CN105656527B - A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system - Google Patents
A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system Download PDFInfo
- Publication number
- CN105656527B CN105656527B CN201510881501.XA CN201510881501A CN105656527B CN 105656527 B CN105656527 B CN 105656527B CN 201510881501 A CN201510881501 A CN 201510881501A CN 105656527 B CN105656527 B CN 105656527B
- Authority
- CN
- China
- Prior art keywords
- matrix
- signal
- transmitting
- algorithms
- antenna
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/0413—MIMO systems
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
- H04B7/0613—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission
- H04B7/0615—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal
- H04B7/0619—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using simultaneous transmission of weighted versions of same signal using feedback from receiving side
- H04B7/0621—Feedback content
- H04B7/063—Parameters other than those covered in groups H04B7/0623 - H04B7/0634, e.g. channel matrix rank or transmit mode selection
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L1/00—Arrangements for detecting or preventing errors in the information received
- H04L1/02—Arrangements for detecting or preventing errors in the information received by diversity reception
- H04L1/06—Arrangements for detecting or preventing errors in the information received by diversity reception using space diversity
- H04L1/0618—Space-time coding
- H04L1/0675—Space-time coding characterised by the signaling
- H04L1/0681—Space-time coding characterised by the signaling adapting space time parameters, i.e. modifying the space time matrix
Abstract
The present invention relates to a kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system, the i-th column signal x of transmitting terminal transmitting antenna transmitting code word matrix X in the extensive mimo system of mobile communication fieldi, after base station received signal, using MRC algorithms come to signal vector xiM-th of signal into row decoding.Advantages of the present invention is embodied in：In extensive mimo system, it is detected using the docking collection of letters number of MRC algorithms, it is not only able to the computation complexity of reduction system, and can ensure the performance of BER of system so that performance of BER of the system performance of BER close to even more than ZF algorithms.
Description
Technical field
The present invention relates to wireless communication technology fields, and in particular to a kind of full diversity based on extensive MIMO systems
VBLAST method for rapidly decoding.
Background technology
Multiple-input and multiple-output (MIMO, multiple-input as next generation mobile communication core technology
Multiple-output) technology, core concept are to be utilized respectively the transmitting of mutiple antennas progress signal in sending and receiving end and connect
It receives, to improve communication quality and improve system channel capacity.With the high speed development of wireless communication technique, to data rate, clothes
Business quality and the demand of number of users are multiplied, and the small-scale mimo system of tradition has been unable to meet requirement, driving wireless communication court
The extensive directions MIMO are developed.
In such a system, a large amount of antennas of base station equipment (being more than 100) are that more mobile subscribers service, higher to obtain
Spectrum efficiency, message transmission rate and handling capacity and better communication quality.When antenna number is very big, complexity is must to examine
The key factor considered, therefore many scholars expand research to extensive mimo system low complex degree detection algorithm.It is anti-to make
With TABU search (RTS, reactive tabusearch) algorithm, which can obtain very for BPSK or QPSK modulation
Good bit error rate performance, but error rate of system performance significantly reduces for high-order QAM modulation, therefore having pertinent literature to be directed to should
Problem proposes layering TABU search (LTS, layered tabu search) algorithm, which can obtain high-order QAM modulation
Preferable bit error rate performance is obtained, but this kind of algorithm is higher in the larger complexity of planisphere dimension.For such case,
There is scholar to propose grid and reduces algorithm (LR, lattice reduction), the instantaneous complexity and constellation sizes of the algorithm
It is unrelated with noise realization, and it is easy to hardware realization, it can effectively improve system error code while increasing system complexity on a small quantity
Rate performance.But LR algorithm is by improving the orthogonality of channel matrix to obtain the improvement of bit error rate performance, and channel matrix
Orthogonality is not directly related to error rate of system performance.For the problem, has and propose the lattice based on element in pertinent literature
Son reduces algorithm (ELR, element-based lattice reduction), the basic thought of the algorithm be by handle be
The directly related noise covariance matrix diagonal entry of system bit error rate performance reduces to obtain the improvement of system performance.But these
Algorithm complexity is still higher or error rate of system performance loss is larger.Most traditional low complex degree detection algorithm be ZF and
MMSE algorithms.Pertinent literature is had studied and is detected using the docking collection of letters number of ZF or MMSE algorithms under high data rate.In document
The performance of ZF and MMSE balanced devices is used to have made intensive studies in the case of high SNR in mimo system.Needle in pertinent literature
To the extensive MIMO systems of nonbinary LDPC code, receiving terminal is detected using the MMSE docking collections of letters number.Also in document
For the macro-diversity mimo system with flat Rayleigh fading channel, collected mail using the docking of ZF and MMSE receivers in receiving terminal
It number is detected.Since ZF and MMSE algorithms are needed to tieing up matrix inversion greatly, complexity is higher, therefore also pertinent literature carries
The signal detection algorithm based on belief propagation is gone out, this kind of algorithm need not calculate big dimension inverse of a matrix, and the algorithm is
Performance of BER of uniting is close with MMSE algorithms.
