Decentralized algorithms for multicell multiuser MIMO wireless mobile networks

Abstract

Small cells deployment is one of key technologies that is introduced to improve cellular communication systems’ performance, since it provides a low-cost approach to reuse system resources. However, densifying cellular systems with small cells increases the inter-cell interference (ICI), which would degrade the system performance if not properly managed. Also, small cells are expected to have a burst-like traffic with strong fluctuation between uplink and downlink traffics, since the number of users served by small cells are expected to vary strongly with time and between adjacent cells. Complementing small cells with multiple-input multiple-output (MIMO) and dynamic TDD (DTDD) technologies can be seen as a key solution to cope with ICI effects and traffic fluctuations. While MIMO technology has great potential to achieve higher throughput, improve system capacity, and enhance spectral efficiency by serving multiple users and spatially eliminate/manage interference, DTDD technology allows each cell to adaptively reconfigure its communication direction based on the prevailing traffic demands and interference levels. This thesis considers a multicell multiuser MIMO wireless network and proposes novel and decentralized algorithms for solving the following research problems. Problem 1: how to design the transmit beamforming vectors that maximize the system weighted sum-rate (WSR), while satisfying the power constraints at transmitters, Problem 2: how to design a robust transmit beamforming vectors that minimize the sum transmit power, while satisfying the users’ qualityof-service (QoS) targets in the presence of channel errors, and Problem 3: how to adaptively select the cells communication directions that maximize the users’ throughput, while jointly considering their traffic conditions and interference levels. In particular, three different and novel algorithms are proposed for solving Problem 1, which are based on the alternating optimization technique and guaranteed to converge to a local WSR-optimum. Further, a novel distributed and robust coordinated beamforming (CBF) algorithm based on alternating direction method of multipliers (ADMM) technique is proposed for solving Problem 2, where the robust beamforming is tackled using a worst-case optimization criterion. For Problem 3, a novel cell reconfiguration technique is proposed that maximizes the users’ throughput, while jointly considering both the prevailing traffic conditions and interference levels. Algorithms evaluations are carried out using computer simulation, from which the effectiveness of the proposed algorithms is evidenced, as compared to reference algorithms, in terms of spectral-efficiency, power-efficiency, convergence rate, signaling overhead, and complexity. The proposed algorithms are decentralized in the sense that each transmitter can act independently, as soon as it has the required information, which makes the proposed algorithms in this thesis especially suitable for current and future wireless networks.