MicrocomputerApplicationsVol.39,No.8,2023文章编号:1007-757X(2023)08-0138-04研究与设计基于张量分解的数据辅助时变信道估计微型电脑应用2023年第39卷第8期魏佳艺,孙超,许崇斌(复旦大学,信息科学与工程学院,上海200433)摘要:毫米波大规模多输入多输出(MIMO)技术是未来无线通信的核心技术之一,但大规模和快时变特性使得其信道估计成为挑战性的问题。文章在基于张量分解的信道估计的基础上,提出一种基于数据辅助的信道估计方法,可以显著降低系统的导频开销。由于数据辅助估计方法的复杂度会随数据长度明显增加,因而研究进一步将数据分段,提出对每段传输应用张量分解的方法进行估计然后对估计结果进行融合的方法,并仔细讨论了不同融合方法对最终估计性能的影响。仿真结果表明,与传统的基于张量分解的信道估计算法相比,所提出方法能够在确保估计精确度的同时显著地降低导频开销,且具有可比甚至更低的复杂度,对毫米波大规模MIMO系统的信道估计具有重要意义。关键词:毫米波大规模MIMO;时变信道估计;张量分解;数据辅助中图分类号:TN928(SchoolofInformationScienceandTechnology,FudanUniversity,Shanghai200433,China)Abstract:Millimeter-wave(mmWave)massivemultiple-inputmultiple-output(MIMO)technologyisoneofthecoretechnolo-giesinthefuturewirelesscommunication.However,itslarge-scaleandfasttime-varyingcharacteristicsmakethechannelesti-mationbeachallengingproblem.Theresearchproposesadata-aidedchannelestimationmethodbasedontensordecomposition,whichcansignificantlyreducethepilotoverheadofthesystem.Sincethecomplexityofthedata-aidedestimationmethodin-creasesobviouslywiththedatalength,theresearchfurtherproposesamethodofsegmentingthedata,estimatingeachtrans-missionsegmentbytensordecompositionandthenfusingtheestimatedresults.Theeffectsofdifferentfusionmethodsonthefinalestimationperformancearediscussedindetail.Simulationresultsshowthatcomparedwiththetraditionalchannelestima-tionalgorithmsbasedontensordecomposition,theproposedschemecansignificantlyreducethepilotoverheadwhileensuringtheestimationaccuracy,andhaslowercomplexity,whichisofgreatsignificancetothechannelestimationofmmWavemassiveMIMOsystem.Keywords:mmWavemassiveMIMO;time-varyingcha...