Some scholars are combined extensive MIMO technology with Space-Time Codes.For extensive in pertinent literature
Mimo system devises high code check non-orthogonal space-time block, then proposes a kind of rise based on multistage likelihood and searches for (M-
LAS, multistage likelihood ascent search) Low Complexity Decoding Algorithm, text in use MMSE algorithms pair
Channel state information is estimated.On this basis, document is directed to extensive MIMO Space-time Block Code Systems, it is proposed that Yi Zhongji
In the low complex degree channel estimation method of particle swarm optimization algorithm (CPSO, particle swarm optimization), the calculation
Method complexity is only in a linear relationship with transmitting antenna and reception antenna number, receiving terminal using based on factor graph belief propagation (BP,
Belief propagation) algorithm docking collects mail number into row decoding.It is calculated with conventional vertical detection algorithm such as ZF algorithms and MMSE
Method is different, parallel VBLAST is directed in some documents, it is proposed that the better Parallel testing algorithm of performance.Due to ZF algorithms needs pair
Big dimension matrix inversion, in order to avoid matrix inversion is to reduce complexity, pertinent literature has studied maximum ratio for VBLAST and closes
And (MRC, maximum ratio combining) algorithm, and have studied that this kind of algorithm is identical as ZF algorithm SINR performances to expire
The condition of foot.
However, in extensive mimo system, since there are many base station reception antenna number, maximum-likelihood decoding algorithm is due to it
Complexity has exponent relation with antenna number, therefore receiving terminal should not dock receipts signal interpretation using maximum-likelihood decoding algorithm.
Traditional linear decoding algorithm such as zero forcing algorithm, when there are many antenna number, the algorithm needs to carry out inversion operation to higher dimensional matrix,
Complexity is still very high.
Invention content
The purpose of the present invention is being directed to deficiency in the prior art, one kind is provided and is not only able to reduce system-computed complexity
Degree, and can ensure the fast decoding sides full diversity VBLAST based on extensive mimo system of system performance of BER
Method.
To achieve the above object, the invention discloses following technical solutions：
A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system, it is extensive in mobile communication field
In mimo system in uplink transmitting terminal antenna code word matrix X the i-th column signal xi, after base station received signal, calculated using MRC
Method is come to signal vector xiM-th of signal into row decoding, specifically comprise the following steps：
S1. it is required to carry out Channel Modeling according to algorithm：Assuming that matrix H be base station reception antenna with transmitting terminal transmitting antenna it
Between channel matrix, then matrix H dimension be Nr×Nt, element h in matrixnmIndicate m-th of transmitting antenna to n-th of reception antenna
Between channel gain coefficient, mean-square value be E [| hnm|2]=Pm,
Wherein, NtIndicate that transmitting terminal is used for the transmitting antenna number to Base Transmitter signal, NrExpression base station reception antenna number, 1
≤m≤Nt, 1≤n≤Nr,
Emit signal to be indicated with matrix X, dimension Nt× T, wherein T are transmission time slot number, element x in matrixniMeet E [|
xni|2]=EsThe i-th row vector x of=1, emission matrix XiIt indicates,
Receipt signal matrix indicate with Y, dimension Nr× T, the i-th row vector yiIt indicates, then receiving signal is
Y=HX+N
1≤i≤T in formula, N indicate the noise matrix between base station reception antenna and transmitting terminal transmitting antenna, dimension Nr
× T, element n in matrixniMeet the distribution of independent same distribution multiple Gauss, mean value 0, variance σ2, therefore noise power be E [|
nni|2]=σ2,
I-th row of receipt signal matrix are expressed as with vector
yi=Hxi+ni=VP1/2xi+ni
It is respectively p that P, which is diagonal entry, in formula1,…,Nt×NtDiagonal matrix is tieed up, element is that mean value is in matrix V
0, the independent same distribution Gaussian random variable that variance is 1, dimension Nr×Nt；
S2. the full diversity VBLAST codes that structure code check is 1：
δ is the plural number that mould is equal to 1, but δ ≠ 1 in formula.There is Nt=T, it is full diversity matrix to obtain X matrix；
S3. the docking collection of letters number of MRC algorithms is used to carry out a fast decoding：It is right
yi=Hxi+ni=VP1/2xi+ni
Reception signal y in formulaiPremultiplication HH, the value H on the equation left side after multiplicationHyiFor the code word that MRC methods detect,
Regard the code word detected as algorithm testing result, into row decoding；
In formula, HHIndicate the Hermitian transformation of channel matrix H, niFor the i-th column vector of noise matrix N.
