Interference management for 4G cellular networks and beyond

Abstract

The next fifth generation (5G) systems expect a traffic volume around 1,000 times greater than fourth generation (4G) systems. In order to reach this requirement, the cellular networks need to be denser and smaller which can lead to a high intercell interference, reducing the expected performance gains. Another feature for next generation networks is the increasing of sub-carriers bandwidth , which motivates the use of time division duplexing (TDD) approach due the flexibility in the use of the available spectrum. Therefore, the dynamic TDD is a natural choice for such networks due the capability of each base station (BS) to choose its own transmission direction adapting to users’ traffic fluctuations. However, if two BSs transmit in opposite directions in addition of intra and intercell interference there will be downlink (DL)- uplink(UL) interference, between BSs, and UL-DL interference between users. Thus, it must exist some kind of interference management to prevent the system performance loss due the presence of interference. In this context, this thesis considers the beamforming (BF), power control (PC) and stream allocation techniques for interference management in order to answer the following questions “How to manage with the interference in dynamic TDD networks?”, “How to determine the best stream allocation for multiple-input multiple-output (MIMO) systems?”. To answer these questions we divided this works in three parts. The first part studies the BF state of art where we compare the classical BF and PC algorithms with some approaches existing in literature in order to motivate the usage of distributed solutions for interference management. The second part addresses the interference problem for dynamic TDD networks, where we propose two novel distributed algorithms based on pricing approach. We compare our proposed algorithms with the classical BF approaches for multiple-input single-output (MISO) and MIMO systems. We also present an analysis regarding to signalization, convergence and complexity aspects. The last part addresses the stream allocation problem for MIMO networks where we present a novel greedy based algorithm which finds the best stream allocation that outperforms any fixed scheme and has a small performance gap when compared to the optimum solution, with less complexity.