. The effective order of time-dispersion diversity is proposed as an efficient and reliable characterization for the wideband fading channel. Two methods of parameterizing the effective order of diversity are proposed, and shown to produce accurate bit error rate predictions from mean power delay profiles. The effect of limiting the number of taps in a wideband diversity combiner is quantified analytically as an increase in symbol outage probability referred to an ideal, unlimited tap combiner.