Further, in the step S1：
As m >=i ＞ 1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm-i+1, for n-th of reception day
Line, using MRC algorithms to signal xm-i+1Into row decoding, other N at this timet- 1 signal is just used as interference signal, then receiving terminal
The ratio between average signal power and interference noise power SINR is
In formula, vector vmIt is the m column vectors of matrix V；
Element pmIt is m-th of diagonal element of diagonal matrix P；
NtIndicate that transmitting terminal is used for the transmitting antenna number to Base Transmitter signal；
NrIndicate base station reception antenna number；
Noise power be E [| nni|2]=σ2；
δ is the plural number that mould is equal to 1, but δ ≠ 1；
As 1≤m≤i-1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time isIt is connect for n-th
Antenna is received, using MRC algorithms to signalInto row decoding, receiving terminal SINR is at this time
As i=1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm, for n-th of reception antenna, adopt
With MRC algorithms to signal xmInto row decoding, receiving terminal SINR is at this time
A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system disclosed by the invention, with ZF algorithms
It compares, in extensive mimo system, is detected using the docking collection of letters number of MRC algorithms, is not only able to the calculating of reduction system
Complexity, and can ensure the performance of BER of system so that system performance of BER is calculated close to even more than ZF
Method.
Description of the drawings
Fig. 1 is Nt=4 and Nr=40, BPSK modulation system, MRC algorithms and the performance of BER of ZF algorithms compare；
Fig. 2 is Nt=4 and Nr=40, QPSK modulation system, MRC algorithms and the performance of BER of ZF algorithms compare；
Fig. 3 is Nt=10 and Nr=100, BPSK modulation system, MRC algorithms and the performance of BER of ZF algorithms compare；
Fig. 4 is Nt=10 and Nr=100, QPSK modulation system, MRC algorithms and the performance of BER of ZF algorithms compare；
Fig. 5 is Nt=40 and Nr=400, BPSK modulation system, MRC algorithms and the performance of BER of ZF algorithms compare；
Fig. 6 is Nt=40 and Nr=400, QPSK modulation system, MRC algorithms and the performance of BER of ZF algorithms compare；
Fig. 7 is system model figure；
Fig. 8 is method implementation flow chart.
Specific implementation mode
With reference to embodiment and with reference to attached drawing, the invention will be further described.
A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system, it is extensive in mobile communication field
The i-th column signal x of transmitting terminal antenna transmitting code word matrix X in mimo system uplinki, after base station received signal, using MRC
Algorithm comes to signal vector xiM-th of signal into row decoding, specifically comprise the following steps：
S1. it is required to carry out Channel Modeling according to algorithm：Assuming that matrix H be base station reception antenna with transmitting terminal transmitting antenna it
Between channel matrix, then matrix H dimension be Nr×Nt, element h in matrixnmIndicate m-th of transmitting antenna to n-th of reception antenna
Between channel gain coefficient, mean-square value be E [| hnm|2]=Pm,
Wherein, NtIndicate that transmitting terminal is used for the transmitting antenna number to Base Transmitter signal, NrExpression base station reception antenna number, 1
≤m≤Nt, 1≤n≤Nr, as shown in Figure 7；
Emit signal to be indicated with matrix X, dimension Nt× T, wherein T are transmission time slot number, element x in matrixniMeet E [|
xni|2]=EsThe i-th row vector x of=1, emission matrix XiIt indicates,
Receipt signal matrix indicate with Y, dimension Nr× T, the i-th row vector yiIt indicates, then receiving signal is
Y=HX+N
1≤i≤T in formula, N indicate the noise matrix between base station reception antenna and transmitting terminal transmitting antenna, dimension Nr
× T, element n in matrixniMeet the distribution of independent same distribution multiple Gauss, mean value 0, variance σ2, therefore noise power be E [|
nni|2]=σ2,
I-th row of receipt signal matrix are expressed as with vector
yi=Hxi+ni=VP1/2xi+ni
It is respectively p that P, which is diagonal entry, in formula1,…,Nt×NtDiagonal matrix is tieed up, element is mean value in matrix V
The independent same distribution Gaussian random variable for being 1 for 0, variance, dimension Nr×Nt；
S2. the full diversity VBLAST codes that structure code check is 1：
δ is the undetermined constant not equal to 1 in formula, therefore has Nt=T, it is full diversity matrix to obtain X matrix；
S3. the docking collection of letters number of MRC algorithms is used to carry out a fast decoding：It is right
yi=Hxi+ni=VP1/2xi+ni
Reception signal y in formulaiPremultiplication HH, the value H on the equation left side after multiplicationHyiFor the code word that MRC methods detect,
Regard the code word detected as algorithm testing result, into row decoding；
In formula, HHIndicate the Hermitian transformation of channel matrix H, niFor the i-th column vector of noise matrix N.
