Downlink-uplink joint beamforming and user scheduling in dynamic TDD system based on Cloud-RAN

Abstract

Since smart phones, tablets, and machine-to-machine devices are being a part of everyday life, the data traffic volume of mobile devices is explosively increased in recent years. As well as the volume of mobile data traffic, the variability of occurrence of mobile data traffic is increased dramatically. The occurrence of mobile data traffic with the time and the place is highly fluid. A massive volume of downlink data traffic (or uplink data traffic) may gather in a certain place at a certain time period. Thus, the evolution system to support variabilities of volumes of downlink data traffic and uplink data traffic actively is required. In this thesis, a dynamic Time Division Duplex (TDD) system using a Cloud Radio Access Network (Cloud-RAN) is considered as a solution system. In the conventional TDD system, the transmission service time is divided up into the downlink transmission time and the uplink transmission time. Also, the same downlink/uplink configuration is to be applied to all the cells to suppress the interference between downlink transmission and uplink transmission. However, a disadvantage that the rapid change of downlink data traffic volume and uplink data traffic volume is not supported effectively occurs because all the cells use the same downlink/uplink configuration. The dynamic TDD system is proposed to make up for this disadvantage. Since each cell can adopt their own downlink/uplink configuration according to the data traffic distribution in the dynamic TDD system, more efficient supporting of downlink data traffic and uplink data traffic is possible. In order to obtain the benefits of the dynamic TDD system, the interference which is generated between downlink transmission and uplink transmission should be managed. A Cloud-RAN consists of a centralized Base Band Unit (BBU) pool and distributed Remote Radio Heads (RRHs). A conventional base station executes all the communication procedures from computational processes to transmission processes. However, in the Cloud-RAN, the complex computational processes are rapidly and accurately calculated in the BBU pools with the centralized manner, then the computed data is delivered to RRHs. RRHs operate simple transmission and reception only. Thus, cooperative operations among RRHs are possible by means of the computational process of the BBU pool. In addition, Cloud-RAN system can achieve the gain of small cells by using low-cost RRHs which are installed massively in a certain area. The dynamic TDD system based on Cloud-RAN is the system which combines the advantages of the dynamic TDD system and the Cloud-RAN. Each RRH can select downlink transmission or uplink transmission by using the characteristics of the dynamic TDD system. Since some RRHs which support downlink transmission and other RRHs which support uplink transmission exist together, downlink data traffic and uplink data traffic can be handled simultaneously. The generated interference from the coincidence of downlink and uplink transmission is effectively managed by virtue of the powerful computations performance of Cloud-RAN. Thus, in this thesis, we consider the interference management in the dynamic TDD system based on Cloud-RAN which supports downlink transmission and uplink transmission simultaneously. For this purpose, we first focus on a beamforming design and a power allocation algorithm for the dynamic TDD system based on Cloud-RAN with single-antenna users. The proposed beamforming consists of the downlink transmit beamforming for downlink RRHs and the uplink receive beamforming for uplink RRHs. Also, the proposed power allocation algorithm is applied to uplink UEs. An iterative method is used to obtain solutions of the proposed beamforming and power allocation. We show that the proposed beamforming and power allocation algorithms can maximize the energy efficiency by appropriately handling the interference generated by simultaneous downlink and uplink transmission. Also, we consider 4 types of beamforming in the dynamic TDD system based on Cloud-RAN with multiple-antenna users. Since both RRHs and users have multiple antennas, more active beamforming can be applied to RRHs and users. The downlink transmit beamforming for downlink RRHs, the downlink receive beamforming for downlink users, the uplink transmit beamforming for uplink users, and the uplink receive beamforming for uplink users are proposed to minimize the total mean square error and the bit error rate. These 4 types of beamforming are iteratively solved to obtain converged solutions. From the simulation results, the proposed algorithm guarantees the lower mean square error and the lower bit error rate compared to conventional beamforming algorithms by managing the interference more effectively. Finally, a user scheduling in the dynamic TDD system based on Cloud-RAN is considered. Since downlink transmission and uplink transmission are supported simultaneously, two types of users should be scheduled at the same time. One is named as downlink scheduled users, and the other is named as uplink scheduled users. The user scheduling is very important to manage the interference between downlink transmission and uplink transmission because 4 types of beamforming are closely connected to each other. Though calculations of beamforming for all combinations of the user scheduling are required to obtain the optimal user scheduling results, the proposed user scheduling algorithm provides the near optimal performance without complex calculations of beamforming.