IMDEA Networks Institute Publications Repository

Enhancements in spectrum management techniques for heterogeneous 5G future networks

Sciancalepore, Vincenzo (2015) Enhancements in spectrum management techniques for heterogeneous 5G future networks. PhD thesis, Universidad Carlos III de Madrid, Spain.

[img] PDF (Enhancements in spectrum management techniques for heterogeneous 5G future networks)
Download (4Mb)


In the last decade, cellular networks are undergoing with a radical change in their basic design foundations. The huge increase in traffic demand requires a novel design of future cellular networks. Driven by this increase, a network densification phenomena is occurring, which in turns requires to devise efficient and reliable mechanisms to deal with the interference problems resulting from such densification. The architecture and mechanisms resulting from such drastic re-design of the network are commonly referred under the term '5G network'. In this context, this work unveils that current networking solutions are no longer sufficient to (i) provide the required network spectral efficiency, and (ii) guarantee the desired level of quality of experience from the user side. In order to address this problem, in this thesis we propose a novel SDN-like framework that incorporates the needed mechanisms to improve spectral efficiency while delivering the desired quality of experience to users. In particular, our architecture includes the following three contributions: Our first approach addresses the intercell interference issues resulting from high network densification. To this end, we propose novel mechanisms to mitigate the inter-cell interference problem. We address the design of such schemes from two angles: (i) a controller-aided mechanism, which gathers all the information of the network at a centralized point and, based on this information, optimally schedules the transmission from different users, and (ii) a semi-distributed mechanism, which reduces the signaling overhead involved in sending the information to a centralized point while providing close to optimal performance. One of the key novelties of our scheduling algorithms is that they are based on the Almost Blank SubFrame (ABSF) scheme; this scheme has been standardized only recently and very little work has addressed the design of algorithm to use it. Our second approach addresses spectral efficiency from a complementary angle: cellular traffic offloading for content update applications. This approach leverages high user mobility to offload the cellular downlink traffic through a device-to-device communication. In this context, we propose an adaptive algorithm to optimally transmit content to base stations while maximizing traffic offload. By relying on control theory techniques, our approach delivers near optimally performance. A third key contribution of this thesis is the design of a solution that combines the above two approaches. In particular, our solution considers that traffic offload is taking place in the network and addresses the design of an optimal scheduling algorithm that leverages on the Almost Blank SubFrame (ABSF) scheme. The combination of these kind of approaches has received little attention from the literature. The feasibility and performance of the approaches described above are thoroughly evaluated and compared against state-of-the-art solutions through an exhaustive simulation campaign. Our results show that the proposed approach outperforms conventional eICIC techniques as well as standard offloading mechanisms, respectively, and confirm their feasibility in terms of overhead and computational complexity. To the best of our knowledge, this thesis is the first attempt to design an unified framework which is able to optimally perform offloading for content-update distribution applications while boosting the network performance in terms of spectral efficiency.

Item Type: Theses (PhD)
Uncontrolled Keywords: ICIC; eICIC; 5G; scheduling; game theory; GBR; BE; control theory; SDN
Depositing User: Jeanet Birkkjaer
Date Deposited: 14 Apr 2016 14:52
Last Modified: 04 May 2016 09:22

Actions (login required)

View Item View Item