It should be noted that in step sl：
As m >=i ＞ 1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm-i+1, for n-th of reception day
Line, using MRC algorithms to signal xm-i+1Into row decoding, other N at this timet- 1 signal is just used as interference signal, then receiving terminal
The ratio between average signal power and interference noise power SINR is
In formula, vector vmIt is the m column vectors of matrix V；
Element pmIt is m-th of diagonal element of diagonal matrix P；
NtIndicate that transmitting terminal is used for the transmitting antenna number to Base Transmitter signal；
NrIndicate base station reception antenna number；
Noise power be E [| nni|2]=σ2；
δ is the plural number that mould is equal to 1, but δ ≠ 1；
As 1≤m≤i-1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time isIt is connect for n-th
Antenna is received, using MRC algorithms to signalInto row decoding, receiving terminal SINR is at this time
As i=1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm, for n-th of reception antenna, adopt
With MRC algorithms to signal xmInto row decoding, receiving terminal SINR is at this time
Prove that matrix X is that the method for full diversity matrix is as follows in step 2：
In the case of Setting signal matrix X and channel matrix H, since the element of noise matrix N meets independent same distribution
Multiple Gauss is distributed, therefore the element of receipt signal matrix Y also Gaussian distributed.Assuming that the signal matrix of transmitting is X1, receive
Signal matrix is Y1, the transmitting signal matrix of decoder decision is X2, pair-wise error probability is in the case of given channel matrix H
In formulaThe Frobenius norms of representing matrix A, wherein Tr () table
The mark for showing matrix will receive signal Y1=HX1+N1It substitutes into above formula, can obtain with matrix F robenius norm expression formulas
B=Tr [N in formula′HH(X1-X2)+(X1-X2)HHN '], in the case where channel matrix H is given, B is a mean value
It is for 0, varianceGaussian random variable, therefore the condition pair-wise error probability in formula is
In formulaFor horse khoum function.Enable D=(X1-X2)H(X1-X2), to matrix D
It carries out singular value decomposition and obtains D=VHThen Λ V carry out proper transformation using decomposition result to above formula, and according to the property of Q functions
MatterCondition pair-wise error probability, which can then be obtained, is
λ in formulam, m=1 ..., NtIt is characteristic value (the obviously λ of matrix Dm>=0), βmnIt is the element of matrix V H, therefore
βmnGaussian random variable, amplitude be obey mean value be 0, the rayleigh distributed stochastic variable that variance is 1, utilize rayleigh distributed
Probability density function asks mathematic expectaion just to obtain pair-wise error probability condition pair-wise error probability, and expression formula is as follows
As signal-to-noise ratio Es/σ2=1/ σ2When larger, and the eigenvalue λ of matrix DmWhen ＞ 0, i.e., when matrix D is non-singular matrix
Then 1 in above formula denominator can be ignored, and above formula becomes at this time
According to above formula it is found that moleculeMiddle index NrNtIt is exactly designed code word matrix
Diversity gain --- be full marks diversity gain.
Will be proven below matrix D is non-singular matrix, as long as proving matrix X1-X2It is non-singular matrix, then matrix D must be full rank
Matrix.If matrix X1-X2Non-singular matrix, then the certain linear independence of matrix column vector.Assuming that matrix X1And X2Column vector
It uses respectivelyWithIt indicates, then matrix X1-X2Column vector useIt indicates.Assuming that matrix X1-X2Column vector
Between it is linearly related, enable the first column vectorOther T-1 column vector linear expressions can be used, enable ki(1≤i≤T) is one
Scalar, wherein k1=1, then matrix X1-X2The first column vectorWith other T-1 column vectorsBetween can use expression
Formula is expressed as
It substitutes into above formula, and passes through proper transformation, due to X1≠X2,
It can obtain
Due to δ ≠ 1, k can be obtained by above formulai=0, i=2 ..., T, therefore above formula is invalid, so matrix X1-X2Institute
There is column vector linear independence, is non-singular matrix, therefore matrix D is also non-singular matrix.
See Fig. 8.Fig. 8 shows the implementation flow charts of MRC algorithms of the present invention and tradition ZF algorithms.
