【作者】 解培中；

【导师】 郑宝玉；

【作者基本信息】 南京邮电大学 ， 信号与信息处理， 2013， 博士

【摘要】 人们对便捷通信不断增长的需求推动了无线通信新技术的迅猛发展。作为3GPP LTE的关键技术,多小区之间的基站协作以虚拟MIMO的方式获得空间分集,能够有效地抑制小区间干扰,提高系统的频谱效率。然而,多小区蜂窝系统具有地理上分散的天线,作为MIMO技术在蜂窝系统中的推广,要让多小区协作处理成为现实还有许多困难要解决。首先是选择哪些基站或用户参与协作,实现以较低的复杂度达到系统整体性能的优化；其次,协作需要基站之间交换数据和信道状态信息,这给回程链路带来了较大的负担。另外,如何抑制小区间干扰也是个新课题。因此,研究蜂窝网络下行链路的基站协作技术,具体研究波束成形、用户调度、有限反馈等手段,以实现多小区协作,达到最大程度干扰抑制从而增加容量的目的依然是重要的研究课题。论文从传统的MIMO技术入手,分析了多小区协作处理的优势,针对协作用户和反馈量的选择、多小区下行链路波束成形算法以及信道状态信息的利用、能量有效性等问题,提出了一些新方法和新思路,得到了相应的研究成果。归纳起来,主要工作如下：第一,在降低复杂度方面,将选择多用户分集的思想应用到多用户MIMO系统中,以SINR作为信道质量的度量,提出了选择最大多用户分集MIMO信道调度方法,每个用户将最大的信干噪比值与设定的门限比较,只有大于门限的值及对应的发射天线序号返回给基站,基站分配独立的信道给具有最大信干噪比的用户,分析了所提方法对应的系统平均容量和归一化反馈负载的理论表达式；所提方法的性能仿真结果与分析一致。注意到MIMO系统的信道容量由信道矩阵的特征值确定,是特征值的函数,提出了基于特征分解的天线选择算法,该算法可以解决多用户MIMO系统中每个用户只有一根接收天线时用户的选择问题。最后介绍了多小区MIMO系统中选择反馈和用户选择方案。第二,针对多小区MIMO系统的小区间干扰,提出了一种新的虚拟信干噪比的度量,根据此度量提出了基于本地CSI的分布式协同波束成形算法,算法不仅考虑使噪声和干扰的和最小,而且考虑了直接链路的信道质量,该算法最大的优势是不需要迭代运行。然后研究了信道状态信息不确定时如何修正波束成形矢量,最后对几种波束成形算法的和速率性能进行了仿真,结果表明这种算法优于数字讨价还价解。第三,在频率复用的多小区多用户无线网络中,为了获得较好的和速率性能,研究了降低同频干扰的协同波束成形设计问题。建立了多小区多用户无线网络波束成形的系统模型,将下行链路波束成形设计为贝叶斯静态博弈,用博弈理论分析,结果表明,性能最优的协同波束成形矢量是自私和利他策略的线性组合。提出了一种最大化和速率的波束成形矢量迭代算法,给出了基于信道状态统计量的组合系数的估计方法。最后,仿真评估了波束成形算法的性能,表明了所提迭代算法的收敛性。第四,介绍了蜂窝网络中将分布式天线和协同波束成形相结合的数据传输方案;给出了考虑回程链路影响的有效和速率和能效函数的定义,给不同协作程度下能量效率的比较提供了统一的性能指标,在此基础上提出了三种不同协作程度对应的能量有效的功率分配方案。仿真结果进一步说明了新定义的必要性。更多还原

【Abstract】 With the growing demands of convenient communication, the rapid development of new wireless communication technology is promoted.As a key technology of3GPP LTE, BSC among multicells can have spatial diversity gain by virtual MIMO, effectively suppress ICI and improve spectral of the systems. However, as the promotion of MIMO technology in cellular system, multi-cell cellular systems usually have geographically separated antennas and several obstacles for its realization. First is the selection of base stations or mobile stations participate in cooperation in order to optimize system performance with lower complexity. Furthermore, in order to multi-cell cooperative process, large amount of data and CSI are exchanged between BSs, which bring serious burden to backhaul. In addition, how to collaborate to suppress the inter-cell interference is a new topic. Therefore, study on technologies in downlink of the cellular network, especially about beamforming, user scheduling, and limited feedback and other means are important research topics in order to achieve the maximum degree of interference suppression so as to increase the capacity. Based on the traditional MIMO technology, the dissertation analyzes the advantages of a multi-cell cooperative process, proposes some novel methods and new ideas and gets some research results about the selection of collaborative users and feedback, a multi-cell downlink beamforming algorithm as well as the application of channel state information. To sum up, the main work is as follows:Firstly, in terms of the reduction of complexity, the idea of multiuser diversity to MU-MIMO system is applied and takes SINR as a measure of the channel quality, and selects a maximum multi-user diversity MIMO channel scheduling method presented, An selective maximum multiuser diversity (SMUD) scheduling scheme which can reduce feedback load while preserving most of the capacity performance in MIMO Channels is proposed, each user compares its maximum SINR with the set-threshold and only those who fall above along with the index transmit antenna are sent back to the base station. The base station allocates independent channels to the users with the highest SINR. Average capacity and feedback load for the proposed method are obtained and shown to match the simulation very well. Note that the channel capacity of MIMO system is a function of the eigenvalues of the channel matrix, an antenna selection algorithm is presented based on the eigenvalue decomposition of channel matrix, which can solve how to choose user in MU-MIMO system when each user has only one receiving antenna. The scheme of select feedback and user options in multi-cell MIMO system are introduced.Secondly, the virtual SINR, a new measurement of inter-cell interference in multi-cell MIMO system is proposed, according to this measurement a distributed collaborative beamforming algorithms based on the local CSI is presented, considering not only to minimum the noise and interference, but also the channel quality of the direct link, the biggest advantage of the algorithm does not need iteration. The problem how to adjust beamforming vector is researched when the CSI is uncertain., several algorithms are simulated in the context of a network with uncoordinated co-channel interferers finally, and simulation results reveal that the proposed algorithm outperforms than the distributive bargaining solution (DBS).Thirdly, in order to obtain better sum-rate performances, beamforming design problem to reduce the same frequency interference is studied in the frequency multiplexing multi-user multi-cell wireless network, Multi-cell multi-user wireless network beamforming system model is established, downlink beamforming is design as Bayesian static game, theoretical analysis results with the game theory show that, the collaborative beamforming vector of the optimal performance is a linear combination of selfishness and altruism strategy, the beamforming vector iteration algorithm to maximize sum-rate is presented, and estimation method about composite coefficient based on the statistic of channel state information is proposed. Finally,the results of simulation assess the performance of the algorithm and show convergence of the proposed algorithm.Fourthly, data transmission scheme that combines distributed antenna and collaborative beamforming in the cellular network is introduced, which can improve the power efficiency of the base station. Sum-rate and energy efficiency function, which is impacted by the backhaul.A unified performance definition can compare energy efficiency of the different level of collaboration. To this end, energy-efficient power allocation strategies corresponding to three different level of collaborations are presented. The simulation results further prove the necessity of the new definition.更多还原