Unified Tractable Model for Large-Scale Networks Using Stochastic Geometry: Analysis and Design
AuthorsAfify, Laila H.
AdvisorsAl-Naffouri, Tareq Y.
MetadataShow full item record
AbstractThe ever-growing demands for wireless technologies necessitate the evolution of next generation wireless networks that fulfill the diverse wireless users requirements. However, upscaling existing wireless networks implies upscaling an intrinsic component in the wireless domain; the aggregate network interference. Being the main performance limiting factor, it becomes crucial to develop a rigorous analytical framework to accurately characterize the out-of-cell interference, to reap the benefits of emerging networks. Due to the different network setups and key performance indicators, it is essential to conduct a comprehensive study that unifies the various network configurations together with the different tangible performance metrics. In that regard, the focus of this thesis is to present a unified mathematical paradigm, based on Stochastic Geometry, for large-scale networks with different antenna/network configurations. By exploiting such a unified study, we propose an efficient automated network design strategy to satisfy the desired network objectives. First, this thesis studies the exact aggregate network interference characterization, by accounting for each of the interferers signals in the large-scale network. Second, we show that the information about the interferers symbols can be approximated via the Gaussian signaling approach. The developed mathematical model presents twofold analysis unification for uplink and downlink cellular networks literature. It aligns the tangible decoding error probability analysis with the abstract outage probability and ergodic rate analysis. Furthermore, it unifies the analysis for different antenna configurations, i.e., various multiple-input multiple-output (MIMO) systems. Accordingly, we propose a novel reliable network design strategy that is capable of appropriately adjusting the network parameters to meet desired design criteria. In addition, we discuss the diversity-multiplexing tradeoffs imposed by differently favored MIMO schemes, describe the relation between the diverse network parameters and configurations, and study the impact of temporal interference correlation on the performance of large-scale networks. Finally, we investigate some interference management techniques by exploiting the proposed framework. The proposed framework is compared to the exact analysis as well as intensive Monte Carlo simulations to demonstrate the model accuracy. The developed work casts a thorough inclusive study that is beneficial to deepen the understanding of the stochastic deployment of the next-generation large-scale wireless networks and predict their performance.