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

TL;DR

It is shown that the fluctuations of the specular components have a detrimental impact on performance, and the best performance is obtained when there is only one specular component.

Abstract

We introduce a comprehensive statistical characterization of the multipath wireless channel built as a superposition of a number of scattered waves with random phases. We consider an arbitrary number $N$ of specular (dominant) components plus other diffusely propagating waves. Our approach covers the cases on which the specular components have constant amplitudes, as well as when these components experience random fluctuations. These propagation scenarios are found in different use cases of 5G networks, as well as in the context of large intelligent surface based communications. We show that this class of fading models can be expressed in terms of a continuous mixture of an underlying Rician (or Rician shadowed) fading model, averaged over the phase distributions of the specular waves. It is shown that the fluctuations of the specular components have a detrimental impact on performance, and the best performance is obtained when there is only one specular component.

Figures (3)

• Figure 1: Probability density function of the received signal amplitude under NWDP fading, for different numbers of dominant specular waves $N$ and different amplitude configurations. Parameter values are $K_{\rm dB}=16$ dB and $\Omega_0=1$. Solid lines correspond to the balanced amplitude cases. Markers denote MC simulations.
• Figure 2: Probability density function of the received signal amplitude under FNR fading, for different numbers of dominant specular waves $N$ and different amplitude configurations. Parameter values are $K_{\rm dB}=20$ dB, $\Omega_0=1$ and $m=8$. Solid lines correspond to the balanced amplitude cases. Markers denote MC simulations.
• Figure 3: Outage capacity vs. average SNR, for different numbers of dominant specular waves $N$. The Rayleigh case is included as a reference (solid black line). Solid colored lines correspond to NWDP fading, and dashed lines correspond to the FNR case. Dotted thin lines are used for the asymptotic approximations using \ref{['eq:15b']} and \ref{['eq:007c']}. Parameter values are $K_{\rm dB}=14$ dB, $\Omega_0=1$, $m=5$ and $R_S=1$ bps/Hz. Markers denote MC simulations