Date of Award:
Doctor of Philosophy (PhD)
Electrical and Computer Engineering
Rose Qingyang Hu
Rose Qingyang Hu
As the complexity of modern cellular networks continuously increases along with the evolution of technologies and the quick explosion of mobile data traffic, conventional large scale system level simulations and analytical tools become either too complicated or less tractable and accurate. Therefore, novel analytical models are actively pursued. In recent years, stochastic geometry models have been recognized as powerful tools to analyze the key performance metrics of cellular networks. In this dissertation, stochastic geometry based analytical models are developed to analyze the performance of some key technologies proposed for 5G mobile networks. Particularly, Device-to-Device (D2D) communication, Non-orthogonal multiple access (NOMA), and ultra-dense networks (UDNs) are investigated and analyzed by stochastic geometry models, more specifically, Poisson Point Process (PPP) models.
D2D communication enables direct communication between mobile users in proximity to each other bypassing base station (BS). Embedding D2D communication into existing cellular networks brings many benefits such as improving spectrum efficiency, decreasing power energy consumption, and enabling novel location-based services. However, these benefits may not be fully exploited if the co-channel interference among D2D users and cellular users is not properly tackled. In this dissertation, various frequency reuse and power control schemes are proposed, aiming at mitigating the interference between D2D users and conventional cellular users. The performance gain of proposed schemes is analyzed on a system modeled by a 2-tier PPP and validated by numerical simulations.
NOMA is a promising radio access technology for 5G cellular networks. Different with widely applied orthogonal multiple access (OMA) such as orthogonal frequency division multiple access (OFDMA) and single carrier frequency division multiple access (SC-FDMA), NOMA allows multiple users to use the same frequency/time resource and offers many advantages such as improving spectral efficiency, enhancing connectivity, providing higher cell-edge throughput, and reducing transmission latency. Although some initial performance analysis has been done on NOMA with single cell scenario, the system level performance of NOMA in a multi-cell scenario is not investigated in existing work. In this dissertation, analytical frameworks are developed to evaluate the performance of a wireless network with NOMA on both downlink and uplink. Distinguished from existing publications on NOMA, the framework developed in this dissertation is the first one that takes inter-cell interference into consideration.
UDN is another key technology for 5G wireless networks to achieve high capacity and coverage. Due to the existence of line-of-sight (LoS)/non-line-of-sight (NLoS) propagation and bounded path loss behavior in UDN networks, the tractability of the original PPP model diminishes when analyzing the performance of UDNs. Therefore, a dominant BS (base station)-based approximation model is developed in this dissertation. By applying reasonable mathematical approximations, the tractability of the PPP model is preserved and the closed form solution can be derived. The numerical results demonstrate that the developed analytical model is accurate in a wide range of network densities.
The analysis conducted in this dissertation demonstrates that stochastic geometry models can serve as powerful tools to analyze the performance of 5G technologies in a dense wireless network deployment. The frameworks developed in this dissertation provide general yet powerful analytical tools that can be readily extended to facilitate other research in wireless networks.
Zhang, Zekun, "Stochastic Geometry Based Performance Study in 5G Wireless Networks" (2019). All Graduate Theses and Dissertations. 7471.
Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .