| As the evolution technology of the 3G technology, LTE(Long Term Evolution) has become the common choice of the world’s major mainstream carriers and has made significant progress in the global commercial process because of its higher system capacity, transmission reliability and spectrum efficiency. Undoubtedly, MIMO((Multiple-Input Multiple-Output, MIMO) technology is undoubtedly one of the core technologies of LTE. However, due to the interference of multiple data streams in a MIMO system and a plurality of common between users and a distributed processing each user receiver, joint detection at the receiving end is already not practical, and therefore the precoding technologies that using the channel state information(CSI) to adjust the transmission policy at the transmitting end have emerged, the precoding technique can effectively suppress interference between multiple data streams or a multiple users in MIMO system. It can significantly reduce the complexity of the receiver processing while significantly improve the channel capacity, thus helps MIMO obtain its unique multiplexing and diversity gain.In this paper, the precoding technique of MIMO in LTE downlink is studied. With respect to the linear pre-coding algorithm, such as ZF(Zero Forcing) precoding and MMSE(Minimum Mean Square Error) precoding algorithms., nonlinear precoding algorithms such as THP(Tomlinson Harashima Preeoding) and VP(Vector Perturbation) precoding can obtain significant performance gains at an acceptable expense of complexity.However, THP and VP precoding algorithms both introduce modulo operation, because modulo operation is used at both sending and receiving end of the THP algorithms, which will introduces the modulo loss, shaping loss and power loss. At the same time, modulo operation is also used at the receiving end of VP which will introduce modulo loss. These losses will bring large negative gain to THP and VP witch seriously affects the performance of non-linear pre-coding in MIMO systems. To solve the above problems, this paper will present improved THP and VP precoding algorithms. The specific works of this dissertation include:1. A brief review of principle and effect of the precoding technology in MIMO system is introduced, after that the design principles of linear precoding algorithms such as ZF and MMSE precoding and non-linear precoding algorithms such as THP and VP precoding are described, and thenwe verify the performance superiority of non-linear pre-coding compared to linear precoding with MATLAB simulation, at last, we summarize and compare their respective advantages and disadvantages.2. This paperanalyzesmodulo loss, shaping loss and power loss of THP algorithm due to the introduction of modulo operation at both sending and receiving end. On the one hand, we propose the Turbo coded expanded soft demapper scheme. On the other hand, the signal end may be drifted to the outside of the Modulo Boundary because of modulo operation at the receiving, we propose to expand the modulo operation factor, however this can decrease modulo loss but will increase power loss at the same time. We have to find an optimal extended modulo factor to compromise modulo loss and power loss so that the overall performance of the system can still have a gain after adding extended modulo factor.3. This paperstudies a simplified VP precoding scheme——LRA precoding, then we analysis modulo loss of VP algorithm due to the introduction of modulo operation at receiving end. On the one hand, we propose the Turbo coded expanded soft demapper scheme. On the other hand, the signal end may be drifted to the outside of the Modulo Boundary because of modulo operation at the receiving, we propose to expand the modulo operation factor, however this can decrease modulo loss but will decrease precoding gain at the same time. We have to find an optimal extended modulo factor to compromise modulo loss and precoding gain so that the overall performance of the system can still have a gain after adding extended modulo factor. |
