CN102510324B - Signal transmission method based on network coding in multi-input and multi-output Y channel - Google Patents

Abstract

The invention discloses a signal transmission method based on network coding in a multi-input and multi-output Y channel, and the method is mainly used for solving the problems of accessibility and smaller rate of an existing technology. The signal transmission method is implemented through the following steps that: a user node pre-codes original signals to align a signal space with a matrix and is attached with a pre-coding matrix, and a power distribution matrix is multiplied by an original signal vector; the user node sends signals obtained by pre-coding to a relay node; the relay node separates the received signals, and the separated signal is multiplied by an interference-eliminating matrix, the pre-coding matrix attached to the relay node and the power distribution matrix so as to get sending signals of the relay node; the relay node broadcasts the sending signals obtained by pre-coding to the user node; and the user node recovers data according to the received signals and the signals sent by the user node. The simulation result shows that compared with the existing signal transmission method, the signal transmission method disclosed by the invention can be used for greatly improving the accessibility and the rate.

Description

The method for transmitting signals of coding Network Based in multiple-input and multiple-output Y channel

Technical field

The invention belongs to wireless communication technology field, relate to precoding and network code, specifically a method for transmitting signals for coding Network Based, can be used in multiple-input and multiple-output Y-channel, to guarantee improving reaching and speed of system on the basis that the maximum degree of freedom can reach.

Background technology

Traditional bidirectional relay channel is that a pair of user or multipair user pass through the mutual photos and sending messages of relaying, and under this model, each user sends information only to one of them user, receives the information sending from this user simultaneously.In actual applications, each user sends information often need to a plurality of users, receives the information from different user simultaneously.In this case, occurred multidirectional trunk channel model, multiple-input and multiple-output Y-channel is exactly a kind of important communication pattern wherein.

Multiple-input and multiple-output Y-channel model as shown in Figure 1.It comprises three user nodes and a via node, and wherein each user node is equipped M root antenna, via node equipment N root antenna.Between different users, there is no direct link, and each user node respectively sends a unicast info to two other user node by via node, concrete communication process is as follows: first, three user nodes send data to via node simultaneously, and this stage is called the MAC stage; Then, via node is broadcast to three user nodes after the information receiving is processed, and this stage is called the BC stage; Finally, by three user nodes, according to the reception signal of oneself and the own information sending in the MAC stage, translate the information that two other user node is issued oneself respectively.

Korea S scholar Namyoon Lee etc. analyzes the degree of freedom of multiple-input and multiple-output Y-channel in " Degrees of Freedom of the MIMO Y Channel:Signal Space Alignment for Network coding; IEEE Trans.Inform.Theory " at article, article, by adopting signal space alignment techniques in the MAC stage and adopting the interference cancellation beam forming technique of coding Network Based in the BC stage, has proved and has worked as time, the maximum degree of freedom of multiple-input and multiple-output Y-channel under decoding forward mode is 3M, each user can send M independently data flow, represent to be more than or equal to the smallest positive integral of 3M/2.

Although provided the maximum degree of freedom of multiple-input and multiple-output Y-channel in article, how to optimize its Precoding Design, thereby be still and need a problem solving with speed what guarantee to improve on the basis that the maximum degree of freedom can reach system.

Summary of the invention:

The object of the invention is to the deficiency for above-mentioned prior art, according to the multiple-input and multiple-output Y-channel model under amplification forwarding pattern, the method for transmitting signals of coding Network Based in a kind of multiple-input and multiple-output Y-channel is proposed, with guarantee to improve on the basis that the maximum degree of freedom can reach system and speed.

Realize object technical scheme of the present invention, comprise the steps:

(1) user node carries out precoding step to primary signal

Each user node carries out precoding by the information of issuing two other user node, and the M * 1 dimension transmitted signal that obtains sending user node is: $x_i=\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}{V}^{[j,i]}{W}_{\mathrm{\π}(j,i)}{\mathrm{\Σ}}^{[j,i]}{s}^{[j,i]},$

In formula, i represents to send user node, and j represents to receive user node, and i, and { M represents to send the antenna number of user node place preparation to j ∈ for 1,2,3}, i ≠ j, and M is even number, s ^{[j, i]}be to send the primary signal vector that user node i issues a M/2 * 1 data flow formation that receives user node j, each element of this vector is that average is mutually independent random variables zero, that variance is 1, ∑ ^{[j, i]}the sending node power division diagonal matrix of M/2 * M/2, W _{π (j, i)}be the additional pre-coding matrix of M/2 * M/2 user node, subscript π (j, i) is an index function, meets π (i, j)=π (j, i), and has π (1,2)=1, π (1,3)=2, and π (2,3)=3, therefore has W _{π (2,1)}=W _{π (1,2)}=W ₁, W _{π (3,1)}=W _{π (1,3)}=W ₂, W _{π (3,2)}=W _{π (2,3)}=W ₃, V ^{[j, i]}be the signal space alignment matrix of M * M/2, meet H ^{[r, i]}v ^{[j, i]}=H ^{[r, j]}v ^{[i, j]}, H ^{[r, i]}represent that user node i is to the channel matrix of N * M of via node r, N represents the antenna number of via node r place configuration, and has N=3M/2, and each element of this channel matrix is separate multiple Gaussian random variable zero, that variance is 1 for obeying average;

(2) user node is to via node transmitted signal step

Send user node i by the transmitted signal x that carries out precoding and obtain _isend to via node r, N * 1 dimension that obtains via node receives signal and is:

$y_r=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{H}^{[r,i]}{x}_i+n_r$

$=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{H}^{[r,i]}\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}{V}^{[j,i]}{W}_{\mathrm{\π}(j,i)}{\mathrm{\Σ}}^{[j,i]}{s}^{[j,i]}+n_r$

$=U_1{W}_1{s}^{[r,1]}+U_2{W}_2{s}^{[r,2]}+U_3{W}_3{s}^{[r,3]}+n_r,$

In formula, s ^{[r, 1]}=∑ ^[2,1]s ^[2,1]+ ∑ ^[1,2]s ^[1,2]the aligned signal that represents user node 1 and user node 2, s ^{[r, 2]}=∑ ^[3,1]s ^[3,1]+ ∑ ^[1,3]s ^[1,3]the aligned signal that represents user node 1 and user node 3, s ^{[r, 3]}=∑ ^[3,2]s ^[3,2]+ ∑ ^[2,3]s ^[2,3]the aligned signal that represents user node 2 and user node 3, U ₁=H ^{[r, 1]}v ^[2,1]=H ^{[r, 2]}v ^[1,2]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 2]}the friendship space of column space, U ₂=H ^{[r, 1]}v ^[3,1]=H ^{[r, 3]}v ^[1,3]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 3]}the friendship space of column space, U ₃=H ^{[r, 2]}v ^[3,2]=H ^{[r, 3]}v ^[2,3]represent channel matrix H ^{[r, 2]}with channel matrix H ^{[r, 3]}the friendship space of column space, n _rbe the noise vector of N * 1, each element of this vector is that average is zero, variance is separate multiple Gaussian random variable;

