Towards Efficient 5G Wireless Networks: Cross-Layer Design and Wireless Networking

Abstract

With the fast growing popularity of smart mobile devices and the explosion of data-intensive services, the 5th generation (5G) wireless system is expected to provide a 1000x mobile data rate in the near future. To provide high data rate services to a large number of user devices, possible approaches include aggressive spectrum reuse, highly efficient multiplexing, and increased spectrum bandwidth. To this end, several promising technologies were proposed including massive MIMO (Multiple Input Multiple Output), small cell, and mmWave (Millimeter Wave) communication.

However, the successful applications of these emerging technologies face new challenges. For example, the high channel estimation overhead of massive MIMO systems due to large number of antennas; the interference issue in small cells network due to the dense deployment of base stations (BS); the vulnerability to blockage for mmWave communication due to short wavelength. To harvest the benefits of each technology, the 5G systems are expected to be an integration of multiple technologies. However, due to the inherent limits of each technology, such integration faces new challenges which need to be addressed with proper design.

In this dissertation work, we aim to address the key challenges of 5G emerging wireless systems from the perspectives of cross-layer design and wireless networking. Through analyzing the special properties of different technologies, we propose solutions to enable efficient integration and enhance the system performance.

The first part of this dissertation investigates the problem of dynamic BS sleep control for energy efficient massive MIMO heterogeneous network (HetNet). To achieve a good balance between data rate and energy consumption, we aim to maximize the energy efficiency of a massive MIMO HetNet through dynamic BS ON-OFF switching. Such a problem is formulated as an integer programming and we proposed centralized and distributed schemes to determine the set of small cell BSs (SBS) to be turned off.

The second part of this dissertation presents a solution of interference management in a massive MIMO HetNet with non-uniform antenna placement. We apply an antenna array configuration and processing technique called nested array for interference management. The design issue is how to use the degree of freedom (DoF) to serve users as well as nullify interference such that the network performance can be optimized. We proposed effective solutions and demonstrated the improved performance of proposed schemes with simulations.

The third part of this dissertation presents a cross layer design for wireless backhaul-based massive MIMO HetNet. We consider a joint frame design, resource allocation, and user association scheme to maximize the sum rate of a massive MIMO HetNet with wireless backhaul. We formulate the problem as an integer programming and propose an iterative solution algorithm. We show that with adaptive pilot length, i.e., the number of symbols dedicated to pilots in each frame, the system performance can be enhanced compared to the schemes with fixed frame structures.

The fourth part of this dissertation investigates the problem of duplex mode selection and resource allocation for full-duplex enabled femtocell networks. We first employ a stable roommate matching algorithm to determine the pairing strategy of users and make a selection between half duplex and full duplex based on the pairing result. We then consider channel and power allocation with the objective improving the sum rate. With the proposed scheme, the interference caused by full duplex transmission is effectively controlled and the system performance is improved compared to the cases without user pairing and adaptive mode selection.

The fifth part of this dissertation presents a BS cooperation architecture for providing high-data-rate service to large number of users with dense deployment of small cells. We propose a cooperative small cell network architecture to mitigate inter-cell interference and improve the network capacity. The key components include adaptive BS deployment and configuration, dynamic resource allocation, and interference coordination. With efficient spectrum reuse and traffic-aware scheduling, the proposed architecture effectively improves the data rate performance of small cell network when serving large number of users.

The sixth part of this dissertation presents a solution of dealing with link blockage with a combination of multiple approaches. We consider a combination of device to device (D2D) relaying and multi-beam reflection to enhance the performance of a mmWave system serving large number of users. We designed an adaptive scheme to select the set of users served by each approach. Simulation results demonstrated the performance gain achieved by adopting a combination of multiple approaches to overcome blockage.