The performance of ZF algorithms and MRC algorithms compares：
Transmitting terminal emits the i-th column signal x of code word matrix XiWhen, ZF and MRC algorithms are respectively adopted after base station received signal
Come to signal vector xiM-th of signal into row decoding.It is the docking collection of letters y when using ZF algorithmsiPremultiplication (HHH)-1HH, this
In HHThe Hermitian transformation for indicating channel matrix H, requires base station reception antenna number to be no less than transmitting terminal antenna number, i.e., at this time
Nr≥Nt；It is the docking collection of letters y when using MRC algorithmsiPremultiplication HH, the receiving terminal SINR that is averaged for comparing two kinds of algorithms is as follows：
(1) as m >=i ＞ 1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm-i+1。
1. n-th of reception antenna is directed to, using MRC algorithms to signal xm-i+1Into row decoding, other N at this timet- 1 signal
With regard to being used as interference signal, then the receiving terminal SINR that is averaged is
2. n-th of reception antenna is directed to, using ZF algorithms to signal xm-i+1Into row decoding, other Nt- 1 signal is to signal
xm-i+1It does not interfere with, when to m-th of signal into row decoding, it is assumed that hmBe former channel matrix H m column elements composition to
Amount, channel matrix H remove the new channel matrix H after m column elementsmIt indicates, receiving terminal SNR is at this time
Therefore, for the SINR performances of MRC algorithms are identical as the SNR performances of ZF algorithms, then to meet
(2) as 1≤m≤i-1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time isFor n-th
MRC algorithms and ZF algorithms is respectively adopted to signal in a reception antennaInto row decoding, receiving terminal SINR and SNR at this time
Respectively
With
Therefore, for the SINR performances of MRC algorithms are identical as the SNR performances of ZF algorithms, then equally to meet
(3) as i=1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm.For n-th of reception antenna,
MRC algorithms and ZF algorithms is respectively adopted to signal xmInto row decoding, receiving terminal SINR and SNR are respectively at this time
With
Therefore, for the SINR performances of MRC algorithms are identical as the SNR performances of ZF algorithms, then to meet
In the case where δ is plural references of the mould equal to 1, due to | δ |2=1, at this time
With above formula equivalence.
Computation complexity compares：
Since transmitting terminal antenna number is N in systemt, base station receiver antenna number is Nr, the plural number of MRC and ZF can be found out
The number of multiplication number, complex addition number and real number extraction of square root.(i) the t column vectors x of antenna for base station emission matrix XtWhen,
Receiving terminal carries out computation complexity when signal interpretation using ZF algorithms, to yiPremultiplication (HHH)-1HH.We calculate matrix first
HHWith matrix H product HHH needsSecondary complex multiplication andSecondary complex addition, is then counted
Calculate inverse matrix (HHH)-1It needsSecondary complex multiplication,Secondary complex addition and NtSecondary extraction of square root fortune
It calculates, then calculating matrix (HHH)-1With matrix HHProduct (HHH)-1HHWhen needSecondary complex multiplication and NtNr(Nt- 1) secondary multiple
Number addition, then calculates (HHH)-1HHWith H products (HHH)-1HHIt is needed when HSecondary complex multiplication andIt is secondary multiple
Number addition, then calculates (HHH)-1HHH and xtProduct (HHH)-1HHHxtWhen needSecondary complex multiplication and Nt(Nt- 1) secondary
Complex addition finally calculates (HHH)-1HHWith ntProduct (HHH)-1HHntWhen need NtNrSecondary complex multiplication and Nt(Nr- 1) secondary multiple
Number addition, therefore emit t column vectors xtWhen need in totalSecondary complex multiplication,Secondary complex addition and NtSecondary extraction of square root operation, when emission matrix X, need in totalSecondary complex multiplication,Secondary complex addition and
NtT extraction of square root operation；(ii) the t column vectors x of antenna for base station emission matrix XtWhen, receiving terminal carries out signal using MRC algorithms
Computation complexity when decoding, to yiPremultiplication HH, we calculate matrix H firstHWith matrix H product HHH needsSecondary complex multiplication andSecondary complex addition, then calculates HHH and xtProduct HHHxtWhen
It needsSecondary complex multiplication and Nt(Nt- 1) secondary complex addition, finally calculates HHWith ntProduct HHntWhen need NtNrSecondary complex multiplication
And Nt(Nr- 1) secondary complex addition, therefore emit t column vectors xtWhen need in totalSecondary complex multiplication
Method andSecondary complex addition, when emission matrix X, need in totalSecondary complex multiplication
Method andSecondary complex addition.Two kinds of algorithm complexities are more as shown in Table 1.
One MRC and ZF algorithm complexities of table compare
As shown in Table 1, compared with ZF algorithms, total operation times are substantially reduced MRC algorithms.
