CN101252418A - Self-adapting transmitting method using channel statistical information in multi-aerial transmission system - Google Patents

Abstract

The self-adaptation transmission method which uses the channel statistical information in a multi-antenna transmission system can improve the frequency spectrum efficiency and the power efficiency of the communication system greatly. The transmission method includes the steps that the result of channel estimation is used to calculate the statistical information of the channel; namely, the channel coupling matrix of the character pattern/the character pattern of the transmission correlative matrix and the SNR; the statistical information can be obtained from the receiver or the transmitter; the power distribution and the precoding transmission are processed in the transmitter according to the channel statistical information of the feedback or the hidden feedback of the receiver. The self-adaptation transmission method provides a self-adaptation transmission proposal of the multi-antenna system, which uses the channel statistical information; the self-adaptation transmission method can adjust the transmission parameters according to the statistic characteristics of the channel; the obtained mutual information quantity approximates to the capacity of the channel.

Description

Utilize the adaptive transmission method of channel statistical information in the multi-aerial transmission system

Technical field

The present invention relates to a kind of mobile communication system, relate in particular to a kind of multiaerial system adaptive transmission method that utilizes channel statistical information by using a plurality of send/receive antennas to come transmitting high speed data.

Background technology

For adapting to the needs of future development, back three generations (B3G) or claim the 4th generation (4G) mobile communication system to require to support up to per second tens of million even go up the high-speed packet data transmission of gigabit, under situation about being becoming tight Radio Resource day, adopt many antenna transmission and many antennas to receive (MIMO) Radio Transmission Technology, abundant digging utilization space resources, improve the availability of frequency spectrum and power efficiency to greatest extent, become the key point of back 3 g mobile communication research.

Compare with the single antenna receiving system with traditional single antenna transmission, the channel circumstance of mimo wireless communication system is more complicated, and the factor that influences channel capacity is more.In the actual propagation environment, because the antenna distance and the limitation of scattering object on every side, channel usually exists decline correlation and line of sight, and under the situation of transmitter unknown channel state information, these factors have directly caused the decline of mimo system channel capacity.The adaptive MIMO transmission is the main means that overcome non-ideal factor in the actual propagation environment.

The adaptive MIMO transmission need utilize the prior information of channel at transmitting terminal.Under the situation of the known channel condition information completely of transmitter, " water filling " method can reach maximum channel capacity.Yet, because the time delay of time variation, channel estimating and the feedback of wireless channel and the skew of frequency make to be difficult to obtain channel condition information completely at transmitting terminal.A kind of method of compromise is to utilize the channel condition information of part, i.e. the statistic behavior information of channel.Channel statistic property, its pace of change are very slow with respect to the instantaneous state of channel, and transmitting terminal can obtain the statistic behavior information of channel reliably.Achievement in research in recent years shows that when transmitting terminal utilized partial channel-state information to carry out the precoding transmission, the channel capacity of mimo system and transmission reliability can be greatly improved.Therefore, it is suitable utilizing channel statistical information to carry out the Adaptive Transmission assessment.

Summary of the invention

Technical problem: the purpose of this invention is to provide the adaptive transmission method that utilizes channel statistical information in a kind of multi-aerial transmission system, a kind of multiaerial system adaptive transmission scheme that utilizes channel statistical information is provided, can send parameter according to the channel statistic property adjustment, the mutual information of acquisition approaches channel capacity.

Technical scheme: the adaptive transmission method of channel statistical information that utilizes of the present invention carries out as follows:

1) obtaining of transmitting terminal channel statistical information: with number of transmit antennas is N _t, the reception antenna number is N _rChannel matrix H be modeled as

Wherein U _tAnd U _rBe respectively N _t* N _tAnd N _r* N _rFixing unitary matrice, D is N _r* N _tFixing reality " diagonal matrix " (N here _r* N _t" diagonal matrix " refer to (i, j) element is 0) of matrix when i ≠ j, M is N _r* N _tFixing real matrix, H _IidBeing one is zero by average, and variance is 1 the independent identically distributed N that answers the gaussian variable composition _r* N _tRandom matrix, represents Hadama product, subscript

Represent conjugate transpose; When adopting feedback model, receiving terminal utilizes the estimated value of channel parameter, uses

With

Calculate and send the relevant battle array with reception of relevant battle array, wherein E{} represents to ask expectation.Next carry out feature decomposition to sending the relevant battle array of relevant battle array respectively with reception:

Channel couples matrix on the calculated characteristics pattern then:

Wherein

Subscript () ^*The conjugate operation of representing matrix; Calculate signal to noise ratio γ=P/ σ ², wherein P is total transmitted power; At last, receiving terminal will send relevant battle array R _t, channel couples matrix Ω and signal to noise ratio γ feed back to transmitting terminal; When employing concealed feedback model, receiving terminal fed back to transmitting terminal with signal to noise ratio γ, and transmitting terminal utilizes transmitting terminal to receive the channel estimation results of link H and the reciprocity H=H of channel ^T, adopt the method identical to calculate and send relevant battle array R with feedback model _tWith channel couples matrix Ω;

2) transmitting terminal carries out feature decomposition to the relevant battle array of the transmission in the channel statistical information:

Obtain sending direction pre-coding matrix U _Q=U _T

3) adopt iteration water-filling algorithm rated output allocation matrix at transmitting terminal according to channel statistical information $\mathrm{\Λ}=\frac{P}{N_t}\mathrm{diag}\left(\mathrm{\λ}\right),$ λ=[λ wherein ₁, λ ₂..., λ _Nt], diag (λ) expression is the diagonal matrix of diagonal entry with the element of vectorial λ, λ ₁, λ ₂..., λ _Nt〉=0; With ${\mathrm{\λ}}^k=[{\mathrm{\λ}}_1^k,{\mathrm{\λ}}_2^k,...,{\mathrm{\λ}}_{N_t}^k]$ The result who represents the power division of the k time iteration gained, λ _i ^kExpression is with λ ^kI element be made as the vector of 1 gained, C _u(λ ^k) previous term of the capacity of lower channel as a result of power division of the k time iteration gained of expression, A _(i)For matrix A is left out its remaining matrix in i row back; Iteration water-filling algorithm step is as follows:

A), initialization: set maximum iteration time K and convergence decision threshold ε, make k=0, λ ^k=1 _{1 * Nt}, C _u(λ ^k)=log ₂ Per(γ Ω), wherein 1 _{1 * Nt}Represent that a dimension is 1 * N _t, component is 1 matrix, Per () represents the permanent operator, for the matrix A of a M * N, defines its expansion permanent Per(A) be Per(A)=Per ([I _MA])=Per ([I _NA ^T]),

B), calculate $p=\left({\mathrm{\λ}}_{\left(i\right)}^k\right)=\underset{\‾}{\mathrm{Per}}\left(\mathrm{\γ}{\mathrm{\Ω}}_{\left(i\right)}\mathrm{diag}\left({\mathrm{\λ}}_{\left(i\right)}^k\right)\right)$ With $q\left({\mathrm{\λ}}_{\left(i\right)}^k\right)=\underset{\‾}{\mathrm{Per}}\left(\mathrm{\γ\Ωdiag}\left({\mathrm{\λ}}_i^k\right)\right)-\underset{\‾}{\mathrm{Per}}\left({\mathrm{\γ\Ω}}_{\left(i\right)}\mathrm{diag}\left({\mathrm{\λ}}_{\left(i\right)}^k\right)\right),$ Wherein, i=1,2 ..., N _t,

C), calculate ${\mathrm{\λ}}_i^{k+1}=\max (0,\tilde{v}-\frac{p\left({\mathrm{\λ}}_{\left(i\right)}^k\right)}{q\left({\mathrm{\λ}}_{\left(i\right)}^k\right)}),$ I=1,2 ..., N _t, wherein

Serve as reasons $\underset{i=1}{\overset{N_t}{\mathrm{\Σ}}}{\mathrm{\λ}}_i^{k+1}=N_t$ The constant of decision,

D), calculate C _u(λ ^K+1)=log ₂ Per(γ Ω diag (λ ^K+1)),

E) if C _u(λ ^K+1)≤C _u(λ ^k), then order ${\mathrm{\λ}}^{k+1}=\frac{1}{N_r}{\mathrm{\λ}}^{k+1}+\frac{N_r-1}{N_r}{\mathrm{\λ}}^k,$ Use formula C _u(λ ^K+1)=log ₂ Per(γ Ω diag (λ ^K+1)) recomputate C _u(λ ^K+1);

F), make k=k+1.If k=K then makes λ=λ ^k, program suspension; Otherwise, enter next procedure;

G), if C _u(λ ^k)-C _u(λ ^K-1)≤ε then makes λ=λ ^k, program suspension; Otherwise, forward step b) to and begin to carry out.

