CN103747515B - It is applicable to the interference alignment schemes of the joint Power distribution of WLAN - Google Patents

Abstract

The invention discloses the interference alignment schemes of a kind of joint Power distribution being applicable to WLAN, the program first step eliminates presence of intercell interference by a part for receiving matrix, equivalent channel after second step user will eliminate presence of intercell interference feeds back to transmitting terminal, transmitting terminal design pre-coding matrix eliminates inter-user interference, 3rd step eliminates the interference between same user stream by singular value decomposition, and each stream is carried out power distribution, receiving matrix is left a part is sent to user simultaneously.Mutual substantial amounts of information is need not between base station of the present invention or between user, each user can confirm orthogonal kernel in the case of pre-coding matrix is launched in the unknown from inter-cell interference signals space, thus design receiving terminal matrix, therefore have only to the feedback of Intra-cell, existing system need not be carried out the biggest change.

Description

It is applicable to the interference alignment schemes of the joint Power distribution of WLAN

Technical field

The present invention relates to wireless communication technology field, be specifically related to a kind of joint Power distribution being applicable to WLAN Interference alignment schemes.

Background technology

Interference alignment techniques is coordinates user, effectively changes the one of signal structure by time slot, area block or antenna Plant transmission strategy.It carries out precoding processing at transmitting terminal to signal so that interference signal is overlapped, and useful signal can be Receiving terminal is separated.Along with sending out of multi-user (Multiple Input Multiple Output, MIMO) communication technology Exhibition, disturbs alignment techniques to become and receives more and more attention.

Have much round the interference alignment research work in theory and practice, but disturbed alignment scheme at these In, it is required for global channel status information to design encoding and decoding.For each node, it is thus achieved that global channel status information is born Carrying on a shoulder pole the heaviest, this ratio in actual wireless LAN system is relatively difficult to achieve.Also document is had to propose alternating iteration method to obtain interference The pre-coding matrix of alignment and decoding matrix, however it is necessary that the iterations close to 200 times obtains preferably coder, also That is transmitter and receiver is required to iteration 200 times.Obviously such iterative scheme is put into ultrahigh speed of future generation wireless LAN can increase the weight of time overhead.

Mainly need to consider following feature, as made in wireless local network interference alignment algorithm of future generation With local channel state information, distributed treatment, linear predictive coding etc., the scheme of interference alignment at present is a lot, but major part ratio More complicated and unfavorable with realizing in systems in practice, so that the codec of design suitably less expense.

Summary of the invention

Goal of the invention: for the feature of WLAN, the present invention devises the ZF interference of a kind of joint Power distribution Alignment schemes, the method need not cooperate between WLAN base station and interactive information between user.

Technical scheme: the interference alignment schemes of a kind of joint Power distribution being applicable to WLAN, including walking as follows Rapid:

(1) user estimates the interference channel of neighbor cell according to pilot tone, and the Part I U of design receiving matrix eliminates The interference of minizone, and feed back to base station；

(2) base station information by user feedback, design is launched the Part I V of pre-coding matrix and is eliminated between user Interference；

(3) multiuser channel is converted into single user channel on the basis of (1), (2) by base station, and is divided by singular value Solve and obtain matrix A and B for eliminating the interference between signal stream；

(4) base station sends signal with the final precoding that sends to user, matrix B is sent to user simultaneously；

(5) user receives useful signal with final receiving matrix.

In described step (3), base station the most also carries out power distribution on equivalence single user channel basis.Due to through step Suddenly (1) eliminates presence of intercell interference, and each community can independently further design power distribution method be improved entirety and fast Rate.

The Part I U of described step (1) receiving matrix is by formula (1.1), (1.2):

$H_{[m,k]}^{\stackrel{\‾}{m}}Q=\left[{\stackrel{\‾}{U}}_{[m,k]}^{\left(1\right)}{\stackrel{\‾}{U}}_{[m,k]}^{\left(0\right)}\right]{\stackrel{\‾}{\mathrm{\Λ}}}_{[m,k]}^{\left(1\right)}{\left({\stackrel{\‾}{V}}_{[m,k]}\right)}^H---\left(1.1\right)$

$U_{[m,k]}=V_{L_s}\left({\stackrel{\‾}{U}}_{[m,k]}^{\left(0\right)}\right)---\left(1.2\right)$

Calculate.

The Part I V of described step (2) pre-coding matrix is by formula (2.1), (2.2):

${\stackrel{\‾}{H}}_{[m,k]}^m{\stackrel{\‾}{U}}_{[m,k]}{\stackrel{\‾}{\mathrm{\Λ}}}_{[m,k]}^{\left(2\right)}{({\stackrel{\‾}{V}}_{[m,k]}^{\left(1\right)},{\stackrel{\‾}{V}}_{[m,k]}^{\left(0\right)})}^H---\left(2.1\right)$

$V_{[m,k]}=V_{L_s}\left({\stackrel{\‾}{V}}_{[m,k]}^{\left(0\right)}\right)---\left(2.2\right)$

Calculate.

Step (3) is realized by following steps: base station according to:Calculate user etc. Effect channel H_SU[m,k], by singular value decomposition:Formula calculates pre-coding matrix A_[m,k]And reception Matrix B_[m,k]。

The distribution realization of described power is to calculate user power distribution factor by water flood.

It is F that described step (4) sends precoding_[m,k]=Q_[m,k]V_[m,k]A_[m,k](p_[m,k])^1/2。

The receiving matrix of described step (5) user is W_[m,k]=U_[m,k]B_[m,k]。

Beneficial effect: need not mutual substantial amounts of information in the present invention between base station or between user, each user can be Confirm orthogonal kernel in the case of unknown transmitting pre-coding matrix from inter-cell interference signals space, thus design receiving terminal Matrix, therefore has only to the feedback of Intra-cell, and existing system need not be carried out the biggest change.Joint Power of the present invention is divided The interference alignment schemes simulation result joined as it is shown in figure 1, in Fig. 1 " Proposed IA Scheme " to represent that the present invention proposes suitable Close the interference alignment schemes of the joint Power distribution of WLAN." GIA " represents interference alignment schemes in the present invention, but does not adds Enter power distribution." Matched Filtering " represents traditional matched filtering device, but for the sake of justice, only with once Iteration.Can be seen that from simulation result the performance of the present invention is more preferable.

Accompanying drawing explanation

Fig. 1 is the interference alignment schemes simulation result figure of joint Power of the present invention distribution；

Fig. 2 is the interference alignment schemes flow chart of joint Power of the present invention distribution.

Detailed description of the invention

Below in conjunction with specific embodiment, it is further elucidated with the present invention, it should be understood that these embodiments are merely to illustrate the present invention Rather than restriction the scope of the present invention, after having read the present invention, the those skilled in the art's various equivalences to the present invention The amendment of form all falls within the application claims limited range.

Considering that implementing scene is that a hot spot region exists G the community interfered, each area of coverage AP has M root sky Line, supports K user, each user's N root antenna.Base station M root antenna sends L to each K user_sIndividual data stream.For simplifying Analyze, first consider the situation of a carrier wave, other be similar to popularization.

The reception signal of the k user of m community can be expressed as:

$y_{[m,k]}=H_{[m,k]}^m{F}_{[m,k]}{s}_{[m,k]}+H_{[m,k]}^m\underset{i\≠k}{\overset{K}{\mathrm{\Σ}}}{F}_{[m,i]}{s}_{[m,i]}+\underset{g\≠m}{\overset{G}{\mathrm{\Σ}}}\underset{i=1}{\overset{K}{\mathrm{\Σ}}}{H}_{[m,k]}^g{F}_{[g,i]}{s}_{[g,i]}+n_{[m,k]}---\left(1\right)$

HereRepresent m-th base station AP_mPre-coding matrix to its kth user STA [m, k],Represent the signal phasor being transferred to STA [m, k],Represent the Gauss of user STA [m, k] Noise,Represent the g base station channel matrix to m community k user STA [m, k].Here false channel is obeyed multiple Gauss independent same distribution.AP_mTo user STA_{[m, k]}Transmitting pre-coding matrix F_[m,k]MeetThis In P_mRepresent AP_mMaximum transmission power.Above formula the 1st represents AP_mIt is sent to the effective information of user STA [m, k], the 2nd table The distracter of other user to users STA [m, k] in the m of Shi Shi community, the 3rd represents that other communities are to user STA [m, k] shape The distracter become.

The receiving matrix making user STA [m, k] isThen its receiving end signal estimated value is

${\hat{s}}_{[m,k]}=W_{[m,k]}^H{y}_{[m,k]}---\left(2\right)$

I.e.

${\hat{s}}_{[m,k]}=W_{[m,k]}^H{H}_{[m,k]}^m{F}_{[m,k]}{s}_{[m,k]}+W_{[m,k]}^H{H}_{[m,k]}^m\underset{i\≠k}{\overset{K}{\mathrm{\Σ}}}{F}_{[m,i]}{s}_{[m,i]}+$

$W_{[m,k]}^H\underset{g\≠m}{\overset{G}{\mathrm{\Σ}}}\underset{i=1}{\overset{K}{\mathrm{\Σ}}}{H}_{[m,k]}^g{F}_{[g,i]}{s}_{[g,i]}+W_{[m,k]}^H{n}_{[m,k]}---\left(3\right)$

Can be seen that by formula (3) signal that receives of user STA [m, k] includes in the interference from other communities, community it Interference between interference that he user causes and stream, it is therefore desirable to design transmitting terminal pre-coding matrix F_{[m, k]}With receiving terminal matrix W_{[m, k]}Eliminate these interference.As in figure 2 it is shown, the scheme that the present invention proposes is: the first step is first come by a part for receiving matrix Eliminating presence of intercell interference, second step eliminates inter-user interference by the part launching pre-coding matrix, through this two stage place After reason system converting for Single User MIMO equivalent channel model, the 3rd step carries out the methods such as Eigenvalues Decomposition again to eliminate between stream Interference.Simultaneously because eliminate presence of intercell interference through the first step, each community can independent further design power distribution side Method improve entirety and speed

Order is launched pre-coding matrix and is given as

F_[m,k]=Q_[m,k]V_[m,k]A_[m,k](p_[m,k])^1/2 (4)

Receiving matrix is

W_[m,k]=U_[m,k]B_[m,k] (5)

Wherein matrixBe introduced for M(M > KL_s) root antenna extension KL_sIndividual data stream is many when considering The when of individual stream, extended matrix Q needs correct selection so thatKernel exist, arbitrary (K+1) L_s× KL_sNon-singular matrix (each row are mutually orthogonal), asJust requirement is met.It is to launch precoding A part, user eliminates the interference of multi-user in same community,For a part for receiving matrix, it is used for disappearing Except presence of intercell interference.Matrix A_[m,k]、B_[m,k]By V_{[m, k]}、U_{[m, k]}WithDetermine, be used for eliminating user STA_[m,k]L_sIndividual stream Between interference.p_{[m, k]}Power allocation factor for user STA [m, k].

The first step: design receiving matrix Part I U_{[m, k]}Eliminate presence of intercell interference

OrderRepresent that user STA [m, k] is by the interference channel from other cell base stations.In order to disappear Except presence of intercell interference, the forward part of receiving matrixBe fromKernel in choose, i.e. need to meet

$U_{[m,k]}^H\left(H_{[m,k]}^{\stackrel{\‾}{m}}Q\right)=0---\left(6\right)$

To matrixCarry out singular value decomposition to obtain

$H_{[m,k]}^{\stackrel{\‾}{m}}Q=\left[{\stackrel{\‾}{U}}_{[m,k]}^{\left(1\right)}{\stackrel{\‾}{U}}_{[m,k]}^{\left(0\right)}\right]{\stackrel{\‾}{\mathrm{\Λ}}}_{[m,k]}^{\left(1\right)}{\left({\stackrel{\‾}{V}}_{[m,k]}\right)}^H---\left(7\right)$

WhereinD before representing_kIndividual left singular vector, d_kRepresentOrder,Represent additionally (N-d_k) an individual left side Singular vector.According to matrix knowledge,It isOne group of orthogonal basis of kernel, i.e. Therefore fromIn choose column vector can be by from other cell base stations as the part of user STA [m, k] receiving matrix Interference to the user of community m eliminates so that (3) Section 3 of formula is 0.I.e. meet the U of condition_[m,k]For

$U_{[m,k]}=V_{L_s}\left({\stackrel{\‾}{U}}_{[m,k]}^{\left(0\right)}\right)---\left(8\right)$

Here V_Ls(X) expression takes the L of X_sIndividual column vector.

Second step: design pre-coding matrix V_[m,k]Eliminate inter-user interference

After eliminating presence of intercell interference, in can eliminating community further with block diagonalization precoding algorithms, other are used The interference that user STA [m, k] is formed by family, i.e. needs to meet

$\left({\left(U_{[m,i]}\right)}^H{H}_{[m,i]}^{m}Q\right)V_{[m,k]}=0$ (i=1,2 ... K and i ≠ k) (9)

U is passed through in order_[m,k]User UE in the m of community after process_m,kEquivalent channel be

$G_{[m,k]}^m={\left(U_{[m,i]}\right)}^H{H}_{[m,k]}^{m}Q---\left(10\right)$

Then except user STA in the m of community_[m,k]The equivalent channel of all users in additionCan be expressed as

${\stackrel{\‾}{H}}_{[m,k]}^m={[{\left(G_{[m,1]}^m\right)}^T,...,{\left(G_{[m,k-1]}^m\right)}^T,{\left(G_{[m,k+1]}^m\right)}^T,...,{\left(G_{[m,K]}^m\right)}^T]}^T---\left(11\right)$

So, formula (9) can be write as in equivalence:

${\stackrel{\‾}{H}}_{[m,k]}^m{V}_{[m,k]}=0---\left(12\right)$

So base station AP_mBlock diagonalization can be utilized to eliminate inter-user interference.Formula (12) is dry for being completely eliminated between user To user STA during the block diagonalization algorithm disturbed_[m,k]Pre-coding matrix V_[m,k]Requirement, namely require that each user's is pre- Encoder matrix V_[m,k]Must be positioned at the channel matrix of other all usersKernel in.After this precoding Equivalence multiuser channel matrix H is a Block diagonal matrix, and after block diagonalization precoding, it is individual independent that system just can be converted into K Single User MIMO system, the independent Single User MIMO system for each equivalence can apply traditional Single User MIMO system The transmission plan of system and decoding scheme.

RightCarry out singular value decomposition to obtain

${\stackrel{\‾}{H}}_{[m,k]}^m={\stackrel{\‾}{U}}_{[m,k]}{\stackrel{\‾}{\mathrm{\Λ}}}_{[m,k]}^{\left(2\right)}{({\stackrel{\‾}{V}}_{[m,k]}^{\left(1\right)},{\stackrel{\‾}{V}}_{[m,k]}^{\left(0\right)})}^H---\left(13\right)$

WhereinL before representing_kIndividual right singular vector, l_kRepresentOrder.Represent additionally (KL_s-l_k) the individual right side Singular vector.It isOne group of orthogonal basis of kernel.Therefore useAs user STA [m, k] precoding square The downlink of remaining multi-user MIMO system can be resolved into multiple independent Single User MIMO channel by a part for battle array. I.e. have

$V_{[m,k]}=V_{L_s}\left({\stackrel{\‾}{V}}_{[m,k]}^{\left(0\right)}\right)---\left(14\right)$

For user STA_[m,k], through U_[m,k]The interference from other communities is eliminated, through V after matrix_{[m, k]}Matrix disappears Except remaining from the interference of other users in the m of community, the Single User MIMO channel matrix of its equivalence is

$H_{\mathrm{SU}[m,k]}=U_{[m,k]}^H{H}_{[m,k]}^m{\mathrm{QV}}_{[m,k]}---\left(15\right)$

3rd step: design matrix A_[m,k]、B_[m,k]Eliminate the matrix design of inter-stream interference

By design transmitting terminal beamforming matrix F_[m,k]Another part A_[m,k]With receiving matrix W_[m,k]Part II B_[m,k] Eliminate the L of user STA [m, k]_sInterference between individual stream.A_[m,k]、B_[m,k]By V_[m,k]、U_[m,k]、Q_[m,k]WithDetermine.Due to Have passed through V_[m,k]And U_[m,k]Eliminate inter-user interference and presence of intercell interference after process respectively, then receiving terminal estimates that signal is:

${\hat{s}}_{[m,k]}{B}_{[m,k]}^H{H}_{\mathrm{SU}[m,k]}{A}_{[m,k]}{s}_{[m,k]}+z_{[m,k]}---\left(16\right)$

Wherein z_[m,k]=B_[m,k]U_[m,k]n_[m,k], and

Equivalence Single User MIMO channel is carried out singular value decomposition:

$H_{\mathrm{SU}[m,k]}=B_{[m,k]}{\mathrm{\Λ}}_{[m,k]}{A}_{[m,k]}^H---\left(17\right)$

Wherein B_[m,k]、A_[m,k]For unitary matrice；Λ_[m,k]The diagonal matrix being made up of singular value, its diagonal values is non-negative Value and descending.

So through above three steps, just being in OBSS(Overlapping BSS) minizone of zone user, between user Interference and inter-stream interference in turn, eliminate.

Further, due to the process by the first step and second step, eliminate user STA_[m,k]Experienced interference, then its speed Rate can be expressed as

$R_{[m,k]}=\underset{s=1}{\overset{L_s}{\mathrm{\Σ}}}{\log }_2(1+\frac{{\left({\mathrm{\Λ}}_{[m,k,s]}\right)}^2{p}_{[m,k,s]}}{{\mathrm{\σ}}_z^2})---\left(18\right)$

Here p_[m,k,s]Represent the power of the upper distribution of user STA [m, k] the s stream；Λ_[m,k,s]For matrix Λ_[m,k]Diagonal angle The s element on line, represents the eigenvalue on equivalence single user channel the s stream,

Because use ZF interference alignment after there is no presence of intercell interference, then m community and speed and other communities Power distribution is relatively independent.So, and the maximized power distribution problems of speed just resolve into single community in and fast Rate maximization problems, i.e. by optimizing power allocation factor { p_[m,k,s]Maximize in the m of community and speed:

$\underset{p_{[m,\mathrm{1,1}]},p_{[m,\mathrm{1,1}]},...,p_{[m,K,L_s]}}{\max }\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{s=1}{\overset{L_s}{\mathrm{\Σ}}}{\log }_2(1+\frac{{\left({\mathrm{\Λ}}_{[m,k,s]}\right)}^2{p}_{[m,k,s]}}{{\mathrm{\σ}}_z^2})---(19)$

Wherein constraints is

$\left\{\begin{array}{c}{p}_{[m,k,s]}\≥0,k=1,...,K;s=1,...,L_s\\ \underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{s=1}{\overset{L_s}{\mathrm{\Σ}}}{p}_{[m,k,s]}\≤P_m\end{array}\right.---\left(20\right)$

Here P_mFor base station AP_mTransmitting power.

Example of the present invention uses water injection power distribution method to try to achieve optimal power allocation.The object function of formula (19) be about The associating concave function of power, this optimization problem can use Lagrangian method to solve.Consideration Lagrangian:

$L(\mathrm{\λ},p_{[m,\mathrm{1,1}]},p_{[m,\mathrm{1,2}]},...,p_{[m,K,L_s]}):=\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{s=1}{\overset{L_s}{\mathrm{\Σ}}}{\log }_2(1+\frac{{\left({\mathrm{\Λ}}_{[m,k,s]}\right)}^2{p}_{[m,k,s]}}{{\mathrm{\σ}}_z^2})-\mathrm{\λ}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{s=1}{\overset{L_s}{\mathrm{\Σ}}}{p}_{[m,k,s]}---\left(21\right)$

Wherein λ is Lagrange multiplier.Kuhn tucker condition (the Karush-Kuhn-Tucker of optimal power allocation Conditions, KKT) be

$\frac{\∂L}{{\∂p}_{[m,k,s]}}\left\{\begin{array}{ccc}=0& \mathrm{If}& p_{[m,k,s]}>0\\ \≤0& \mathrm{If}& p_{[m,k,s]}=0\end{array}\right.---\left(22\right)$

Definition x⁺:=max (x, 0), then power distribution

$p_{[m,k,s]}^{*}={(\frac{1}{\mathrm{\λ}}-\frac{{\mathrm{\σ}}_z^2}{{\left({\mathrm{\Λ}}_{[m,k,s]}\right)}^2})}^{+}---\left(23\right)$

Meeting the condition of formula (20), Lagrange multiplier λ being therefore optimum, selected makes power constraint be expired Foot:

$\underset{k=1}{\overset{K}{\mathrm{\Σ}}}\underset{s=1}{\overset{L_s}{\mathrm{\Σ}}}{(\frac{1}{\mathrm{\λ}}-\frac{{\mathrm{\σ}}_z^2}{{\left({\mathrm{\Λ}}_{[m,k,s]}\right)}^2})}^{+}=P_m---\left(24\right)$

The transmitting pre-coding matrix finally combining power distribution and interference alignment can be

F_[m,k]=Q_[m,k]V_[m,k]A_[m,k](p_[m,k])^1/2 (25)

Wherein diagonal matrix P_[m,k]For the power allocation matrix of user STA [m, k], its diagonal entry is (p_[m,k,1],…, p_[m,k,Ls])。