(3) via node carries out precoding step to received signal

(3a) via node r y to received signal _rin the aligned signal s that comprises ^{[r, 1]}, s ^{[r, 2]}and s ^{[r, 3]}carry out separation, use s ^{[r, 1]}m/2 * N dimension separation matrix H ₁be multiplied by y _r, meet H ₁[U ₂u ₃]=0 _{m/2 * M}, obtain the aligned signal y after first separation ^{[r, 1]}=H ₁y _r=H ₁u ₁w ₁s ^{[r, 1]}+ H ₁n _r, 0 _{m/2 * M}the full null matrix that represents M/2 * M, similarly, selects s ^{[r, 2]}, s ^{[r, 3]}separation matrix H ₂, H ₃, meet respectively H ₂[U ₁u ₃]=0 _{m/2 * M}, H ₃[U ₁u ₂]=0 _{m/2 * M}, obtain second aligned signal y after separation ^{[r, 2]}=H ₂y _r=H ₂u ₂w ₂s ^{[r, 2]}+ H ₂n _raligned signal after separated with the 3rd, y ^{[r, 3]}=H ₃y _r=H ₃u ₃w ₃s ^{[r, 3]}+ H ₃n _r;

(3b) to the aligned signal y after separation ^{[r, 1]}, y ^{[r, 2]}, y ^{[r, 3]}carry out precoding, obtain the transmitted signal of N * 1 $x_r=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{V}^{[i,r]}{\mathrm{\Σ}}^{[i,r]}{U}^{[i,r]}{y}^{[r,i]},$

In formula, V ^{[i, r]}for interference cancellation matrix, ∑ ^{[i, r]}for the power division diagonal matrix of via node, the value of its diagonal element is according to aligned signal s ^{[r, i]}, i=1, each data flow adopts average power allocation to determine in 2,3, U ^{[i, r]}for the additional pre-coding matrix of via node;

(4) via node broadcast transmission signals step

Via node r is by the transmitted signal x that carries out precoding and obtain _rbe broadcast to user node, M * 1 dimension that obtains user node i receives signal and is: y _i=H ^{[i, r]}x _r+ n _i, i ∈ 1,2,3},

N in formula _ibe the noise vector of M * 1, each element of this vector is that average is zero, variance is separate multiple Gaussian random variable, H ^{[i, r]}be via node to the channel matrix of M * N of user node i, each element of this matrix is to obey the separate multiple Gaussian random variable that average is zero, variance is 1;

(5) user node recovers data step

User node i is according to receiving signal y _iand the transmitted signal x of oneself _iremove the interference of self, from receiving signal y _imiddle by x _iby the part signal forming after channel, cut; Then adopt ZF to detect and recover the primary signal s that user j sends to user i ^{[i, j]}, j ∈ { 1,2,3}, j ≠ i.

The present invention compared with prior art tool has the following advantages:

Method for transmitting signals in existing multiple-input and multiple-output Y-channel is from the angle design of the degree of freedom, the additional pre-coding matrix of user node and via node is unit matrix, can reach with speed less, the present invention is guaranteeing on the basis that the degree of freedom can reach, the additional pre-coding matrix of user node and via node has been designed to unitary matrice, compare with existing additional pre-coding matrix, increased the diversity that additional pre-coding matrix element is selected, therefore improved reaching and speed of system.

Accompanying drawing explanation

Fig. 1 is existing multiple-input and multiple-output Y-channel model;

Fig. 2 is flow chart of the present invention;

Fig. 3 is the performance simulation figure of the present invention under fixed relay station signal to noise ratio condition;

Fig. 4 is the performance simulation figure of the present invention under fixed-line subscriber node signal to noise ratio condition.

Embodiment

Method for transmitting signals proposed by the invention is applicable to the multiple-input and multiple-output Y-channel model described in Fig. 1, this channel model comprises three user nodes and a via node, wherein each user node is equipped M root antenna, via node equipment N root antenna, each user node respectively sends a unicast info to two other user node by via node; Concrete communication process is as follows: first, three user nodes send information to via node simultaneously; Then, via node is broadcast to three user nodes after the information receiving is processed; Finally, the information being received according to oneself respectively by three user nodes and the own information sending recover the information that two other user node is issued oneself.

With reference to Fig. 2, it is as follows that the present invention utilizes the channel model in Fig. 1 to carry out the step of signal transmission:

Step 1, user node carries out precoding to primary signal.

(1.1) each user node carries out precoding by the information of issuing two other user node, be about to signal space alignment matrix, user node adds pre-coding matrix, and user node power division matrix and primary signal multiplication of vectors, the M * 1 dimension transmitted signal that obtains sending user node is:

In formula, i represents to send user node, and j represents to receive user node, and i, and { M represents to send the antenna number of user node place preparation to j ∈ for 1,2,3}, i ≠ j, and M is even number, s ^{[j, i]}be to send the primary signal vector that user node i issues a M/2 * 1 data flow formation that receives user node j, each element of this vector is that average is mutually independent random variables zero, that variance is 1, ∑ ^{[j, i]}the sending node power division diagonal matrix of M/2 * M/2, W _{π (j, i)}be the additional pre-coding matrix of M/2 * M/2 user node, subscript π (j, i) is an index function, meets π (i, j)=π (j, i), and has π (1,2)=1, π (1,3)=2, and π (2,3)=3, therefore has W _{π (2,1)}=W _{π (1,2)}=W ₁, W _{π (3,1)}=W _{π (1,3)}=W ₂, W _{π (3,2)}=W _{π (2,3)}=W ₃, V ^{[j, i]}be the signal space alignment matrix of M * M/2, meet H ^{[r, i]}v ^{[j, i]}=H ^{[r, j]}v ^{[i, j]}, H ^{[r, i]}represent that user node i is to the channel matrix of N * M of via node r, N represents the antenna number of via node r place configuration, and has N=3M/2, and each element of this channel matrix is separate multiple Gaussian random variable zero, that variance is 1 for obeying average;

(1.2) according to channel matrix H ^{[r, i]}and H ^{[r, j]}, determine signal space alignment matrix V ^{[j, i]}:

Mode one, issues user 2 primary signal s for user 1 ^[2,1]required signal space alignment matrix V ^[2,1], and user 2 issues user 1 primary signal s ^[1,2]required signal space alignment matrix V ^[1,2], selection mode is as follows:

If V ^[2,1], V ^[1,2]and U ₁k row be respectively u _{k, 1}, k ∈ 1,2 ..., M/2}, according to signal space alignment matrix V ^[2,1]and V ^[1,2]need satisfied condition: H ^{[r, 1]}v ^[2,1]=H ^{[r, 2]}v ^[1,2]=U ₁, solving of these three column vectors is equivalent to the equation solving below:

$\left[\begin{array}{ccc}{I}_N& {-H}^{[r,1]}& 0_{N\×M}\\ I_N& 0_{N\×M}& {-H}^{[r,2]}\end{array}\right]\left(\begin{array}{c}{u}_{k,1}\\ v_k^{\left[\mathrm{2,1}\right]}\\ v_k^{\left[\mathrm{1,2}\right]}\end{array}\right)=0_{2N\×1},---\left(1\right)$

Solve above-mentioned equation (1) and obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as U ₁, V ^[2,1], V ^[1,2]m/2 row, in formula, I _nthe unit matrix that represents N * N, 0 _{n * M}the full null matrix that represents N * M, H ^{[r, 1]}and H ^{[r, 2]}represent that respectively user node 1 and user node 2 are to the channel matrix of via node;

Mode two, issues user 3 primary signal s for user 1 ^[3,1]required signal space alignment matrix V ^[3,1], and user 3 issues user 1 primary signal s ^[1,3]required signal space alignment matrix V ^[1,3], selection mode is as follows:

If V ^[3,1], V ^[1,3]and U ₂k row be respectively u _{k, 2}, k ∈ 1,2 ..., M/2}, according to signal space alignment matrix V ^[3,1]and V ^[1,3]need satisfied condition: H ^{[r, 1]}v ^[3,1]=H ^{[r, 3]}v ^[1,3]=U ₂, solving of these three column vectors is equivalent to the equation solving below:

$\left[\begin{array}{ccc}{I}_N& {-H}^{[r,1]}& 0_{N\×M}\\ I_N& 0_{N\×M}& {-H}^{[r,3]}\end{array}\right]\left(\begin{array}{c}{u}_{k,2}\\ v_k^{\left[\mathrm{3,1}\right]}\\ v_k^{[1,3]}\end{array}\right)=0_{2N\×1},---\left(2\right)$

Solve above-mentioned equation (2) and obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as U ₂, V ^[3,1], V ^[1,3]m/2 row, in formula, H ^{[r, 1]}and H ^{[r, 3]}represent that respectively user node 1 and user node 3 are to the channel matrix of via node;

Mode three, issues user 3 primary signal s for user 2 ^[3,2]required signal space alignment matrix V ^[3,2], and user 3 issues user 2 primary signal s ^[2,3]required signal space alignment matrix V ^[2,3], selection mode is as follows:

If V ^[3,2], V ^[2,3]and U ₃k row be respectively u _{k, 3}, k ∈ 1,2 ..., M/2}, according to signal space alignment matrix V ^[3,2]and V ^[2,3]need satisfied condition: H ^{[r, 2]}v ^[3,2]=H ^{[r, 3]}v ^[2,3]=U ₃, solving of these three column vectors is equivalent to the equation solving below:

$\left[\begin{array}{ccc}{I}_N& {-H}^{[r,2]}& 0_{N\×M}\\ I_N& 0_{N\×M}& {-H}^{[r,3]}\end{array}\right]\left(\begin{array}{c}{u}_{k,3}\\ v_k^{[3,2]}\\ v_k^{[2,3]}\end{array}\right)=0_{2N\×1},---\left(3\right)$

Solve above-mentioned equation (3) and obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as U ₃, V ^[3,2], V ^[2,3]m/2 row, in formula, H ^{[r, 2]}and H ^{[r, 3]}represent that respectively user node 2 and user node 3 are to the channel matrix of via node;

(1.3) determine that user node adds pre-coding matrix W _i:

First, according to following three condition: H ₁[U ₂u ₃]=0 _{m/2 * M}, H ₂[U ₁u ₃]=0 _{m/2 * M}, H ₃[U ₁u ₂]=0 _{m/2 * M}, obtain respectively first separation matrix H ₁, second separation matrix H ₂, and the 3rd separation matrix H ₃, in formula, 0 _{m/2 * M}the full null matrix that represents M/2 * M, U ₁=H ^{[r, 1]}v ^[2,1]=H ^{[r, 2]}v ^[1,2]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 2]}the friendship space of column space, U ₂=H ^{[r, 1]}v ^[3,1]=H ^{[r, 3]}v ^[1,3]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 3]}the friendship space of column space, U ₃=H ^{[r, 2]}v ^[3,2]=H ^{[r, 3]}v ^[2,3]represent channel matrix H ^{[r, 2]}with channel matrix H ^{[r, 3]}the friendship space of column space;

Then, by matrix H _iu _i, { 1,2,3} carries out singular value decomposition to i ∈, obtains in formula, subscript H represents conjugate transpose, U _[i]and V _[i]the unitary matrice of M/2 * M/2, ∑ _[i]it is the diagonal matrix of M/2 * M/2;

Finally, the unitary matrice V obtaining according to above-mentioned singular value decomposition _[i], select the additional pre-coding matrix of user node to be: W _i=V _[i], i ∈ { 1,2,3};

(1.4) determine sending node power division diagonal matrix sigma ^{[j, i]}diagonal element:

If send user node power division matrix ∑ ^{[j, i]}k diagonal element be set the transmitted signal x of user node i _imeeting power constraint condition is $E\left\{\mathrm{tr}\left(x_i{x}_i^H\right)\right\}{P}_i,i\∈\left\{\mathrm{1,2,3}\right\},$ ?

$\mathrm{tr}\left\{\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}{V}^{[j,i]}{V}_{\left[\mathrm{\π}(j,i)\right]}{\mathrm{\Σ}}^{[j,i]}{\left(V^{[j,i]}{V}_{\left[\mathrm{\π}(j,i)\right]}{\mathrm{\Σ}}^{[j,i]}\right)}^H\right\}$

$=\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}\underset{k=1}{\overset{M/2}{\mathrm{\Σ}}}{w}_k^{[j,i]}\left|\right|{\tilde{v}}_k^{[j,i]}\left|\right|\≤P_i$

In formula, symbol E represents to ask expectation, and tr represents to ask matrix trace, V ^{[j, i]}represent signal space alignment matrix, V _{[π (j, i)]}represent primary signal vector s ^{[j, i]}additional pre-coding matrix, representing matrix V ^{[j, i]}v _{[π (j, i)]}k row, || || represent to ask vectorial euclideam norm, primary signal vector s ^{[j, i]}the power of k data flow;

According to primary signal vector s ^{[j, i]}, j ∈ 1,2,3}, in j ≠ i, each data flow adopts the condition of average power allocation: obtain sending node power division matrix ∑ ^{[j, i]}k diagonal element be: $w_k^{[j,i]}=\sqrt{P_i/\left(M{\left|\right|{\tilde{v}}_k^{[j,i]}\left|\right|}^2\right)},$ k∈{1，2，…，M/2}。

Step 2, user node is to via node transmitted signal.

Send user node i by the transmitted signal x that carries out precoding and obtain _isend to via node r, N * 1 dimension that obtains via node receives signal and is:

$y_r=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{H}^{[r,i]}{x}_i+n_r$

$=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{H}^{[r,i]}\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}{V}^{[j,i]}{W}_{\mathrm{\π}(j,i)}{\mathrm{\Σ}}^{[j,i]}{s}^{[j,i]}+n_r$

$=U_1{W}_1{s}^{[r,1]}+U_2{W}_2{s}^{[r,2]}+U_3{W}_3{s}^{[r,3]}+n_r,$

In formula, s ^{[r, 1]}=∑ ^[2,1]s ^[2,1]+ ∑ ^[1,2]s ^[1,2]the aligned signal that represents user node 1 and user node 2, s ^{[r, 2]}=∑ ^[3,1]s ^[3,1]+ ∑ ^[1,3]s ^[1,3]the aligned signal that represents user node 1 and user node 3, s ^{[r, 3]}=∑ ^[3,2]s ^[3,2]+ ∑ ^[2,3]s ^[2,3]the aligned signal that represents user node 2 and user node 3, U ₁=H ^{[r, 1]}v ^[2,1]=H ^{[r, 2]}v ^[1,2]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 2]}the friendship space of column space, U ₂=H ^{[r, 1]}v ^[3,1]=H ^{[r, 3]}v ^[1,3]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 3]}the friendship space of column space, U ₃=H ^{[r, 2]}v ^[3,2]=H ^{[r, 3]}v ^[2,3]represent channel matrix H ^{[r, 2]}with channel matrix H ^{[r, 3]}the friendship space of column space, n _rbe the noise vector of N * 1, each element of this vector is that average is zero, variance is separate multiple Gaussian random variable.

Step 3, via node carries out precoding to received signal.

(3.1) via node r y to received signal _rin the aligned signal s that comprises ^{[r, 1]}, s ^{[r, 2]}and s ^{[r, 3]}carry out separation, use s ^{[r, 1]}m/2 * N dimension separation matrix H ₁be multiplied by y _r, meet H ₁[U ₂u ₃]=0 _{m/2 * M}, obtain the aligned signal y after first separation ^{[r, 1]}=H ₁y _r=H ₁u ₁w ₁s ^{[r, 1]}+ H ₁n _r, 0 _{m/2 * M}the full null matrix that represents M/2 * M, similarly, selects s ^{[r, 2]}, s ^{[r, 3]}separation matrix H ₂, H ₃, meet respectively H ₂[U ₁u ₃]=0 _{m/2 * M}, H ₃[U ₁u ₂]=0 _{m/2 * M}, obtain second aligned signal after separated with the 3rd:

y ^[r，2]＝H ₂y _r＝H ₂U ₂W ₂s ^[r，2]+H ₂n _r，y ^[r，3]＝H ₃y _r＝H ₃U ₃W ₃s ^[r，3]+H ₃n _r；

(3.2) to the aligned signal y obtaining after separation ^{[r, i]}, i ∈ 1,2,3} carries out precoding:

(3.2a) establish the aligned signal y after separation ^{[r, i]}, { the additional pre-coding matrix of 1,2,3} is U to i ∈ ^{[i, r]}, this matrix obtains as follows:

By matrix H _iu _i, { 1,2,3} carries out singular value decomposition to i ∈, obtains in formula, H _irepresent aligned signal s ^{[r, i]}separation matrix, U _irepresent that corresponding user node is to the friendship space of the column space of a pair of channel matrix of via node, subscript H represents conjugate transpose, U _[i]and V _[i]the unitary matrice of M/2 * M/2, ∑ _[i]it is the diagonal matrix of M/2 * M/2;

The unitary matrice U obtaining according to above-mentioned singular value decomposition _[i], select the additional pre-coding matrix of via node to be: $U^{[i,r]}=U_{\left[i\right]}^H,i\∈\left\{\mathrm{1,2,3}\right\}.$

(3.2b) establish the aligned signal y after separation ^{[r, i]}, { interference cancellation matrix of 1,2,3} is V to i ∈ ^{[i, r]}, this matrix is according to channel matrix H ^{[j, r]}determine one of as follows:

Mode 1, establishes interference cancellation matrix V ^{[1, r]}k classify as k ∈ 1,2 ..., M/2}, according to interference cancellation matrix V ^{[1, r]}need satisfied condition: by V ^{[1, r]}solving of M/2 column vector be equivalent to the equation solving below:

$H^{[3,r]}{v}_k^{[1,r]}=0_{M\×1},---1)$

Solve above-mentioned equation 1) obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as V ^{[1, r]}m/2 row, in formula, 0 _{m * 1}the full null vector that represents M * 1;

Mode 2, establishes interference cancellation matrix V ^{[2, r]}k classify as k ∈ 1,2 ..., M/2}, according to interference cancellation matrix V ^{[2, r]}need satisfied condition: will be to V ^{[2, r]}solving of M/2 column vector be equivalent to the equation solving below:

$H^{[2,r]}{v}_k^{[2,r]}=0_{M\×1},---2)$

Solve above-mentioned equation 2) obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as V ^{[2, r]}m/2 row;

Mode 3, establishes interference cancellation matrix V ^{[3, r]}k classify as k ∈ 1,2 ..., M/2}, according to interference cancellation matrix V ^{[3, r]}need satisfied condition: by V ^{[3, r]}solving of M/2 column vector be equivalent to the equation solving below:

$H^{[1,r]}{v}_k^{[3,r]}=0_{M\×1},---3)$

Solve above-mentioned equation 3) obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as V ^{[3, r]}m/2 row.

(3.2c) establish the aligned signal y after separation ^{[r, i]}, { the power division diagonal matrix of 1,2,3} is ∑ to i ∈ ^{[i, r]}, this matrix obtains as follows:

First, establish the power division diagonal matrix sigma of via node ^{[i, r]}k diagonal element be i ∈ 1,2,3}, k ∈ 1,2 ... M/2}, according to the transmitted signal of via node is:

$x_r=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{V}^{[i,r]}{\mathrm{\Σ}}^{[i,r]}{\mathrm{\Σ}}_{\left[i\right]}{s}^{[r,i]}+\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{V}^{[i,r]}{\mathrm{\Σ}}^{[i,r]}{U}_{\left[i\right]}^H{H}_i{n}_r$

Note represent via node transmitted signal x _rin useful signal;

Secondly, establish useful signal power be be expressed as:

$P_{x_r^s}=E\left\{\mathrm{tr}\left(x_r^s{\left(x_r^s\right)}^H\right)\right\}$

$=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}\underset{k=1}{\overset{M/2}{\mathrm{\Σ}}}{\left({\mathrm{\α}}_k^{[\mathrm{\π}(j,i),r]}{\mathrm{\β}}_{[k,\mathrm{\π}(j,i)]}\right)}^2\{{\left(w_k^{[j,i]}\right)}^2+{\left(w_k^{[i,j]}\right)}^2\}$

In formula, subscript H represents conjugate transpose, and symbol tr represents to ask matrix trace, β _{[k, π (j, i)}, π (j, i) ∈ 1,2,3}, j ≠ i, k ∈ 1,2 ..., M/2} representing matrix ∑ _{[π (j, i)]}k diagonal element, j, i ∈ 1,2,3}, j ≠ i, k ∈ 1,2 ..., M/2} represents user node power division matrix ∑ ^{[j, i]}k diagonal element;

Finally, establish useful signal the power of the 3M/2 a comprising data flow is λ ², according to via node transmitted signal x _rsatisfied power constraint condition: try to achieve the value of λ; According to relational expression:

${\left({\mathrm{\α}}_k^{[\mathrm{\π}(j,i),r]}{\mathrm{\β}}_{[k,\mathrm{\π}(j,i)]}\right)}^2\{{\left(w_k^{\left[\mathrm{1,2}\right]}\right)}^2+{\left(w_k^{\left[\mathrm{2,1}\right]}\right)}^2\}={\mathrm{\λ}}^2$

Obtain via node place power division matrix ∑ ^{[π (j, i), r]}, π (j, i) ∈ { the diagonal element value of 1,2,3} k ∈ 1,2 ... M/2} is:

${\mathrm{\α}}_k^{[\mathrm{\π}(j,i),r]}=\frac{\mathrm{\λ}}{{\mathrm{\β}}_{[k,\mathrm{\π}(j,i)]}\sqrt{\{{\left(w_k^{[j,i]}\right)}^2+{\left(w_k^{[j,i]}\right)}^2\}}}---\left(a\right)$

According to above-mentioned formula (a), can obtain respectively power division diagonal matrix sigma ^{[1, r]}k diagonal element be:

${\mathrm{\α}}_k^{[1,r]}=\frac{\mathrm{\λ}}{{\mathrm{\β}}_{[k,1]}\sqrt{\{{\left(w_k^{\left[\mathrm{2,1}\right]}\right)}^2+{\left(w_k^{\left[\mathrm{1,2}\right]}\right)}^2\}}},k\∈\{\mathrm{1,2},...M/2\},$

Power division diagonal matrix sigma ^{[2, r]}k diagonal element be:

${\mathrm{\α}}_k^{[2,r]}=\frac{\mathrm{\λ}}{{\mathrm{\β}}_{[k,2]}\sqrt{\{{\left(w_k^{\left[\mathrm{3,1}\right]}\right)}^2+{\left(w_k^{\left[\mathrm{1,3}\right]}\right)}^2\}}},k\∈\{\mathrm{1,2},...M/2\},$

Power division diagonal matrix sigma ^{[3, r]}k diagonal element be:

${\mathrm{\α}}_k^{[3,r]}=\frac{\mathrm{\λ}}{{\mathrm{\β}}_{[k,3]}\sqrt{\{{\left(w_k^{[3,2]}\right)}^2+{\left(w_k^{[2,3]}\right)}^2\}}},k\∈\{\mathrm{1,2},...M/2\}.$

(3.2d) by the additional pre-coding matrix U of the via node obtaining in step (3.2a)-(3.2c) ^{[i, r]}, interference cancellation matrix V ^{[i, r]}, and via node power division diagonal matrix sigma ^{[i, r]}, i ∈ 1,2,3} with separated after aligned signal y ^{[r, i]}, i ∈ 1,2,3} multiplies each other, and obtains the transmitted signal of N * 1 of via node r:

Step 4, via node broadcast transmission signal.

Via node r is by the transmitted signal x that carries out precoding and obtain _rbe broadcast to user node, M * 1 dimension that obtains user node i receives signal and is: y _i=H ^{[i, r]}x _r+ n _i, i ∈ 1,2,3},

N in formula _ibe the noise vector of M * 1, each element of this vector is that average is zero, variance is separate multiple Gaussian random variable, H ^{[i, r]}be via node to the channel matrix of M * N of user node i, each element of this matrix is to obey the separate multiple Gaussian random variable that average is zero, variance is 1.

Step 5, user node recovers data.

User node i is according to receiving signal y _iand the transmitted signal x of oneself _iremove the interference of self, from receiving signal y _imiddle by x _iby the part signal forming after channel, cut; Then adopt ZF to detect and recover the primary signal s that user j sends to user i ^{[i, j]}, j ∈ { 1,2,3}, j ≠ i.

Effect of the present invention can further illustrate by following emulation:

1. simulated conditions:

Each user node of setting multiple-input and multiple-output Y-channel and the antenna number of via node configuration are respectively: M=6, and N=9, the noise variance of establishing each user node and via node place is σ ², definition user node i, i ∈ the signal to noise ratio of 1,2,3} and via node r is respectively: $\mathrm{SNR}\left(i\right)=\frac{P_i}{{\mathrm{\σ}}^2},\mathrm{SNR}\left(r\right)=\frac{P_r}{{\mathrm{\σ}}^2}.$

2. emulation content

(a) under the condition of signal to noise ratio snr (the r)=15dB at fixed relay place, compare reaching and speed of method for transmitting signals proposed by the invention under constant power condition and existing method for transmitting signals, simulation result as shown in Figure 3, abscissa in Fig. 3 represents the signal to noise ratio of user node, ordinate represents reaching of multiple-input and multiple-output Y-channel and speed, with the unit of speed be bits/trans, the bit number that every transmission primaries can send.

The meaning that in Fig. 3, two curves represent is as follows:

" existing method " represents the existing method for transmitting signals under constant power condition, and " the present invention " represents method for transmitting signals proposed by the invention.

As seen from Figure 3, under the condition of the signal to noise ratio at fixed relay place, method for transmitting signals proposed by the invention is significantly improved than reaching with speed of existing method for transmitting signals, and the increase along with user node place signal to noise ratio, improve more obvious, when signal to noise ratio is 20dB, the of the present invention and existing method for transmitting signals of speed ratio approximately increases by 4 bits.

(b) fixing the signal to noise ratio snr of each user node (i)=15dB, i ∈ { 1,2, under the condition of 3}, compare reaching and speed of method for transmitting signals proposed by the invention under constant power condition and existing method for transmitting signals, as shown in Figure 4, the abscissa in Fig. 4 represents the signal to noise ratio of via node to simulation result, and ordinate represents reaching of multiple-input and multiple-output Y-channel and speed.

The meaning that in Fig. 4, two curves represent is as follows:

" existing method " represents the existing method for transmitting signals under constant power condition, and " the present invention " represents method for transmitting signals proposed by the invention.

As seen from Figure 4, under the condition of signal to noise ratio of fixing each user node place, method for transmitting signals proposed by the invention also has obvious gain than reaching with speed of existing method for transmitting signals, when being 10bits/trans with speed, the present invention compares with existing method for transmitting signals, has the gain of 4dB.

Simulation result by Fig. 3, Fig. 4 can be found out, the in the situation that of difference fixed relay station signal to noise ratio and fixed-line subscriber node signal to noise ratio, reaching with speed of method for transmitting signals proposed by the invention is all significantly improved than reaching with speed of existing method for transmitting signals, this is because method for transmitting signals proposed by the invention is now to carry out additional prelisting on methodical basis, thereby increased the diversity that pre-coding matrix element is selected, therefore can obtain higher and speed.

Claims (5)

1. a method for transmitting signals for coding Network Based in multiple-input and multiple-output Y-channel, comprising:

(1) user node carries out precoding step to primary signal

Each user node carries out precoding by the information of issuing two other user node, and the M * 1 dimension transmitted signal that obtains sending user node is: $x_i=\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}{V}^{[j,i]}{W}_{\mathrm{\π}(j,i)}{\mathrm{\Σ}}^{[j,i]}{s}^{[j,i]},$

In formula, i represents to send user node, and j represents to receive user node, and i, and { M represents to send the antenna number of user node place preparation to j ∈ for 1,2,3}, i ≠ j, and M is even number, s ^{[j, i]}be to send the primary signal vector that user node i issues a M/2 * 1 data flow formation that receives user node j, each element of this vector is that average is mutually independent random variables zero, that variance is 1, Σ ^{[j, i]}the sending node power division diagonal matrix of M/2 * M/2, W _{π (j, i)}be the additional pre-coding matrix of M/2 * M/2 user node, subscript π (j, i) is an index function, meets π (i, j)=π (j, i), and has π (1,2)=1, π (1,3)=2, and π (2,3)=3, therefore has W _{π (2,1)}=W _{π (1,2)}=W ₁, W _{π (3,1)}=W _{π (1,3)}=W ₂, W _{π (3,2)}=W _{π (2,3)}=W ₃, V ^{[j, i]}be the signal space alignment matrix of M * M/2, meet H ^{[r, i]}v ^{[j, i]}=H ^{[r, j]}v ^{[i, j]}, H ^{[r, i]}represent that user node i is to the channel matrix of N * M of via node r, N represents the antenna number of via node r place configuration, and has N=3M/2, and each element of this channel matrix is separate multiple Gaussian random variable zero, that variance is 1 for obeying average;

(2) user node is to via node transmitted signal step

Send user node i by the transmitted signal x that carries out precoding and obtain _isend to via node r, N * 1 dimension that obtains via node receives signal and is:

$\begin{array}{c}{y}_r=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{H}^{[r,i]}{x}_i+n_r\\ =\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{H}^{[r,i]}\underset{j=1,j\≠i}{\overset{3}{\mathrm{\Σ}}}{V}^{[j,i]}{W}_{\mathrm{\π}(j,i)}{\mathrm{\Σ}}^{[j,i]}{s}^{[j,i]}+n_r\\ =U_1{W}_1{s}^{[r,1]}+U_2{W}_2{s}^{[r,2]}+U_3{W}_3{s}^{[r,3]}+n_r\end{array},$

In formula, s ^{[r, 1]}=Σ ^[2,1]s ^[2,1]+ Σ ^[1,2]s ^[1,2]the aligned signal that represents user node 1 and user node 2,

S ^{[r, 2]}=Σ ^[3,1]s ^[3,1]+ Σ ^[1,3]s ^[1,3]the aligned signal that represents user node 1 and user node 3,

S ^{[r, 3}]=Σ ^[3,2]s ^[3,2]+ Σ ^[2,3]s ^[2,3]the aligned signal that represents user node 2 and user node 3,

U ₁=H ^{[r, 1]}v ^[2,1]=H ^{[r, 2]}v ^[1,2]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 2]}the friendship space of column space, U ₂=H ^{[r, 1]}v ^[3,1]=H ^{[r, 3]}v ^[1,3]represent channel matrix H ^{[r, 1]}with channel matrix H ^{[r, 3]}the friendship space of column space, U ₃=H ^{[r, 2}] V ^[3,2]=H ^{[r, 3]}v ^[2,3]represent channel matrix H ^{[r, 2]}with channel matrix H ^{[r, 3]}the friendship space of column space, n _rbe the noise vector of N * 1, each element of this vector is that average is zero, variance is separate multiple Gaussian random variable;

(3) via node carries out precoding step to received signal

(3a) via node r y to received signal _rin the aligned signal s that comprises ^{[r, 1]}, s ^{[r, 2]}and s ^{[r, 3]}carry out separation, use s ^{[r, 1]}m/2 * N dimension separation matrix H ₁be multiplied by y _r, meet H ₁[U ₂u ₃]=0 _{m/2 * M}, obtain the aligned signal y after first separation ^{[r, 1]}=H ₁y _r=H ₁u ₁w ₁s ^{[r, 1]}+ H ₁n _r, 0 _{m/2 * M}the full null matrix that represents M/2 * M, similarly, selects s ^{[r, 2]}, s ^{[r, 3]}separation matrix H ₂, H ₃, meet respectively H ₂[U ₁u ₃]=0 _{m/2 * M}, H ₃[U ₁u ₂]=0 _{m/2 * M}, obtain second aligned signal y after separation ^{[r, 2]}=H ₂y _r=H ₂u ₂w ₂s ^{[r, 2}]+H ₂n _raligned signal after separated with the 3rd, y ^{[r, 3]}=H ₃y _r=H ₃u ₃w ₃s ^{[r, 3]}+ H ₃n _r;

(3b) to the aligned signal y after separation ^{[r, 1]}, y ^{[r, 2]}, y ^{[r, 3]}carry out precoding, obtain the transmitted signal of N * 1 $x_r=\underset{i=1}{\overset{3}{\mathrm{\Σ}}}{V}^{[i,r]}{\mathrm{\Σ}}^{[i,r]}{U}^{[i,r]}{y}^{[r,i]},$

In formula, V ^{[i, r]}for interference cancellation matrix, Σ ^{[i, r]}for the power division diagonal matrix of via node, the value of its diagonal element is according to aligned signal s ^{[r, i]}, i=1, each data flow adopts average power allocation to determine in 2,3, U ^{[i, r]}for the additional pre-coding matrix of via node;

(4) via node broadcast transmission signals step

Via node r is by the transmitted signal x that carries out precoding and obtain _rbe broadcast to user node, M * 1 dimension that obtains user node i receives signal and is: y _i=H ^{[i, r]}x _r+ n _i, i ∈ 1,2,3},

N in formula _ibe the noise vector of M * 1, each element of this vector is that average is zero, variance is separate multiple Gaussian random variable, H ^{[i, r]}be via node to the channel matrix of M * N of user node i, each element of this matrix is to obey the separate multiple Gaussian random variable that average is zero, variance is 1;

(5) user node recovers data step

User node i is according to receiving signal y _iand the transmitted signal x of oneself _iremove the interference of self, from receiving signal y _imiddle by x _iby the part signal forming after channel, cut; Then adopt ZF to detect and recover the primary signal s that user j sends to user i ^{[i, j]}, j ∈ { 1,2,3}, j ≠ i;

Signal space alignment matrix V in described step (1) ^{[j, i]}according to channel matrix H ^{[r, i]}and H ^{[r, j]}determine one of as follows:

Mode one, establishes V ^[2,1], V ^[1,2]and U ₁k row be respectively u _{k, 1}, k ∈ 1,2 ..., M/2}, according to signal space alignment matrix V ^[2,1]and V ^[1,2]need satisfied condition: H ^{[r, 1]}v ^[2,1]=H ^{[r, 2]}v ^[1,2]=U ₁, solving of these three column vectors is equivalent to the equation solving below:

$\left[\begin{array}{ccc}{I}_N& -H^{[r,1]}& 0_{N\&times;M}\\ I_N& 0_{N\&times;M}& -H^{[r,2]}\end{array}\right]\left(\begin{array}{c}{u}_{k,1}\\ v_k^{\left[\mathrm{2,1}\right]}\\ v_k^{\left[\mathrm{1,2}\right]}\end{array}\right)=0_{2N\&times;1},---<1>$

Solve above-mentioned equation <1> and obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as U ₁, V ^[2,1], V ^[1,2]m/2 row, in formula, I _nthe unit matrix that represents N * N, 0 _{n * M}the full null matrix that represents N * M, H ^{[r, 1]}and H ^{[r, 2]}represent that respectively user node 1 and user node 2 are to the channel matrix of via node;

Mode two, establishes V ^[3,1], V ^[1,3]and U ₂k row be respectively u _{k, 2}, k ∈ 1,2 ..., M/2}, according to signal space alignment matrix V ^[3,1]and V ^[1,3]need satisfied condition: H ^{[r, 1]}v ^[3,1]=H ^{[r, 3]}v ^[1,3]=U ₂, solving of these three column vectors is equivalent to the equation solving below:

$\left[\begin{array}{ccc}{I}_N& -H^{[r,1]}& 0_{N\&times;M}\\ I_N& 0_{N\&times;M}& -H^{[r,3]}\end{array}\right]\left(\begin{array}{c}{u}_{k,2}\\ v_k^{\left[\mathrm{3,1}\right]}\\ v_k^{\left[\mathrm{1,3}\right]}\end{array}\right)=0_{2N\&times;1},---<2>$

Solve above-mentioned equation <2> and obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as U ₂, V ^[3,1], V ^[1,3]m/2 row, in formula, H ^{[r, 1]}and H ^{[r, 3]}represent that respectively user node 1 and user node 3 are to the channel matrix of via node;

Mode three, establishes V ^[3,2], V ^[2,3]and U ₃k row be respectively u _{k, 3}, k ∈ 1,2 ..., M/2}, according to signal space alignment matrix V ^[3,2]and V ^[2,3]need satisfied condition:

H ^[r,2]V ^[3,2]＝H ^[r,3]V ^[2,3]＝U ₃

Solving of these three column vectors is equivalent to the equation solving below:

$\left[\begin{array}{ccc}{I}_N& -H^{[r,2]}& 0_{N\&times;M}\\ I_N& 0_{N\&times;M}& -H^{[r,3]}\end{array}\right]\left(\begin{array}{c}{u}_{k,3}\\ v_k^{\left[\mathrm{3,2}\right]}\\ v_k^{\left[\mathrm{2,3}\right]}\end{array}\right)=0_{2N\&times;1},---<3>$

Solve above-mentioned equation <3> and obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as U ₃, V ^[3,2], V ^[2,3]m/2 row, in formula, H ^{[r, 2]}and H ^{[r, 3]}represent that respectively user node 2 and user node 3 are to the channel matrix of via node;

Interference cancellation matrix V in described step (3b) ^{[i, r]}, be according to channel matrix H ^{[j, r]}determine one of in the following manner:

Mode 1, establishes interference cancellation matrix V ^{[1, r]}k classify as k ∈ 1,2 ..., M/2}, according to interference cancellation matrix V ^{[1, r]}need satisfied condition: $\mathrm{span}(v_1^{[1,r]},...,v_{M/2}^{[1,r]})\&Subset;\mathrm{null}\left(H^{[3,r]}\right),$ By V ^{[1, r]}solving of M/2 column vector be equivalent to the equation solving below:

$H^{[3,r]}{v}_k^{[1,r]}=0_{M\&times;1},---1)$

Solve above-mentioned equation 1) obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as V ^{[1, r]}m/2 row, in formula, 0 _{m * 1}the full null vector that represents M * 1;

Mode 2, establishes interference cancellation matrix V ^{[2, r]}k classify as k ∈ 1,2 ..., M/2}, according to interference cancellation matrix V ^{[2, r]}need satisfied condition: $\mathrm{span}(v_1^{[2,r]},...,v_{M/2}^{[2,r]})\&Subset;\mathrm{null}\left(H^{[2,r]}\right),$ Will be to V ^{[2, r]}solving of M/2 column vector be equivalent to the equation solving below:

$H^{[2,r]}{v}_k^{[2,r]}=0_{M\&times;1},---2)$

Solve above-mentioned equation 2) obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as V ^{[2, r]}m/2 row;

Mode 3, establishes interference cancellation matrix V ^{[3, r}] k classify as k ∈ 1,2 ..., M/2}, according to interference cancellation matrix V ^{[3, r]}need satisfied condition: $\mathrm{span}(v_1^{[3,r]},...,v_{M/2}^{[3,r]})\&Subset;\mathrm{null}\left(H^{[1,r]}\right),$ By V ^{[3, r]}solving of M/2 column vector be equivalent to the equation solving below:

$H^{[1,r]}{v}_k^{[3,r]}=0_{M\&times;1},---3)$

Solve above-mentioned equation 3) obtain M/2 solution, the unit vector that this M/2 solution is turned to pairwise orthogonal is as V ^{[3, r]}m/2 row.

2. the method for transmitting signals of coding Network Based in multiple-input and multiple-output Y-channel according to claim 1, the additional pre-coding matrix W of user node in wherein said step (1) _idetermine in the following way:

By matrix H _iu _i, { 1,2,3} carries out singular value decomposition to i ∈, obtains in formula, H _irepresent aligned signal s ^{[r, i]}separation matrix, U _irepresent that corresponding user node is to the friendship space of the column space of a pair of channel matrix of via node, subscript H represents conjugate transpose, U _[i]and V _[i]the unitary matrice of M/2 * M/2, Σ _[i]the diagonal matrix of M/2 * M/2,

The unitary matrice V obtaining according to above-mentioned singular value decomposition _[i], select the additional pre-coding matrix of user node to be: W _i=V _[i], i ∈ { 1,2,3}.

