空时相关MIMO信道下的空时联合Huffman有限反馈预编码

34 11 2013 11

Journal on Communications

Vol.34 No. 11 November 2013

doi:10.3969/j.issn.1000-436x.2013.11.020

MIMO Huffman └ 8

ē ē ē

┦ 211100

MIMO J Huffman └ 8 8 └ ┑ : Ё Huffman Huffman

Ё TN911.22 A 1000-436X(2013)11-0179-05

Joint space-time Huffman limited feedback precoding for

spatially and temporally correlated MIMO channels

JU Mei-yan, GE Xin, LI Yue-heng, TAN Guo-ping

College of Computer and Information, Hohai University, Nanjing 211100, China

Abstract: For the MIMO channels with space correlation and time correlation, a novel joint space-time Huffman limited feedback precoding scheme was proposed which improves the system performance and reduces the amount of feedback. Based on space correlation, the precoding structure under zero-forcing (ZF) criterion was derived and the rotating quan-tization codebook was designed which reduces the effect of space correlation on system performance. In addition, in view of time correlation of channels, the scheme reduces the feedback data of channel state information (CSI) in the slow fad-ing channel by using neighborhood-based limited feedback. Due to different probabilities of codewords in the neighbor-hood, Huffman coding was adopted to further reduce the amount of feedback.

Key words: quantization codebook; space and time correlation; neighborhood; Huffman coding

1

MIMO └ 8 ˊ 8 (CSI, channel state information)[1] ┉ Ё ▂ E ⑤ CSI [2]

MIMO Grassmann

┉ Ё [3]

Grassmann [4] [5] Grassmann [6] 8

zero-forcing) (ZF

8 Grassmann MIMO 8

2013-04-02 2013-05-16 - - (61001068)

Foundation Item: The National Natural Science Foundation of China(61001068)

g180g 34

┉ Ё MIMO г ┑ [7] └ Ё ┑ [7]Ё 8 ┑ E Huffman [8] Huffman ┑

MIMO Huffman ┑

2 8

( )T ( )H

Tr( ) F Frobenius CN(µ,σ2) µ σ2

MIMO 1 Nt Nr

1 MIMO └ 8

MIMO Kronecker [9]

H=R1/2rGR1/2t

(1)

Ё RH/2t=RtR1/2t

R1/2

H/2

r=RrRr

G

εx M

8 F Nt×M (M Nt) G ˊ

x=

+Gn (2)

Ё n CN(0,N0) 8 FHF=IM Tr(FHF)=M Grassmann 8 [5] Grassmann

[10] А SNRmin ZF8

SNRmin σmin{HF}

εx

MN (3)

Ё σmin{HF} HF SNRmin σmin{HF}

[11] σmin{HF} σmax{HF}

(4)

HF

222F

=σmin{HF}+σmax{HF}

(5)

σ211min{HF} 2

(σ2min{HF}+σ2max{HF})=

2

HF2

F

(6)

HF

2F

=H1/22=∑H1/2

2wRtF

F

wwVwRtF

F

H2=∑Vw

R1/2t

F

σ2

HR1/221

w

t

F

(7)

F

F

Ё H1/2w=RrG Uw ∑w Vw Hw ∑ Vw ∑w Vw M σ1 Hw R1/2t=∑tVHt ∑t Vt Rt

VHH

wR1/2tF=Vw∑HtVtF

(8)

(4)~ (8)

σ2

{HF} 1

Hminσ21w∑tVH22tF

(9)

F

11 MIMO Huffman └ 8 g181g

А F Hw

∑H

2tVtF

FF=cVH/2H∑H

2t∑tVw=cRt

Vw w

tVtF

F

c=1 RH/2tVw Nt×M

RH/2tVw 8

RH/2tVw 2 Rt Vw Hw Hw Grassmann

Grassmann Vw Grassmann Vw

1) 8 Grassmann P P={P1, ,Pn} n Ё 2) RH/2t Grassmann C C={C1, ,Cn}

Ё CH/2i=RtPi i=1, ,n 3) Ё

CHC=IM Ё F

F={F1, ,Fn}

Ё FH0.5i=Ci(CiCi) i=1, ,n 3 Huffman

┉ Ё

г г [7]┑

└ ε Fi Wi Н

Wi={Fj|d(Fi,Fj)<ε} i,j∈{1, ,n} (10) Ё d(Fi,Fj) Fi FjП Chordal [12] Fi Fi Fi ε Ё ┑