4) utilize step 2) and step 3) the sending direction pre-coding matrix and the power division matrix that calculate carry out power division, and precoding transmission.

The obtaining of transmitting terminal channel statistical information is divided into feedback and two kinds of patterns of latent feedback, when adopting latent feedback model, directly in the statistical information of transmitting terminal calculating channel; When adopting feedback model, transmitting terminal obtains channel statistical information by the feedback of receiving terminal.The channel statistical information that is obtained refers to the feature mode and the signal to noise ratio of the relevant battle array of channel couples matrix/transmission on the feature mode.

Beneficial effect: the invention provides a kind of adaptive multi-antenna transmission method that utilizes channel statistical information, this method has following advantage:

1, this method only needs the statistical information of channel, is applicable to various typical wireless communication systems;

2, the channel model in this method has been considered the line of sight of channel, transmission is relevant, reception is relevant and transmit-receive combination is relevant, more approaches actual channel;

3, the power distribution algorithm fast convergence rate in this method, iteration can restrain for several times, and only an iteration can obtain near optimum separating;

4, the mutual information that this method obtained approaches the channel capacity that the optimum power distribution is obtained.

Embodiment

Method of the present invention mainly may further comprise the steps:

Step 1), when adopting latent feedback model, receiving terminal utilizes the result of channel estimating to calculate signal to noise ratio, and signal to noise ratio is fed back to transmitting terminal, and transmitting terminal calculates it and receives the link channel statistical information, utilize the reciprocity of channel, directly obtain sending the channel statistical information of link; When adopting feedback model, utilize the statistical information of the calculating channel as a result of channel estimating at receiving terminal, and send it to transmitting terminal;

Step 2), calculate sending direction pre-coding matrix U at transmitting terminal according to channel statistical information _Q

Step 3), at the channel statistical information rated output allocation matrix Λ of transmitting terminal according to receiving terminal feedback;

Step 4), the sending direction pre-coding matrix and the power division matrix that utilize preceding two steps to calculate carry out power division and precoding transmission.

Consider that a number of transmit antennas is N _t, the reception antenna number is N _rMimo wireless communication system, on the basis that its channel capacity is analyzed, can construct following transmitting terminal precoding transmission plan by a upper bound of maximum channel capacity:

At receiving terminal: if system adopts feedback model, then to digital baseband receiving signals y (n)=[y ₁(n) y ₂(n) ... y _NR(n)] ^TCarry out channel estimating, wherein y _i(n) received signal of i reception antenna of expression, subscript () ^TThe expression conjugate transpose.Utilize the statistical information of the calculating channel as a result of channel estimating, and channel statistical information is fed back to transmitting terminal.If what adopt is latent feedback model, then calculate signal to noise ratio, signal to noise ratio is fed back to transmitting terminal.

At transmitting terminal:, then at first utilize its channel estimation results that receives link H and the reciprocity H=H of channel if system adopts latent feedback model ^T, calculate the channel statistical information that it sends link H, utilize the channel statistical information that obtains, computer memory power division matrix Λ and sending direction pre-coding matrix U _QIf adopt feedback model, then directly utilize the channel statistical information computer memory power division matrix Λ and the sending direction pre-coding matrix U of receiving terminal feedback _QThen to incoming symbol stream d (n)=[d ₁(n) d ₂(n) ... d _Nt(n)] ^TCarry out linear predictive coding, obtain sending signal s (n)=[s ₁(n) s ₂(n) ... s _Nt(n)] ^T, d wherein _i(n) i incoming symbol stream of expression, s _i(n) the transmission signal of i transmitting antenna of expression.Satisfy following relation between d (n) and the s (n):

s(n)＝Fd(n)， 【1】

Wherein,

F＝U _QΛ ^1/2， 【2】

It is pre-coding matrix.

