The increasing demand for connectivity and broadband wireless access is leading to the fifth generation (5G) of cellular networks. The overall scope of 5G is greater in client width and diversity than in previous generations, requiring substantial changes to network topologies and air interfaces. This divergence from existing network designs is prompting a massive growth in research, with the U.S. government alone investing $400 million in advanced wireless technologies. 5G is projected to enable the connectivity of 20 billion devices by 2020, and dominate such areas as vehicular networking and the Internet of Things. However, many challenges exist to enable large scale deployment and general adoption of the cellular industries. In this dissertation, we propose three new additions to the literature to further the progression 5G development. These additions approach 5G from top down and bottom up perspectives considering interference modeling and physical layer prototyping. Heterogeneous deployments are considered from a purely analytical perspective, modeling co-channel interference between and among both macrocell and femtocell tiers. We further enhance these models with parameterized directional antennas and integrate them into a novel mixed point process study of the network. At the air interface, we examine Software-Defined Radio (SDR) development of physical link level simulations. First, we introduce a new algorithm acceleration framework for MATLAB, enabling real-time and concurrent applications. Extensible beyond SDR alone, this dataflow framework can provide application speedup for stream-based or data dependent processing. Furthermore, using SDRs we develop a localization testbed for dense deployments of 5G smallcells. Providing real-time tracking of targets using foundational direction of arrival estimation techniques, including a new OFDM based correlation implementation.