3GPP Long Term Evolution: Performance Analysis and Evolution towards 4G with Coordinated Multi-Point Transmission
David Martín-Sacristán Gandía
In today's information society, there is a growing need to access data communication services ubiquitously, with mobility and increasingly higher data rates. This society's demand has motivated the development of the fourth generation of mobile communications (4G) and its evolution towards the fifth generation (5G). This development has required a revolution on the radio interface of the mobile communications systems, and, consequently, has significantly modified their capabilities and their radio resource management. This is the case of the technology known as Long Term Evolution (LTE) and its 4G version called LTE-Advanced. This Doctoral Thesis addresses the modelling, the radio resource management analysis, and the performance evaluation of the downlink of LTE and LTE-Advanced where, among the different features of LTE-Advanced, the focus is on the Coordinated Multi-Point (CoMP) transmission. The Thesis provides a detailed description of the main characteristics of LTE and LTE-Advanced. The high complexity of these systems, has prompted the use of computer simulations as the primary research methodology. The Thesis makes a detailed description of the simulation methodology and the system modelling required, including some contributions of the author in this field. Among them, it is of significant relevance the link-level simulation results used in the European project WINNER + for the LTE evaluation. With regard to the analysis of the radio resource management in LTE, the fundamentals of link adaptation and scheduling are explained in the first place. In relation to the scheduling, the Thesis includes a thorough study of the proportional fairness concept and the suboptimal implementation typically used in LTE to maximize this metric. This study has resulted in a series of ideas embodied in a modification of the typical implementation, which has proved to be capable of increasing the proportional fairness of the resource allocations. Moreover, the link adaptation analysis has revealed the "flash-light" effect problem, which is characterized by a high interference variability due to rapid changes in the scheduling decisions. The Thesis demonstrates that a particular implementation that stabilizes the scheduling decisions can improve the system performance. The radio resource management analysis of this Thesis is completed with the study of CoMP. Specifically, the CoMP scheme studied in this Thesis is a solution with coordinated scheduling and beamforming (CS/CB), that takes into account realistic and robust assumptions concerning the knowledge that the coordinated points have about the channel state. The Thesis proposes this solution for its simplicity and its ability to improve high data rates coverage and capacity even with incomplete channel knowledge. Concerning LTE and LTE-Advanced evaluation, it is performed in two different types of scenarios. On the one hand, the scenarios defined in the process of evaluation of IMT-Advanced. In this framework, it is evaluated the importance of different multi-antenna techniques, including CoMP, considering full-buffer traffic models. The most important conclusions in these scenarios are the significant performance improvement achieved with spatial multiplexing of users and the fact that CoMP mechanisms provide a reduced benefit. The second group of scenarios are those defined by the European project METIS for the evaluation of 5G technologies. Specifically, an indoor office scenario and an outdoor sports stadium have been selected. In these scenarios, a realistic traffic model is used, and it has been demonstrated the utility of CoMP to satisfy the first 5G requirement definitions with feasible frequency bandwidths. In these scenarios with less homogeneous deployments, or with a limited number of transmitters originating the major part of interference, is where this Thesis has found CoMP to be more useful and where the Thesis promotes its use.