Modeling and Coverage Analysis of K-Tier Integrated Satellite-Terrestrial Downlink Networks

TL;DR

This paper develops a mathematical framework to analyze the downlink coverage of multi-tier integrated satellite-terrestrial networks, considering spatial distribution, fading effects, and frequency orthogonality, to optimize network performance.

Contribution

It introduces a tractable analytical approach modeling base stations on concentric spheres and derives coverage expressions considering Shadowed-Rician fading and orthogonal bands, filling a gap in performance analysis of ISTNs.

Findings

Coverage is maximized at an optimal user density ratio.

Derived analytical expressions accurately predict coverage performance.

Simulation results validate the theoretical model.

Abstract

Integrated satellite-terrestrial networks (ISTNs) can significantly expand network coverage while diminishing reliance on terrestrial infrastructure. Despite the enticing potential of ISTNs, there is no comprehensive mathematical performance analysis framework for these emerging networks. In this paper, we introduce a tractable approach to analyze the downlink coverage performance of multi-tier ISTNs, where each network tier operates with orthogonal frequency bands. The proposed approach is to model the spatial distribution of cellular and satellite base stations using homogeneous Poisson point processes arranged on concentric spheres with varying radii. Central to our analysis is a displacement principle that transforms base station locations on different spheres into projected rings while preserving the distance distribution to the typical user. By incorporating the effects of Shadowed-Rician fading on satellite channels and employing orthogonal frequency bands, we derive analytical expressions for coverage in the integrated networks while keeping full generality. Our primary discovery is that network performance reaches its maximum when selecting the optimal density ratio of users associated with the network according to the density and the channel parameters of each network. Through simulations, we validate the precision of our derived expressions.