Performance simulation
Assuming that transmitting terminal antenna number is respectively N in systemt=4,10 and 40, base station receiver antenna number is respectively Nr=40,
100 and 400, the full diversity VBLAST code word matrix such as formula of transmittingIt is shown, formula
Middle δ is the plural number that mould is 1, is enabledAssuming that channel is independent identically distributed rayleigh fading channel.The MRC that the application uses
Compared with detection algorithm has carried out performance of BER with traditional ZF linear detection algorithms.The availability of frequency spectrum is respectively 1bit/s/
BPSK and QPSK modulation systems are respectively adopted in Hz and 2bits/s/Hz.When transmitting terminal antenna number is N in systemt=4, base station connects
Receipts machine antenna number is NrWhen=40, simulation result difference is as depicted in figs. 1 and 2.When transmitting antenna number is Nt=10, receiver day
Line number is NrWhen=100, simulation result difference is as shown in Figure 3 and Figure 4.When transmitting antenna number is Nt=40, receiver antenna number is
NrWhen=400, simulation result difference is as shown in Figure 5 and Figure 6.
It is by Fig. 1 and Fig. 2 as it can be seen that less in transmitting terminal transmitting antenna number and base station reception antenna number, work as BER=
10-5When, the performance of BER point compared with traditional ZF algorithms of MRC algorithms used in the present invention under BPSK and QPSK modulation systems
There are not the gain of -0.5dB and -2dB, in other words, although MRC algorithms its computation complexity compared with ZF algorithms is relatively low, can lead
Performance of BER is caused to reduce；By Fig. 3 and Fig. 4 as it can be seen that with transmitting antenna number and reception antenna number increase, when SNR is relatively low
In the case of, the performance of BER compared with traditional ZF algorithms of MRC algorithms used in the present invention under BPSK and QPSK modulation systems
Preferably, and MRC algorithms performance of BER compared with ZF algorithms can reduce as SNR higher；By Fig. 5 and Fig. 6 as it can be seen that with hair
Penetrate continuing growing for antenna number and reception antenna number, the mistake ratio of MRC algorithms used in the present invention under BPSK and QPSK modulation systems
The performance of BER of special rate performance ratio ZF algorithms will be got well.When mobile subscriber's transmitting terminal antenna and base station receiving terminal antenna number are continuous
When increase, matrix HHWith the product H of channel matrix HHOff diagonal element constantly reduces compared with diagonal entry in H, works as base station
When antenna number tends to infinity, off-diagonal element can be ignored compared with diagonal element.Matrix H at this timeHH is approximately equal to E
HHH, i.e. matrix HHH is similar to diagonal matrix, in other words, works as Nr,NtWhen → ∞,HereIt is Nt×NtList
Position battle array, therefore MRC algorithms are docking collection of letters yiPremultiplication HH, the first item after multiplication on the right of equation be similar to diagonal matrix with
Signal vector xiProduct, make vector xiThe coefficient of middle element is more than 1；And ZF algorithms are docking collection of letters yiPremultiplication (HHH)- 1HH, the first item after multiplication on the right of equation is equal to unit matrix and signal vector xiProduct, make vector xiMiddle element is
Number is equal to 1.So in extensive mimo system, compared with traditional ZF algorithms, carried out using the docking collection of letters number of MRC algorithms
Detection, is not only able to the computation complexity of reduction system, and can ensure the performance of BER of system.
As m >=i ＞ 1, the detailed calculating process of SINR is as follows：
For n-th of reception antenna, using MRC algorithms to signal xm-i+1Into row decoding, then receiving terminal SINR is
According to formula yi=Hxi+ni=VP1/2xi+niIt is found that
Vector v in formulamIt is formula yi=Hxi+ni=VP1/2xi+niThe m column vectors of middle matrix V, element pmIt is formula yi=Hxi+
ni=VP1/2xi+niM-th of diagonal element of middle diagonal matrix P.By formulaSubstitution formulaIn, according to E [| xm-i+1|2]=1, transmitting signal xm-i+1With
Element in noise vector n all with vectorial hmIn element statistical iteration, and
Above formula is substituted into formulaIn, it can obtain
Using Laplace approximations, can obtain
To the molecule denominator in above formula simultaneously divided byIt can obtain
Using Laplace approximations, can obtain
Due toAndIt obeysDistribution, therefore above formula becomes
As m >=i ＞ 1, the detailed calculating process of SNR is as follows：
For n-th of reception antenna, using ZF algorithms to signal xm-i+1Into row decoding, then receiving terminal SNR is
In formula []*mM-th of element for indicating column vector, E in above formula [| [(HHH)-1HHHx]*m|2]=E [|xm-i+1
|2], denominator in above formula is unfolded, can be obtained
In formula []m*Indicate m-th of element of row vector, []mmM-th of diagonal entry of representing matrix.Due to making an uproar
Element in sound vector n all with vectorial hmIn element statistical iteration, and noise power be E [| nni|2]=σ2, therefore above formula can
To become
According to formula yi=Hxi+ni=VP1/2xi+niIt is found that
H=VP1/2
By H=VP1/2Formula substitutes into formulaIn, have
ReferenceIn above formula
Due to vmElement be mean value is 0, variance is 1 independent same distribution (i.i.d.) Gaussian random variable, therefore element
It obeysDistribution, i.e.,So formulaBecome
For the extensive mimo system of uplink, calculated separately MRC algorithms and ZF algorithms receiving terminal SINR and
SNR, analyzes the equal condition that should meet of two kinds of algorithm SINR/SNR performances, and to two kinds of algorithms carried out computation complexity and
Performance of BER compares.The result shows that in extensive mimo system, compared with ZF algorithms, using MRC algorithms to receiving
Signal is detected, and is not only able to the computation complexity of reduction system, and can ensure the performance of BER of system.
