Downlink Multi-User Beamforming Technology In TD-LTE System

3GPP Long Term Evolution (LTE) is the evolution of UMTS. Spectral efficiency has been greatly enhanced because of the introduction of physical layer technology, such as MIMO/OFDM, and link layer adaptive technology, such as frequency domain scheduling, HARQ.As a downlink transmission mode, beamforming plays a major role in the elimination of inter-cell interference and performance improvement of cell edge users. Single-user dual-layer beamforming in LTE Release 9, which can further increase the system throughput, is the direct evolution of single-layer beamforming technology in Release 8. Based on SDMA, multi-user beamforming can achieve multi-user diversity gain from the different channel feature. Same time-frequency resource is allocated to different users and inter-user interference is cancelled through effective multi-user scheduling.Firstly the dissertation describes the main features of LTE as well as the research background and content. Then the principles of link adaptive technology and beamforming are specifically presented. The paper discusses the design and modeling process of TD-LTE system simulator, placing emphasis on radio resource management framework design, which pave the way for the following simulation.Then we disscuss problems aobut simulation process, parameter selection and performance evaluation metric in this paper. After that we evaluate the performance of single-user dual layer beamforming. Two different dual layer beamforming schemes are proposed. We analyze the performance gain with different channel quality indicator feedback period.Finally this dissertation details the model of beamforming weight vector and three antenna pattern curves are pictured showing the relationship between beam width and the number of antenna elements, direction of arrival, azimuth spread. Then a pairing algorithm based on Geometry is introduced, which is implemented in multi-user beamforming. Scheduling framework is improved after including the pairing algorithm. Then we evaluate the gain of multi-user beamforming over single-user beamforming and 1x2 MRC transmission. Finally we analysis and compare three different kind of pairing algorithms (Geometry, random, maximum angle). Simulation results show that average performance of multi-user beamforming is similar between random pairing and maximum angle pairing but Geometry can bring 20% performance gain.