Estimation of the Composite Fast Fading and Shadowing Distribution Using the Log-Moments in Wireless Communications

Abstract

In this work, we propose a framework to obtain estimators from a variety of distributions used in composite fast fading and shadowing modeling with applications in wireless communications: the Suzuki (Rayleigh-lognormal), Nakagami-lognormal, K (Rayleigh-gamma), generalized-K (Nakagami-gamma) and α-μ (generalized gamma) distributions. These estimators are derived from the method of moments of these distributions in logarithmic units, usually known as log-moments. The goodness-of-fit of these estimators to experimental distributions has been checked from a measurement campaign carried out in an urban environment. Moreover a new method to separate fast fading and shadowing based on the Rathgeber procedure is proposed. The results conclude that the best-fitting distribution to the measurements is the Nakagami-lognormal. Also, the α-μ distribution provides an acceptable matching with the advantage of its simplicity.