The above is only a preferred embodiment of the present invention, it is noted that for those of ordinary skill in the art,
Without departing from the invention herein, can also be to several improvement of the invention made and supplement, these are improved and supplement, should also regard
For protection scope of the present invention.
Claims (2)
1. a kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system, which is characterized in that in mobile communication
The i-th column signal x of transmitting terminal antenna transmitting code word matrix X in the extensive mimo system uplink in fieldi, base station received signal
Afterwards, using MRC algorithms come to signal vector xiM-th of signal into row decoding, specifically comprise the following steps：
S1. it is required to carry out Channel Modeling according to algorithm：Assuming that matrix H is between base station reception antenna and transmitting terminal transmitting antenna
Channel matrix, then matrix H dimension is Nr×Nt, element h in matrixnmIndicate m-th of transmitting antenna between n-th of reception antenna
Channel gain coefficient, mean-square value be E [| hnm|2]=Pm,
Wherein, NtIndicate that transmitting terminal is used for the transmitting antenna number to Base Transmitter signal, NrIndicate base station reception antenna number, 1≤m
≤Nt, 1≤n≤Nr,
Emit signal to be indicated with matrix X, dimension Nt× T, wherein T are transmission time slot number, element x in matrixniMeet E [| xni
|2]=EsThe i-th row vector x of=1, emission matrix XiIt indicates,
Receipt signal matrix indicate with Y, dimension Nr× T, the i-th row vector yiIt indicates, then receiving signal is
Y=HX+N
1≤i≤T in formula, N indicate the noise matrix between base station reception antenna and transmitting terminal transmitting antenna, dimension Nr× T, square
Element n in battle arrayniMeet the distribution of independent same distribution multiple Gauss, mean value 0, variance σ2, therefore noise power be E [| nni|2]
=σ2,
I-th row of receipt signal matrix are expressed as with vector
yi=Hxi+ni=VP1/2xi+ni
It is respectively p that P, which is diagonal entry, in formula1,…,Nt×NtTie up diagonal matrix, in matrix V element be mean value be 0, side
The independent same distribution Gaussian random variable that difference is 1, dimension Nr×Nt；
S2. the full diversity VBLAST codes that structure code check is 1：
δ is the plural number that mould is equal to 1 in formula, there is N in but δ ≠ 1t=T, it is full diversity matrix to obtain X matrix；
S3. the docking collection of letters number of MRC algorithms is used to carry out a fast decoding：It is right
yi=Hxi+ni=VP1/2xi+ni
Reception signal y in formulaiPremultiplication HH, the value H on the equation left side after multiplicationHyiFor the code word that MRC methods detect, will detect
The code word gone out regards algorithm testing result as, into row decoding；
In formula, HHIndicate that the Hermitian transformation of channel matrix H, ni are the i-th column vector of noise matrix N.