For making the technical scheme among the present invention clearer, below this programme is specifically described:

One, the acquisition of channel statistical information

Receiving terminal calculates noise variance σ on each antenna of receiving terminal according to the result of channel estimating in the described scheme ², channel sends relevant battle array and receives relevant battle array, and the channel couples matrix Ω on the feature mode.

We use N _r* N _tMatrix H represent channel matrix, channel matrix H is modeled as:

【3】

Wherein

U _tAnd U _rBe respectively N _t* N _tAnd N _r* N _r, fixing unitary matrice, D is N _r* N _tFixing reality " diagonal matrix " (N here _r* N _t" diagonal matrix " refer to (i, j) element is O) of matrix when i ≠ j, M is N _r* N _tFixing real matrix, H _IidBeing one is zero by average, and variance is 1 the independent identically distributed N that answers the gaussian variable composition _r* N _tRandom matrix, represents Hadama product, subscript

Represent conjugate transpose.Utilize the estimated value of channel parameter, calculate respectively and send the relevant battle array of relevant battle array with reception:

E[wherein] expression asks expectation.Next carry out feature decomposition to sending the relevant battle array of relevant battle array respectively with reception

Channel couples matrix on the calculated characteristics pattern then

Wherein

Subscript () ^*The conjugate operation of representing matrix.Calculate signal to noise ratio

γ＝P/σ ²，【9】

Wherein P is total transmitted power.When adopting feedback model, receiving terminal will send relevant battle array R _t, channel couples matrix Ω and signal to noise ratio γ feed back to transmitting terminal; When employing concealed feedback model, receiving terminal fed back to transmitting terminal with signal to noise ratio γ, and transmitting terminal calculates and sends a relevant gust R _tWith channel couples matrix Ω.

Two, sending direction pre-coding matrix

Transmitting terminal carries out feature decomposition to sending relevant battle array

Obtain U _t, the sending direction pre-coding matrix U in the described scheme _QBe chosen as U _Q=U _T

Three, power division matrix

Power division matrix Λ in this programme can be expressed as:

$\mathrm{\Λ}=\frac{P}{N_t}\mathrm{diag}\left(\mathrm{\λ}\right),$ 【10】

Wherein, λ=[λ ₁, λ ₂..., λ _Nt], diag (λ) expression is the diagonal matrix of diagonal entry with the element of vectorial λ, λ ₁, λ ₂..., λ _Nt〉=0.Represent the permanent operator with Per ().For the matrix A of a M * N, define its expansion permanent Per(A) be

Per(A)＝Per([I _M?A])＝Per([I _N?A ^T])【11】

Definition A _(i)For matrix A is left out its remaining matrix in i row back.

A kind of algorithm of iteration water filling is adopted in the acquisition of the power division matrix in this programme, uses ${\mathrm{\λ}}^k=[{\mathrm{\λ}}_1^k,{\mathrm{\λ}}_2^k,...,{\mathrm{\λ}}_{N_t}^k]$ The result who represents the power division of the k time iteration gained, λ _i ^kExpression is with λ ^kI element be made as the vector of 1 gained, C _u(λ ^k) previous term of the capacity of lower channel as a result of power division of the k time iteration gained of expression.The concrete steps of this iteration water-filling algorithm are described below:

Step 1), initialization: set maximum iteration time K and convergence decision threshold ε, make k=0, λ ^k=1 _{1 * Nt}, C _u(λ ^k)=log ₂ Per(λ Ω), wherein 1 _{1 * Nt}Represent that a dimension is 1 * N _t, component is 1 matrix;

Step 2), calculate

$p\left({\mathrm{\λ}}_{\left(i\right)}^k\right)=\underset{\‾}{\mathrm{Per}}\left(\mathrm{\γ}{\mathrm{\Ω}}_{\left(i\right)}\mathrm{diag}\left({\mathrm{\λ}}_{\left(i\right)}^k\right)\right),$ 【12】