4th step: transmitting precoding F [m, k] of user STA [m, k] is by base station

F_[m,k]=Q_[m,k]V_[m,k]A_[m,k](p_[m,k])^1/2

5th step: B [m, k] is sent to user STA [m, k] by base station, and the receiving matrix of user STA [m, k] is

W_[m,k]=U_[m,k]B_[m,k] 。

Claims (7)

1. the interference alignment schemes of the joint Power distribution being applicable to WLAN, it is characterised in that include walking as follows Rapid:

(1) user estimates the interference channel of neighbor cell according to pilot tone, designs receiving matrix W_[m,k]=U_[m,k]B_[m,k](m represents Community, k represents user) Part I U to eliminate the interference of minizone, and feed back to base station；Design receiving matrix first Point U eliminate presence of intercell interference particularly as follows:

OrderRepresent that user STA [m, k] is by the interference channel from other cell base stations；Little in order to eliminate Interval interference, the forward part of receiving matrixBe fromKernel in choose, i.e. need to meet

$U_{\[m,k\]}^H\left(H_{\[m,k\]}^{\stackrel{\‾}{m}}Q\right)=0$

To matrixCarry out singular value decomposition to obtain

$H_{\[m,k\]}^{\stackrel{\‾}{m}}Q=\[\begin{array}{cc}{\stackrel{\‾}{U}}_{\[m,k\]}^{\left(1\right)}& {\stackrel{\‾}{U}}_{\[m,k\]}^{\left(0\right)}\end{array}\]{\stackrel{\‾}{\mathrm{\Λ}}}_{\[m,k\]}^{\left(1\right)}{\left({\stackrel{\‾}{V}}_{\[m,k\]}\right)}^H$

WhereinD before representing_kIndividual left singular vector, d_kRepresentOrder,Represent additionally (N-d_k) an individual left side is unusual Vector.According to matrix knowledge,It isOne group of orthogonal basis of kernel, i.e.Therefore fromIn choose column vector as user STA [m, k] receiving matrix a part can by from other cell base stations to community The interference of the user of m eliminates so thatThe Section 3 of formula It is 0；I.e. meet the U of condition_[m,k]For

$U_{\[m,k\]}=V_{L_s}\left({\stackrel{\‾}{U}}_{\[m,k\]}^{\left(0\right)}\right)$

Here V_Ls(X) expression takes the L of X_sIndividual column vector, L_sRepresent data stream；

(2) base station information by user feedback, design is launched the Part I V of pre-coding matrix and is eliminated doing between user Disturb, particularly as follows:

Eliminating after presence of intercell interference, other users couple in community can eliminated further with block diagonalization precoding algorithms The interference that user STA [m, k] is formed, i.e. needs to meet

(i=1,2 ... K and i ≠ k)

U is passed through in order_[m,k]User UE in the m of community after process_m,kEquivalent channel be

$G_{\[m,k\]}^m={\left(U_{\[m,i\]}\right)}^H{H}_{\[m,k\]}^{m}Q$

Then except user STA in the m of community_[m,k]The equivalent channel of all users in additionCan be expressed as

${\stackrel{\‾}{H}}_{\[m,k\]}^m={\[{\left(G_{\[m,1\]}^m\right)}^T,...,{\left(G_{\[m,k-1\]}^m\right)}^T,{\left(G_{\[m,k+1\]}^m\right)}^T,...,{\left(G_{\[m,K\]}^m\right)}^T\]}^T$

So,(i=1,2 ... K and i ≠ k) can equivalence be write as:

${\stackrel{\‾}{H}}_{\[m,k\]}^m{V}_{\[m,k\]}=0$

So base station AP_mBlock diagonalization can be utilized to eliminate inter-user interference；Above formula is the block pair that inter-user interference is completely eliminated To user STA during keratinization algorithm_[m,k]Pre-coding matrix V_[m,k]Requirement, namely require the pre-coding matrix of each user V_[m,k]Must be positioned at the channel matrix of other all usersKernel in.Equivalence after this precoding is multiplex Family channel matrix H is a Block diagonal matrix, and after block diagonalization precoding, system just can be converted into K independent single user Mimo system, can apply the biography of traditional Single User MIMO system for the independent Single User MIMO system of each equivalence Transmission scheme and decoding scheme；

