A Generalized Fading Model with Multiple Specular Components

Abstract

The wireless channel of 5G communications will have unique characteristics that can not be fully apprehended by the traditional fading models. For instance, the wireless channel may often be dominated by finite number of specular components, the conventional Gaussian assumption may not be applied to the diffuse scattered waves and the point scatterers may be inhomogeneously distributed.These physical attributes were incorporated into the state-of-the-art fading models, such as the κ-µ shadowed fading model, the generalized two-ray fading model and the fluctuating two ray fading model.Unfortunately, much of the existing published work commonly imposed arbitrary assumptions on the channel parameters to achieve theoretical tractability, thereby limiting their application to represent diverse range of propagation environments. This motivates us to find a more general fading model that incorporates multiple specular components with clusterized diffuse scattered waves, but achieves analytical tractability at the same time. To this end, we introduced the Multiple-Waves with Generalized Diffuse Scatter (MWGD) and Fluctuating Multiple-Ray (FMR) model that allow arbitrary number of specular components and assume generalized diffuse scattered model. We derive the distribution functions of the signal envelop in closed form and calculate second order statistics of the proposed fading model. Furthermore, we evaluate the performance metrics of wireless communications systems, such as the capacity, outage probability and average bit error rate. Through numerical simulations, we obtain important new insights into the link performance of the 5G communications while considering a diverse range of fading conditions and channel characteristics.

Stochastic Fading Channel Models with Multiple Dominant Specular Components for 5G and Beyond

“…This kind of computation is rather common in communication theory [21] and actually does not pose any numerical challenge because the integrand is a continuous bounded function and the integration interval is finite. In addition, the case N > 4 will rarely need to be considered, as the proposed model will rapidly converge to either Rayleigh or Rician fading by virtue of the central limit theorem [14,18]. As an additional benefit of this approach, we will now show that the MGF expression can be manipulated to eliminate one nested integral, so that N − 2 integrations need to be performed, i.e.…”

“…The wireless channel model given in ( 21)-( 22) for the particular case when N = 1 corresponds to the Rician shadowed fading model [23] and when N = 2 corresponds to the FTR fading model [16]. The general proposed model will be referred to as the Fluctuating N -Ray (FNR) fading model and, to the authors' knowledge, its statistical characterization has not been presented for N > 2 in a tractable form [18].…”

“…However, as a larger number of specular components is considered, i.e., N > 2, the evaluation of the PDF and CDF becomes rather hard to be numerically evaluated [9]. Expansions based on Laguerre polynomial series [17], multivariate hypergeometric functions or using the Hankel transform (which requires an improper integration of a highly oscillatory function) are no exception [18].…”