2. a kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system according to claim 1,
It is characterized in that, in the step S1：
As m >=i ＞ 1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm-i+1, for n-th of reception antenna, adopt
With MRC algorithms to signal xm-i+1Into row decoding, other N at this timet- 1 signal is just used as interference signal, then the average letter of receiving terminal
Number power is with the ratio between interference noise power SINR
In formula,
Element pmIt is m-th of diagonal element of diagonal matrix P；
NtIndicate that transmitting terminal is used for the transmitting antenna number to Base Transmitter signal；
NrIndicate base station reception antenna number；
Noise power be E [| nni|2]=σ2；
δ is the plural number that mould is equal to 1, but δ ≠ 1；
As 1≤m≤i-1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time isFor n-th of reception day
Line, using MRC algorithms to signalInto row decoding, receiving terminal SINR is at this time
As i=1, the signal of m-th of transmitting antenna of transmitting terminal transmitting at this time is xm, for n-th of reception antenna, calculated using MRC
Method is to signal xmInto row decoding, receiving terminal SINR is at this time
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510881501.XA CN105656527B (en) | 2015-12-03 | 2015-12-03 | A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510881501.XA CN105656527B (en) | 2015-12-03 | 2015-12-03 | A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105656527A CN105656527A (en) | 2016-06-08 |
| CN105656527B true CN105656527B (en) | 2018-09-28 |
Family
ID=56481928
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510881501.XA Expired - Fee Related CN105656527B (en) | 2015-12-03 | 2015-12-03 | A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105656527B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106549898B (en) * | 2016-09-27 | 2020-02-18 | 广东顺德中山大学卡内基梅隆大学国际联合研究院 | MIMO-OFDM system-based SSFE signal detection method and device |
| CN107248877B (en) * | 2017-05-22 | 2020-06-16 | 西安电子科技大学 | MassiveMIMO channel modeling method based on measured data |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101409604A (en) * | 2007-10-09 | 2009-04-15 | 鼎桥通信技术有限公司 | Method for eliminating serial interference in multi-input multi-output system |
| CN102035630A (en) * | 2010-12-30 | 2011-04-27 | 浙江大学 | Linear dispersion space-time coding method applicable to any antenna number |
| CN102067496A (en) * | 2008-06-24 | 2011-05-18 | 高通股份有限公司 | Methods and systems for space-time coding signal decoding using MIMO decoder |
| CN103326825A (en) * | 2013-06-26 | 2013-09-25 | 中国计量学院 | Low complexity decoding method of quasi-orthogonal space-time block codes |
-
2015
- 2015-12-03 CN CN201510881501.XA patent/CN105656527B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101409604A (en) * | 2007-10-09 | 2009-04-15 | 鼎桥通信技术有限公司 | Method for eliminating serial interference in multi-input multi-output system |
| CN102067496A (en) * | 2008-06-24 | 2011-05-18 | 高通股份有限公司 | Methods and systems for space-time coding signal decoding using MIMO decoder |
| CN102035630A (en) * | 2010-12-30 | 2011-04-27 | 浙江大学 | Linear dispersion space-time coding method applicable to any antenna number |
| CN103326825A (en) * | 2013-06-26 | 2013-09-25 | 中国计量学院 | Low complexity decoding method of quasi-orthogonal space-time block codes |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105656527A (en) | 2016-06-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101304300B (en) | Method and device for quantizing multiuser MIMO system channel based on limiting feedback | |
| CN100589597C (en) | Method and system for determining a signal vector | |
| CN101394254B (en) | Linear pre-coding method in multi-user MIMO system | |
| EP2410707A1 (en) | Space time coding method in orthogonal network and relay transmission system | |
| CN104779988B (en) | A kind of method of iteratively faster beam forming | |
| CN104935367B (en) | A kind of iteratively faster beam-forming method based on channel space sparse characteristic | |
| CN109743086A (en) | A kind of channel estimation methods of extensive mimo system | |
| CN101442351B (en) | Multiuser MIMO system with combined block diagonalization and united channel analyze feedback algorithm | |
| CN102281089B (en) | Signal detection method and device thereof of multioutput system | |
| CN106452662B (en) | A kind of method for precoding and device | |
| CN105656527B (en) | A kind of full diversity VBLAST method for rapidly decoding based on extensive mimo system | |
| CN102510324A (en) | Signal transmission method based on network coding in multi-input and multi-output Y channel | |
| CN101567716B (en) | Orthogonal random beam forming transmission method based on partial channel information feedback | |
| CN102158267A (en) | Method and system for processing signals | |
| US9479237B2 (en) | Vector selection modulation-based multi-antenna transmission method, receiving method and device | |
| CN107359920A (en) | A kind of extensive MIMO method for precoding based on tchebycheff's iteration method | |
| US9838105B2 (en) | Multiple-input multiple-output MIMO detection method, apparatus, and systems for transmit signal | |
| CN104320171A (en) | Three-dimensional beam forming space division multi-access adaptive transmission method adopting statistic channel state information | |
| CN101964667B (en) | High-efficiency multi-antenna detection method for long term evolution scheme | |
| CN102238733B (en) | User scheduling method in mobile communication system | |
| CN104113399B (en) | User choosing method based on Matrix condition number in multi-user MIMO system | |
| CN101615942A (en) | A kind of data communications method, Apparatus and system | |
| CN104868945A (en) | Multi-user downlink three-dimensional statistics beam forming adaptive transmission method using channel mean information | |
| Lu et al. | Performance based receive antenna selection for V-BLAST systems | |
| CN106330280A (en) | Large-scale MIMO pre-coding method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20180928 Termination date: 20191203 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |