CN101741448A - Minimum mean square error beam forming-based information transmission method in two-way channel - Google Patents

Abstract

The invention discloses a minimum mean square error beam forming-based information transmission method in a two-way channel, belongs to the technical field of wireless communication, and mainly solves the problem of high technical complexity of the prior art. Particularly, the invention provides the information transmission method capable of improving channel capacity and error ratio performance, and reducing complexity at the same time by aiming at a wireless communication network of two source nodes and a relay node. The method comprises the following steps that: the two source nodes simultaneously send encoding information to the relay node; the relay node builds a beam forming matrix according to a minimum mean square error rule; the relay node carries out beam forming processing on the received source node mixed information, and then broadcasts the source node mixed information after the beam forming processing to the source nodes; and the source nodes are decoded to acquire information of other parts. The method reduces the complexity and ensures the correctness of transmission without great influence on system performance, and can be used for reliable information transmission in a two-way relay system.

Description

In the two-way channel based on the information transferring method of least mean-square error beam forming

Technical field

The invention belongs to wireless communication technology field, relate to network code and beam forming, specifically, can be used for improving channel capacity and performance of BER at the cordless communication network of two source nodes and a via node.

Background technology

Utilizing relaying to help the mobile subscriber to transmit data, can obtain extra diversity gain, improve the errored bit BER performance of receiving terminal, is to improve the mobile subscriber in one of effective means of cell edge speech quality.At wireless both-way trunk channel shown in Figure 1, two source nodes need four time slots during by a via node exchange message usually, and promptly two source nodes take a time slot respectively and via node communicates, via node takies two time slots, is respectively two source nodes and transmits data.

Network code carries out transmitting after certain linearity or the non-uniform encoding as the signal that a kind of design in the modern communication networks allows network node that multilink is received, reduced the number of transmissions of packet, improved network throughput, strengthen the fault-tolerance and the robustness of network, for the efficiency of transmission that improves two-way trunk channel provides an effective method.Particularly in wireless network,, more effectively solved the interference problem in the wireless network because its broadcast characteristic in wireless transmission makes it can better utilize network code.

Usually the exchange primary information needs four time slots, we can finish once exchange in three time slots by utilizing network code at via node, its basic principle is two source nodes to be taken a time slot respectively and via node communicates, recompile was transmitted after the data message that decoding is obtained two source nodes by via node carried out Modulo-two operation then, need three time slots just can exchange a packet, utilization can be finished this process based on the TWRC of analog network coding ANC in two time slots, source node sends information to relaying simultaneously, amplify at relaying then and transmit operation, thereby realized the exchange of information between two source nodes.

Beam forming method in the cordless communication network is according to fading channel state design beam forming matrix, utilize the beam forming matrix to carry out preliminary treatment elimination interchannel interference to sending signal, improve the transmission performance in the multidiameter fading channel, the capacity optimal beam forming method of two-way trunk channel is different from the design in the legacy network, the capacity limit problem that all is based on system is designed the beam forming matrix, though existing optimal beam forming method can be eliminated interference when satisfying power system capacity, but there is the complexity problem of higher, and the suboptimum beam forming method though but to have reduced the complexity bit error rate performance relatively poor, give in the transmission course and bring more error bit, the correctness of influence transmission.

Summary of the invention

The objective of the invention is to overcome the shortcoming of above-mentioned prior art, information transferring method based on the least mean-square error beam forming has been proposed in a kind of two-way channel, with under relaying power limited condition, reduce the complexity of beam forming matrix, improve the error rate of system performance, avoid the error bit in the transmission course, guaranteed the correctness of transmission.

