李航偉 唐加山
摘 要:上行鏈路的大規(guī)模多入多出(MIMO)系統(tǒng)信號檢測中,最小均方誤差(MMSE)算法由于涉及矩陣求逆運算使得算法復雜度過高。為改善這一缺點,在傳統(tǒng)MMSE算法的基礎上,通過設計一種新的迭代方式代替矩陣求逆,使原算法的復雜度降低了一個數(shù)量級。理論分析及仿真表明,改進算法顯著改善了傳統(tǒng)MMSE算法的復雜度,僅需少量迭代次數(shù)就能快速向理想MMSE矩陣求逆檢測算法收斂。
關鍵詞:大規(guī)模MIMO;信號檢測;MMSE;迭代
DOI:10. 11907/rjdk. 182084
中圖分類號:TP312文獻標識碼:A文章編號:1672-7800(2019)003-0070-03
0 引言
隨著移動數(shù)據(jù)業(yè)務量爆發(fā)式增加,傳統(tǒng)的多輸入多輸出技術(multiple-input multiple-output,MIMO)因其僅能提供4×4或者8×8天線規(guī)模的系統(tǒng)而顯得力不從心,大規(guī)模MIMO技術應運而生。大規(guī)模MIMO系統(tǒng)指在基站端配置多達幾十甚至數(shù)百根天線陣列同時服務于多個單天線用戶終端,大大提高了系統(tǒng)的頻譜和能量效率[1-3]。但是,隨著天線數(shù)量的增加,大規(guī)模MIMO系統(tǒng)也面臨一些問題,如何實現(xiàn)高效可靠的上行鏈路信號檢測就是其中之一[4-5]。
隨著基站端天線數(shù)量的大幅度增加,信道之間逐漸趨于正交,基于這個特性,線性檢測算法如最小均方誤差(minimum mean square error,MMSE)等在大規(guī)模MIMO系統(tǒng)中也具有很好的性能[6-7]。但是,這些線性檢測算法涉及復雜的矩陣求逆運算從而導致復雜度過高[8]。為降低矩陣求逆帶來的計算復雜度,采用Neumann級數(shù)展開算法用于信號檢測,但當?shù)螖?shù)大于2時,其計算復雜度又回到[O(k3)],且當基站端天線和用戶天線數(shù)量之比接近1時,會帶來明顯的BER性能損失[9]。GAO X[10]提出了Richardson迭代算法,但在迭代參數(shù)計算量較大且迭代次數(shù)較低時算法性能很差。TANG C[11]和DAI L[12]提出了Gauss-Seidel算法和Newton算法,它更多地關注精度,所以計算復雜度也較大。
本文提出了一種新的迭代算法,通過新的迭代矩陣構(gòu)造方式避免了MMSE算法復雜的矩陣求逆問題。對該迭代方式的收斂性給出理論證明并分析復雜度,證明該算法復雜度為[O(k2)],比傳統(tǒng)MMSE算法降低了一個數(shù)量級。
1 系統(tǒng)模型
本文研究對象是大規(guī)模MIMO系統(tǒng)上行鏈路,系統(tǒng)模型如圖1所示。該系統(tǒng)由部署N根天線的基站和K個單天線用戶組成,所有用戶以相同的時頻資源向基站發(fā)送數(shù)據(jù),通常設置為N>>K[13-14]。
因為在大規(guī)模MIMO系統(tǒng)中,濾波矩陣W是一個嚴格對角占優(yōu)且正定的矩陣,且矩陣[A=PH-Q]保留了矩陣W的所有對角線元素,所以A是一個嚴格對角占優(yōu)的Hermit矩陣,故下列不等式成立:
2.2 算法復雜度分析
Neumann級數(shù)展開算法與本文算法在不同迭代次數(shù)情況下的復雜度對比見表1。
3 仿真分析
為驗證改進算法的復雜度和檢測性能,本文對比分析了傳統(tǒng)MMSE算法和本文算法的誤比特率,并與Neumann級數(shù)展開算法進行比較。設置仿真時的傳輸信道為快衰落瑞利信道,基帶信號調(diào)制方式為QPSK,天線規(guī)模[N×K]為128×16。
如圖2所示,隨著迭代次數(shù)增加,本文算法的誤碼率隨之下降,當?shù)螖?shù)為3時,要達到[10-3]的誤比特率,本文需要的SNR與傳統(tǒng)MMSE算法相差大約0.8dB。當?shù)螖?shù)為4時,與傳統(tǒng)算法的誤碼率基本相同,說明本文算法僅需要少量的迭代次數(shù)就能快速向理想MMSE矩陣求逆檢測算法的性能收斂,即經(jīng)過幾次簡單的迭代,本文算法的檢測性能就能接近最優(yōu)。
如圖3所示,隨著迭代次數(shù)的增加,Neumann算法也可向理想MMSE矩陣求逆檢測算法的性能收斂,但該算法迭代次數(shù)為7時檢測性能才接近最佳,而本文算法僅需4次迭代,故本文算法可在更低復雜度的情況下得到相同的最佳信號檢測性能。
4 結(jié)語
本文在傳統(tǒng)MMSE信號檢測算法基礎上,利用大規(guī)模MIMO系統(tǒng)特性和迭代運算設計了一種新算法,避免了傳統(tǒng)MMSE算法所需要的復雜度較高的矩陣求逆運算。通過理論分析,該算法可使MMSE算法的復雜度從[O(k3)]降低到[O(k2)]。仿真結(jié)果表明,本文算法經(jīng)過較少的迭代次數(shù)就可達到接近理想MMSE矩陣求逆的檢測性能。
參考文獻:
[1] 尤肖虎,潘志文,高西奇,等. 5G移動通信發(fā)展趨勢與若干關鍵技術[J]. 中國科學:信息科學,2014(5):551-563.