Ё г Ё Huffman [8] М Huffman

1) + 2 8 F={F1, ,Fn}

2) Fi i=1, ,n Wi Fi ┑ Wi={1, ,N} Ё N 1 Fi 2 Fi П

3) Wi Huffman N=8 1 8 Huffman 01,11,000,001,101,1000,10010, 10011

4) WiЁ

Hw

∑H

2tVtFi

F

8

5) 2)~4) Ё [7] : Huffman ┑

Ё Fi ā01ā Fi)

Huffman ┑

4

MIMO Grassmann Ё Nt=4

g182g 34

Nr=2 M=2 Jakes dt=2λ dr=0.5λ λ 0° 60° 2° PAS QPSK Grassmann 4 bit Kronecker ZF ˊ 8 SVD8 [1], ˊ 2 3 Grassmann (BER, bit error rate)

2 8

3 8

2 3Ё MIMO BER Grassmann

ˊ SVD8

Ё Huffman N=8 Huffman 2 4 5Ё

Huffman Huffman BER

4

5

4 : Щ 4Ё :

11 MIMO Huffman └ 8 g183g

: : 2 П г Huffman Huffman

5 2 8 5Ё ┑ Ё П Щ Huffman Huffman 5Ё Huffman: г

5

MIMO Huffman └ 8 Grassmann ┑ : Huffman ┑

[1] MAI H V, PAULRAJ A J. MIMO wireless linear precoding[J]. IEEE

Signal Processing Magazine, 2007, 24(5):86-105.

[2] PITAVAL R A, MAATTANEN H L, SCHOBER K, et al. Beamform-ing codebooks for two transmit antenna systems based on optimum grassmannian packings[J]. IEEE Transactions on Information Theory, 2011, 57 (10):6591-6602.

[3] TURKI I, KAMMOUN I, SIALA M. Precoding for MIMO systems

with Grassmannian codebooks[A]. 2012 Third International Confer-ence on Communications and Networking[C]. Hammamet, 2012. 1-7. [4] LI S H, JIA H Y, KANG J F. Robust codebook design based on unitary

rotation of grassmannian codebook[A]. IEEE 72nd Vehicular Tech-nology Conference Fall[C]. Ottawa, Canada, 2010. 1-9.

[5] LOVE D J, HEATH R W J. Grassmannian beamforming on correlated

MIMO channels[A]. IEEE Global Telecommunications Conference[C]. Texas, USA, 2004. 106-110.

[6] RAGHAVAN V, VEERAVALLI V V, SAYEED A M. Quantized mul-timode precoding in spatially correlated multiantenna channels[J]. IEEE Trans on Signal Procesing, 2008, 56(12):6017-6030.

[7] , , . MIMO Ё └ [J].

, 2010, 32(7):2343-2347.

WU M, SHEN C, QIU Z D. Neighborhood-based limited feedback approach in MIMO systems[J]. Journal of Systems Engineering and Technology, 2010, 32(7):2343-2347.

[8] REN W Z, WANG H B, XU L M, et al. Research on a quasi-lossless

compression algorithm based on huffman coding[A]. 2011 Interna-tional Conference on Transportation, Mechanical, and Electrical En-gineering (TMEE)[C]. Changchun, China, 2011. 1729-1732. [9] POURAHMADI V, KOHANDANI F, MOBASHER A. On the accu-racy of channel modeling based on the Kronecker product[A]. IEEE 72nd Vehicular Technology Conference Fall[C]. Ottawa, Canada, 2010. 1-5.

[10] HEATH R W J, SANDHU S, PAULRAJ A J. Antenna selection for

spatial multiplexing systems with linear receivers[J]. IEEE Communi-cations Letters, 2001, 5(4):142-144.

[11] . [M]. : , 2004.

ZHANG X D. Matrix Analysis and Applications[M]. Beijing:Tsinghua University Press, 2004.

[12] PITAVAL R A, TIRKKONEN O, BLOSTEIN S D. Density and

bounds for Grassmannian codes with chordal distance[A]. IEEE In-ternational Symposium on Information Theory Proceedings[C]. Saint Petersburg, Russia, 2011. 2298-2302.

1979- ˊ

1987- ˊ

1971- ˊ

1975-

本文来源：https://www.bwwdw.com/article/p324.html