$q\left({\mathrm{\λ}}_{\left(i\right)}^k\right)=\underset{\‾}{\mathrm{Per}}\left(\mathrm{\γ\Ωdiag}\left({\mathrm{\λ}}_i^k\right)\right)-\underset{\‾}{\mathrm{Per}}\left(\mathrm{\γ}{\mathrm{\Ω}}_{\left(i\right)}\mathrm{diag}\left({\mathrm{\λ}}_{\left(i\right)}^k\right)\right),$ 【13】

Wherein, i=1,2 ..., N _t

Step 3), calculating

${\mathrm{\λ}}_i^{k+1}=\max (0,\tilde{v}-\frac{p\left({\mathrm{\λ}}_{\left(i\right)}^k\right)}{q\left({\mathrm{\λ}}_{\left(i\right)}^k\right)}),i=\mathrm{1,2},...,N_t,$ 【14】

Wherein

Serve as reasons $\underset{i=1}{\overset{N_t}{\mathrm{\Σ}}}{\mathrm{\λ}}_i^{k+1}=N_t$ The constant of decision;

Step 4), calculating

C _u(λ ^k+1)＝log ₂ Per(γΩdiag(λ ^k+1))；【15】

If step 5) C _u(λ ^K+1)≤C _u(λ ^k), then order ${\mathrm{\λ}}^{k+1}=\frac{1}{N_r}{\mathrm{\λ}}^{k+1}+\frac{N_r-1}{N_r}{\mathrm{\λ}}^k,$ Recomputate C with formula [15] _u(λ ^K+1);

Step 6), make k=k+1.If k=K then makes λ=λ ^k, program suspension; Otherwise, enter next procedure;

Step 7), if C _u(λ ^k)-C _u(λ ^K-1)≤ε then makes λ=λ ^k, program suspension; Otherwise, forward step 2 to) begin to carry out.

The specific embodiment of the invention is as follows:

Receiving terminal:

1) if adopt latent feedback model, then calculate signal to noise ratio γ, γ feeds back to transmitting terminal with signal to noise ratio, and skips to step 6); Otherwise, enter step 2).

2) utilize received signal to carry out channel estimating, calculate signal to noise ratio γ, utilize formula [4] and formula [5] to calculate and send relevant battle array R _tRelevant battle array R with reception _r

3) carry out feature decomposition to sending the relevant battle array of relevant battle array, obtain U with reception _tAnd U _r

4) utilize the U as a result of feature decomposition _t, U _rAnd formula [8] calculating channel coupling matrix Ω.

5) with R _t, Ω and γ feed back to transmitting terminal, enter step 9).

Transmitting terminal:

6) utilize transmitting terminal to receive the channel estimation results of link H and the reciprocity H=H of channel ^T, calculate the relevant battle array of transmission R according to formula [4] and formula [5] _tRelevant battle array R with reception _r

7) carry out feature decomposition to sending the relevant battle array of relevant battle array, obtain U with reception _tAnd U _r

8) utilize the U as a result of feature decomposition _t, U _rAnd formula [8] calculating channel coupling matrix Ω, and enter step 10).

9) carry out feature decomposition to sending relevant battle array, obtain U _t

10) set maximum iteration time K and convergence decision threshold ε, make k=0, λ ^k=1 _{1 * Nt}, C _u(λ ^k)=log ₂ Per(λ Ω), wherein 1 _{1 * Nt}Represent that a dimension is 1 * N _t, component is 1 matrix.

11) utilize formula [12] and formula [13] to calculate p (λ _(i) ^k)) and q (λ _(i) ^k)), i=1,2 ..., N _t

12) utilize formula [14] to calculate λ _i ^K+1, i=1,2 ..., N _t

13) utilize formula [15] to calculate C _u(λ ^K+1).

14) if C _u(λ ^K+1)≤C _u(λ ^k), then order ${\mathrm{\λ}}^{k+1}=\frac{1}{N_r}{\mathrm{\λ}}^{k+1}+\frac{N_r-1}{N_r}{\mathrm{\λ}}^k,$ Recomputate C with formula [15] _u(λ ^K+1); Otherwise, enter next procedure.