[2] DAI L,WANG Z,YANG Z. Next-generation digital television terrestrial broadcasting systems: key technologies and research trends[J]. IEEE Communication Magazine, 2012,50(6):150-158.
[3] NGO H Q,LARSSON E G,MARZETTA T L. Uplink power efficiency of multiuser mimo with very large antenna arrays[C]. Allerton Conference on Communication,Control,and Computing,2011:1272-1279.
[4] DAI L, WANG Z, YANG Z.Spectrally efficient time-frequency training ofdm for mobile large-scale mimo systems[J]. IEEE Journal on Selected Areas in Communication, 2013,31(2):251-263.
[5] HEALTH R W,LOZANO A,MARZETTA T L. Five disruptive technology directions for 5G[J]. IEEE Communications Magazine, 2014,52(2): 74-80.
[6] RUSEK F,PERSSON D,LAU B K. Scaling up MIMO: opportunities and challenges with very large arrays[J]. IEEE Signal Processing Magazine,2012,30(1): 40-60.
[7] ZIMMERMANN E,F(xiàn)ETTWEIS G. Unbiased mmse tree search detection for multiple antenna systems[C]. Proceedings of the International Symposium on Wireless Personal Multimedia Communications,2006:1017-1023.
[8] NGO H, LARSSON E, MARZETTA T. Energy and spectral efficiency of very large multiuser MIMO systems[J]. IEEE Communications Magazine, 2012(61):1436-1449.
[9] WU M,YIN B,WANG G,et al. Large-scale MIMO detection for 3GPP lte: algorithms and FPGA implementations[J]. IEEE Journal of Selected Topics in Signal Processing, 2014,8(5):916-929.
[10] GAO X, DAI L, YUAN C, et al. Low-complexity MMSE signal detection based on Richardson method for large-scale MIMO systems[C] Vehicular Technology Conference, 2014: 1-5.
[11] TANG C,LIU C,YUAN L,et al. High precision low complexity matrix inversion based on Newton iteration for data detection in the massive mimo[J]. IEEE Communications Letters,2016,20(3): 490-493.
[12] DAI L, GAO X, SU X,et al. Low-complexity soft-output signal detection based on Gauss-Seidel method for uplink multiuser large-scale MIMO systems[J]. IEEE Transactions on Vehicular Technology, 2014, 64(10):4839-4845.
[13] MULLER R R, VERDU S. Design and analysis of low-complexity interference mitigation on vector channels[J]. IEEE Journal on Selected Areas in Communications,2001,19(8):1429-1441.
[14] YANG Z, LIU C. A new approach for fast generalized sphere decoding in MIMO systems[J]. IEEE Signal Processing Magazine, 2005,12(1):41-44.
[15] LIANG Y C, PAN G, BAI Z D. Asymptotic performance of MMSE receivers for large systems using random matrix theory[J]. Information Theory, 2007,53(11): 4173-4190.
[16] LARSSON E G. Mimo detection methods: how they work[J]. IEEE Signal Process Magazine, 2009,26(3): 91-95.
[17] MARZETTA T L. Noncooperative cellular wireless with unlimited numbers of base station antennas[J]. IEEE Transactions on Wireless Communications, 2010, 9(11): 3590-3600.
[18] LIN B, JIN H. Low complexity multiuser detection in MIMO-OFDM systems [C]. Annual IEEE India Conference,2012:1228-1233.
[19] LIU T H. Some results for the fast MMSE-SIC detection in spatially multiplexed MIMO systems[J]. IEEE Transactions on Wireless Communications, 2009,8(11): 5443-5448.
[20] NAGARAJA S,DABEER O,CHOCKALINGAM A. Large-MIMO receiver based on liner regression of MMSE residual[C]. Vehicular Technology Conference, 2013:1-5.
(責任編輯:杜能鋼)