摘要多基站多用户MIMO（Multi-user MIMO, MU-MIMO）系统由于频谱利用率较高在无线通信持续发展的当今成为无线通信的研究热点，但此系统存在严重的用户间干扰和基站间干扰，通常通过发射端的预编码和接收端波束成型能有效地抑制或消除这些干扰，从而提高系统速率与误码性能。本文针对发射端预编码技术进行了研究，获得如下结论：62968

1）构建了协作多基站MU-MIMO系统预编码仿真模型，并在此基础上研究和比较了六种线性预编码算法：奇异值分解（Singular Value Decomposition, SVD）算法、迫零（Zero Forcing, ZF）、块对角化（Block-Diagonalization, BD）算法、最大化信泄噪比（Maximum Signal-to-Leakage-and-Noise Rate, Max-SLNR）算法、最小均方误差（Minimum Mean Square Error, MMSE）算法。通过仿真分析得出： SVD的性能是最优的，但它需要用户端充分协作，在实际应用中很难实现；MMSE与Max-SLNR的性能差于SVD，但是它们在具有与BD相同计算复杂度时速率与性能均优于BD算法。

2）考虑到信道状态信息存在误差时，系统性能会发生变化。我们仿真了修正后的Max-SLNR算法和MMSE算法，发现随着信道估计误差的增大，系统性能会逐渐变差。

3）为了进一步提升MIMO协作多基站MU-MIMO系统性能，本文在1）中构建的模型基础上研究了非线性预编码算法：脏纸编码和Tomlinson Harashima Precoding算法。研究获得如下结论：与线性预编码算法相比，非线性算法能提供更高的系统吞吐量，但其计算复杂度较高。

关键词  多基站多用户，MIMO，线性预编码，信道估计误差，功率分配，非线性预编码

毕业设计说明书（论文）外文摘要

Title    Beamforming Schemes for Cooperative Multi-base Station Communication System                     

Abstract Due to higher spectral efficiency, multi-cell multi-user MIMO (Multi-user MIMO, MU-MIMO) system has currently became hot research topic with rapid development of wireless communication .However, this system brings the interference between users and the interference between base stations, transmit precoding and receive beamforming are usually adopted to suppress or eliminate such interference effectively, to enhance the sum-rate and error performance. In this paper, we studied the transmit precoding techniques and obtained the following conclusions:

1) Simulation model of cooperative multi-cell MU-MIMO system with precoding are constructed, and base on this model we studied and compared six linear precoding algorithms: singular value decomposition (SVD), zero-forcing (ZF)、block  diagonalization (BD), maximum signal-to- leakage-and-noise ratio (Max-SLNR), minimum mean square error (MMSE). From simulation analysis, we find: the performance of SVD is optimal, but it requires a full user cooperation which is difficult to be achieved in practice; MMSE and Max-SLNR are worse than SVD but their sum-rates and performance are better than BD when they have the same computational complexity as BD.

2) In the presence of channel estimation error, the system performance will become worse. We find that as channel estimation error increases, the system performance will gradually deteriorate.

3) In order to further improve the performance of MU-MIMO system, we studied two nonlinear precoders : dirty paper coding and Tomlinson Harashima precoding. After studying we obtained the following conclusions: compared with those linear precoding schemes, nonlinear precoders can provide a higher system throughput, however their computational complexity is higher.

Keywords  multi-base station, multi-user, MIMO, the linear precoding, channel estimation error, power distribution, nonlinear precoding

目录

1  绪论