基于Min-Sum-MSE的多用户MIMO训练序列与数据功率联合优化  

Joint optimization of training and data power based on Min-Sum-MSE for multiuser MIMO

作　　者：王丽[1,2] 胡捍英[1] 唐其辉[2] 张建安[2] 

机构地区：[1]信息工程大学信息工程学院,河南郑州450002 [2]信息工程大学电子技术学院昆明分院,云南昆明650031

出　　处：《电路与系统学报》2011年第3期44-49,共6页Journal of Circuits and Systems

基　　金：国家863计划重点项目(2009AA011504);国家重大专项(2009ZX03003-008)

摘　　要：本文研究了多用户MIMO系统中基于Sum-MSE最小准则预编码的训练序列与数据联合功率分配问题。考虑信道空间相关性,采用MMSE准则进行信道估计,在训练序列与数据总功率一定的条件下,通过分析上下行链路存在MSE对偶性,将下行Sum-MSE最小化转化为上行Sum-MSE最小化问题。然后推导了在发送端存在相关性时该功率联合优化的闭式解。相关性越强或者平坦块衰落块长度越长,分配给训练序列的总功率就越多。仿真结果也表明,这种联合功率分配在相关信道及非相关信道下,其系统的平均BER性能方面均优于数据、训练序列平均功率分配时的性能。In this paper,joint power allocation of training and data for multiuser MIMO based on minimum Sum-MSE precoding with the constraint of total power of training and data is investigated,which adopts MMSE channel estimation and considers spatial correlations.By utilizing MSE-duality,the downlink Sum-MSE minimum problem is converted into uplink Sum-MSE minimum problem.And closed-form expressions for the optimal training power are also concluded with transmit correlation.More powers for total training sequences are needed if the spatial correlation is increased or the block length of the channel is increased.Finally,simulation results also show that the optimal power allocation derived here outperforms average power allocation in terms of average BER.

关 键 词：多用户MIMO 功率分配 Sum-MSE最小准则 MMSE信道估计 

分 类 号：TN911.22[电子电信—通信与信息系统]