3. the method for transmitting signals of coding Network Based in multiple-input and multiple-output Y-channel according to claim 1, the sending node power division matrix Σ in wherein said step (1) ^{[j, i]}diagonal element, determine in the following manner:

If send user node power division matrix Σ ^{[j, i]}k diagonal element be k ∈ 1,2 ..., M/2}, the transmitted signal x of setting user node i _imeeting power constraint condition is $E\left\{\mathrm{tr}\left(x_i{x}_i^H\right)\right\}\&le;P_i,i\&Element;\left\{\mathrm{1,2,3}\right\},$ ?

$\begin{array}{c}\mathrm{tr}\left\{\underset{j=1,j\&NotEqual;i}{\overset{3}{\mathrm{\&Sigma;}}}{V}^{[j,i]}{V}_{\left[\mathrm{\&pi;}(j,i)\right]}{\mathrm{\&Sigma;}}^{[j,i]}{\left(V^{[j,i]}{V}_{\left[\mathrm{\&pi;}(j,i)\right]}{\mathrm{\&Sigma;}}^{[j.i]}\right)}^H\right\}\\ =\underset{j=1,j\&NotEqual;i}{\overset{3}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{M/2}{\mathrm{\&Sigma;}}}{\left(w_k^{[j,i]}\right)}^2{\left|\right|{\tilde{v}}_k^{[j,i]}\left|\right|}^2\&le;P_i\end{array}$

In formula, symbol E represents to ask expectation, and tr represents to ask matrix trace, V ^{[j, i]}represent signal space alignment matrix, V _{[π (j, i)]}represent primary signal vector s ^{[j, i]}additional pre-coding matrix, representing matrix V ^{[j, i]}v _{[π (j, i)]}k row, || || represent to ask vectorial euclideam norm, primary signal vector s ^{[j, i]}the power of k data flow;

According to primary signal vector s ^{[j, i]}, j ∈ 1,2,3}, in j ≠ i, each data flow adopts the condition of average power allocation: obtain sending user node power division matrix Σ ^{[j, i]}k diagonal element be $w_k^{[j,i]}=\sqrt{P_i/\left(M{\left|\right|{\tilde{v}}_k^{[j,i]}\left|\right|}^2\right)},k\&Element;\{\mathrm{1,2},...,M/2\}.$

4. the method for transmitting signals of coding Network Based in multiple-input and multiple-output Y-channel according to claim 1, the additional pre-coding matrix U of via node in wherein said step (3b) ^{[i, r]}, determine in the following manner:

By matrix H _iu _i, { 1,2,3} carries out singular value decomposition to i ∈, obtains in formula, H _irepresent aligned signal s ^{[r, i]}separation matrix, U _irepresent that corresponding user node is to the friendship space of the column space of a pair of channel matrix of via node, subscript H represents conjugate transpose, U _[i]and V _[i]the unitary matrice of M/2 * M/2, Σ _[i]it is the diagonal matrix of M/2 * M/2;

The unitary matrice U obtaining according to above-mentioned singular value decomposition _[i], select the additional pre-coding matrix of via node to be: $U^{[i,r]}=U_{\left[i\right]}^H,i\&Element;\left\{\mathrm{1,2,3}\right\}.$

5. the method for transmitting signals of coding Network Based in multiple-input and multiple-output Y-channel according to claim 1, wherein the value of the diagonal element described in step (3b) is according to aligned signal s ^{[r, i]}, i=1, each data flow adopts average power allocation to determine in 2,3, carries out as follows:

First, establish the power division matrix Σ of via node ^{[i, r]}, { k the diagonal element of 1,2,3} is i ∈ i ∈ 1,2,3}, k ∈ 1,2 ... M/2}, according to the transmitted signal of via node is:

$x_r=\underset{i=1}{\overset{3}{\mathrm{\&Sigma;}}}{V}^{[i,r]}{\mathrm{\&Sigma;}}^{[i,r]}{\mathrm{\&Sigma;}}_{\left[i\right]}{s}^{[r,i]}+\underset{i=1}{\overset{3}{\mathrm{\&Sigma;}}}{V}^{[i,r]}{\mathrm{\&Sigma;}}^{[i,r]}{U}_{\left[i\right]}^H{H}_i{n}_r$

Note $x_r^s=\underset{i=1}{\overset{3}{\mathrm{\&Sigma;}}}{V}^{[i,r]}{\mathrm{\&Sigma;}}^{[i,r]}{\mathrm{\&Sigma;}}_{\left[i\right]}{s}^{[r,i]},$ represent via node transmitted signal x _rin useful signal;

Secondly, establish useful signal power be be expressed as:

$\begin{array}{c}{P}_{x_r^s}=E\left\{\mathrm{tr}\left(x_r^s{\left(x_r^s\right)}^H\right)\right\}\\ =\underset{i=1}{\overset{3}{\mathrm{\&Sigma;}}}\underset{j=1,j\&NotEqual;i}{\overset{3}{\mathrm{\&Sigma;}}}\underset{k=1}{\overset{M/2}{\mathrm{\&Sigma;}}}{\left({\mathrm{\&alpha;}}_k^{[\mathrm{\&pi;}(j,i),r]}{\mathrm{\&beta;}}_{[k,\mathrm{\&pi;}(j,i)]}\right)}^2\{{\left(w_k^{[j,i]}\right)}^2+{\left(w_k^{[j,i]}\right)}^2\}\end{array}$

In formula, subscript H represents conjugate transpose, and symbol tr represents to ask matrix trace, β _{[k, π (j, i)]}, π (j, i) ∈ 1,2,3}, j ≠ i, k ∈ 1,2 ..., M/2} representing matrix Σ _{[π (j, i)]}k diagonal element, j, i ∈ 1,2,3}, j ≠ i, k ∈ 1,2 ..., M/2} represents user node power division diagonal matrix Σ ^{[j, i]}k diagonal element;

Finally, establish useful signal the power of the 3M/2 a comprising data flow is λ ², according to via node transmitted signal x _rsatisfied power constraint condition: try to achieve the value of λ; According to relational expression

${\left({\mathrm{\&alpha;}}_k^{[\mathrm{\&pi;}(j,i),r]}{\mathrm{\&beta;}}_{[k,\mathrm{\&pi;}(j,i)]}\right)}^2\{{\left(w_k^{\left[\mathrm{1,2}\right]}\right)}^2+{\left(w_k^{\left[\mathrm{2,1}\right]}\right)}^2\}={\mathrm{\&lambda;}}^2$

Obtain the power division matrix Σ of via node place ^{[π (j, i), r]}, π (j, i) ∈ { the diagonal element value of 1,2,3} for:

${\mathrm{\&alpha;}}_k^{[\mathrm{\&pi;}(j,i),r]}=\frac{\mathrm{\&lambda;}}{{\mathrm{\&beta;}}_{[k,\mathrm{\&pi;}(j,i)]}\sqrt{\{{\left(w_k^{[j,i]}\right)}^2+{\left(w_k^{[j,i]}\right)}^2\}}},k\&Element;\{\mathrm{1,2},...,M/2\}.$