15) make k=k+1.If k=K then makes λ=λ ^k, and forward step 17 to); Otherwise, enter next procedure.

16) if C _u(λ ^k)-C _u(λ ^K+1)≤ε then makes λ=λ ^k, and forward step 17 to); Otherwise, forward step 11) to and begin to carry out.

17) utilize formula [10] rated output allocation matrix Λ.

18) make sending direction precoding battle array U _Q=U _t

19) utilize 17) and 18) in the U that calculates _QAnd Λ, calculate the linear predictive coding matrix according to formula [2], send control according to formula [1].

Claims (3)

1. utilize the adaptive transmission method of channel statistical information in the multi-aerial transmission system, it is characterized in that carrying out as follows:

1) obtaining of transmitting terminal channel statistical information: with number of transmit antennas is N _t, the reception antenna number is N _rChannel matrix H be modeled as

Wherein

U _tAnd U _rBe respectively N _t* N _tAnd N _r* N _rFixing unitary matrice, D is N _r* N _tFixing reality " diagonal matrix ", M is N _r* N _tFixing real matrix, H _IidBeing one is zero by average, and variance is 1 the independent identically distributed N that answers the gaussian variable composition _r* N _tRandom matrix, represents Hadama product, subscript

Represent conjugate transpose; When adopting feedback model, receiving terminal utilizes the estimated value of channel parameter, uses

With

Calculate and send the relevant battle array with reception of relevant battle array, wherein E{} represents to ask expectation; Next carry out feature decomposition to sending the relevant battle array of relevant battle array respectively with reception:

Channel couples matrix on the calculated characteristics pattern then:

Wherein

Subscript () ^*The conjugate operation of representing matrix; Calculate signal to noise ratio γ=P/ σ ², wherein P is total transmitted power; At last, receiving terminal will send relevant battle array R _t, channel couples matrix Ω and signal to noise ratio γ feed back to transmitting terminal; When employing concealed feedback model, receiving terminal fed back to transmitting terminal with signal to noise ratio γ, and transmitting terminal utilizes transmitting terminal to receive the channel estimation results of link H and the reciprocity H=H of channel ^T, adopt the method identical to calculate and send relevant battle array R with feedback model _tWith channel couples matrix Ω;

2) transmitting terminal carries out feature decomposition to the relevant battle array of the transmission in the channel statistical information: Obtain sending direction pre-coding matrix U _Q=U _T

3) adopt iteration water-filling algorithm rated output allocation matrix at transmitting terminal according to channel statistical information $\mathrm{\Λ}=\frac{P}{N_t}\mathrm{diag}\left(\mathrm{\λ}\right),$ λ=[λ wherein ₁, λ ₂..., λ _Nt], diag (λ) expression is the diagonal matrix of diagonal entry with the element of vectorial λ, λ ₁, λ ₂..., λ _Nt〉=0; With ${\mathrm{\λ}}^k={[\mathrm{\λ}}_1^k,{\mathrm{\λ}}_2^k,...,{\mathrm{\λ}}_{N_t}^k]$ The result who represents the power division of the k time iteration gained, λ _i ^kExpression is with λ ^kI element be made as the vector of 1 gained, C _u(λ ^k) previous term of the capacity of lower channel as a result of power division of the k time iteration gained of expression, A _(i)For matrix A is left out its remaining matrix in i row back;

4) utilize step 2) and step 3) the sending direction pre-coding matrix and the power division matrix that calculate carry out power division, and precoding transmission.

2. utilize the adaptive transmission method of channel statistical information in the multi-aerial transmission system according to claim 1, it is characterized in that: obtaining of transmitting terminal channel statistical information is divided into feedback and two kinds of patterns of latent feedback, when employing conceals feedback model, direct statistical information at the transmitting terminal calculating channel; When adopting feedback model, transmitting terminal obtains channel statistical information by the feedback of receiving terminal.

3. utilize the adaptive transmission method of channel statistical information in the multi-aerial transmission system according to claim 1, it is characterized in that the channel statistical information that is obtained refers to the feature mode and the signal to noise ratio of the relevant battle array of channel couples matrix/transmission on the feature mode.