RightCarry out singular value decomposition to obtain

${\stackrel{\‾}{H}}_{\[m,k\]}^m={\stackrel{\‾}{U}}_{\[m,k\]}{\stackrel{\‾}{\mathrm{\Λ}}}_{\[m,k\]}^{\left(2\right)}{({\stackrel{\‾}{V}}_{\[m,k\]}^{\left(1\right)},{\stackrel{\‾}{V}}_{\[m,k\]}^{\left(0\right)})}^H$

WhereinL before representing_kIndividual right singular vector, l_kRepresentOrder；Represent additionally (KL_s-l_k) the individual right side is unusual Vector；It isOne group of orthogonal basis of kernel；Therefore useAs user STA [m, k] pre-coding matrix The downlink of remaining multi-user MIMO system can be resolved into multiple independent Single User MIMO channel by a part；I.e. have

$V_{\[m,k\]}=V_{L_s}\left({\stackrel{\‾}{V}}_{\[m,k\]}^{\left(0\right)}\right)$

For user STA_[m,k], through U_[m,k]The interference from other communities is eliminated, through V after matrix_[m,k]Matrix eliminates Remaining from the interference of other users in the m of community, the Single User MIMO channel matrix of its equivalence is

$H_{SU\[m,k\]}=U_{\[m,k\]}^H{H}_{\[m,k\]}^m{\mathrm{QV}}_{\[m,k\]};$

(3) multiuser channel is converted into single user channel on the basis of (1), (2) by base station, and is obtained by singular value decomposition Matrix A and B are for eliminating the interference between signal stream；

(4) base station sends signal with the final precoding that sends to user, matrix B is sent to user simultaneously；

(5) user receives useful signal with final receiving matrix, and the receiving matrix of user is W_[m,k]=U_[m,k]B_[m,k]。

The interference alignment schemes of the joint Power distribution being applicable to WLAN the most according to claim 1, its feature Being, in described step (3), base station the most also carries out power distribution on equivalence single user channel basis.

The interference alignment schemes of the joint Power distribution being applicable to WLAN the most according to claim 1, its feature Being, the Part I U of described step (1) receiving matrix is by formula (1.1), (1.2):

$H_{\[m,k\]}^{\stackrel{\‾}{m}}Q=\[\begin{array}{cc}{\stackrel{\‾}{U}}_{\[m,k\]}^{\left(1\right)}& {\stackrel{\‾}{U}}_{\[m,k\]}^{\left(0\right)}\end{array}\]{\stackrel{\‾}{\mathrm{\Λ}}}_{\[m,k\]}^{\left(1\right)}{\left({\stackrel{\‾}{V}}_{\[m,k\]}\right)}^H---\left(1.1\right)$

$U_{\[m,k\]}=V_{L_s}\left({\stackrel{\‾}{U}}_{\[m,k\]}^{\left(0\right)}\right)---\left(1.2\right)$

Calculate.

The interference alignment schemes of the joint Power distribution being applicable to WLAN the most according to claim 1, its feature Being, the Part I V of described step (2) pre-coding matrix is by formula (2.1), (2.2):

${\stackrel{\‾}{H}}_{\[m,k\]}^m={\stackrel{\‾}{U}}_{\[m,k\]}{\stackrel{\‾}{\mathrm{\Λ}}}_{\[m,k\]}^{\left(2\right)}{({\stackrel{\‾}{V}}_{\[m,k\]}^{\left(1\right)},{\stackrel{\‾}{V}}_{\[m,k\]}^{\left(0\right)})}^H---\left(2.1\right)$

$V_{\[m,k\]}=V_{L_s}\left({\stackrel{\‾}{V}}_{\[m,k\]}^{\left(0\right)}\right)---\left(2.2\right)$

Calculate.

The interference alignment schemes of the joint Power distribution being applicable to WLAN the most according to claim 1, its feature Being, step (3) is realized by following steps: base station according to:Calculate the equivalence letter of user Road H_SU[m,k], by singular value decomposition:Formula calculates pre-coding matrix A_[m,k]And receiving matrix B_[m,k]。

The interference alignment schemes of the joint Power distribution being applicable to WLAN the most according to claim 2, its feature Being, the distribution realization of described power is to calculate user power distribution factor by water flood.

The interference alignment schemes of the joint Power distribution being applicable to WLAN the most according to claim 1, its feature Being, it is F that described step (4) sends precoding_[m,k]=Q_[m,k]V_[m,k]A_[m,k](p_[m,k])^1/2。