The present invention is achieved in that

(1) adopts the data message d of Turbo code to two source nodes ₁, d ₂Encode respectively, obtain coded sequence x ₁, x ₂, this coded sequence is modulated, and the signal s after will modulating ₁(n) and s ₂(n) with power p ₁And p ₂Be sent to via node simultaneously;

(2) via node makes up the beam forming matrix based on minimum mean square error criterion:

(2a) make H _UL=[h ₁, h ₂], H _DL=[h ₂, h ₁] ^T, h ₁And h ₂Be respectively the channel of source node, to H to via node _ULCarry out singular value decomposition H _UL=U ∑ V ^H, U ∈ C ^{M * 2}, be A=U with relaying least mean-square error beam forming matrix notation ^*BU ^H, B ∈ C ^{2 * 2}

(2b) adopt minimum mean square error criterion to make up junction waves beam forming matrix:

Receiver And Transmitter expression formula according to minimum mean square error criterion makes up relaying least mean-square error beam forming matrix:

$A_{\mathrm{MMSE}}=H_{\mathrm{DL}}^H{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}{a}_{\mathrm{<figure-callout\; id="0"\; label="MMSE"\; filenames="H2009102193421E0000031.png"\; state="\{\{state\}\}">MMSE</figure-callout>}}& 0\\ 0& b_{\mathrm{MMSE}}\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}{H}_{\mathrm{UL}}^H$

σ wherein _n ²And σ _a ²Be respectively noise power and signal power, a _MMSEAnd b _MMSEBe to make up two complex variables that the beam forming matrix need be found the solution, I is multiple unit matrix,

Be that the least mean-square error receiver is expressed formula,

It is least mean-square error transmitter expression formula;

Utilize the singular value decomposition formula

With

Respectively to H _DL ^HAnd H _UL ^HDecompose, obtain junction waves beam forming matrix and be:

$A_{\mathrm{MMSE}}=H^{*}{\mathrm{\&Sigma;V}}^T{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}0& a_{\mathrm{MMSE}}\\ b_{\mathrm{MMSE}}& 0\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}{\mathrm{V\&Sigma;U}}^H$

Order $B_{\mathrm{MMSE}}={\mathrm{\&Sigma;V}}^T{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}0& a_{\mathrm{MMSE}}\\ b_{\mathrm{MMSE}}& 0\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\mathrm{V\&Sigma;},$

Make A _MMSE=U ^*B _MMSEU ^H

B in the formula _MMSEMatrix passes through source node S ₁And S ₂Achievable rate find the solution and obtain U ^*Be the conjugate matrices of U, U ^HAssociate matrix for U;

(3) utilize the least mean-square error beam forming matrix that obtains, the mixed signal that via node receives is carried out the least mean-square error beam forming, be expressed as x _R(n)=A _MMSEy _R(n), n=1 ..., N, wherein y _R(n) mixed information of two source nodes that receive for via node;

(4) via node at second time slot with beam forming information x _R(n) be broadcast to two source nodes, these two source nodes deduct from the information that receives separately respectively after disturbing mutually, obtain the information of the other side's source node by demodulation, to realize the exchange of information between the source node in two time slots.

The present invention compared with prior art has following advantage:

1. the present invention is owing to adopt achievable rate optimization to make up least mean-square error beam forming matrix at via node, systematic function is not being affected greatly down, greatly reduce complexity than the optimal beam forming method, guaranteed the correctness of transmission simultaneously.

2. the present invention is owing to adopted the least mean-square error beam forming on via node, make via node not need to translate the transmission information of two source nodes and directly amplification forwarding, can finish the information interaction of two source nodes at two time slots, make throughput reach 1/2 symbol/user/symbol period, improved throughput of system;

Description of drawings

Fig. 1 is existing wireless both-way trunk communication network model schematic diagram;

Fig. 2 is information transmission process figure of the present invention;

Fig. 3 is that matrix B is found the solution in achievable rate optimization of the present invention _MMSEFlow chart;

Fig. 4 is a capacity of the present invention territory performance comparison diagram;

Fig. 5 is a bit error rate performance comparison diagram of the present invention.

Embodiment

With reference to Fig. 2, based on the information transferring method of least mean-square error beam forming, comprise the steps: in the two-way channel

Step 1, two source nodes send information simultaneously

With reference to Fig. 1, adopt the data message d of Turbo code to two source nodes ₁, d ₂Encode respectively, obtain coded sequence x ₁, x ₂, this coded sequence is modulated, and the signal s after will modulating ₁(n) and s ₂(n) with power p ₁And p ₂Be sent to via node simultaneously.

Step 2, via node makes up the beam forming matrix

Via node is constructed the beam forming matrix according to minimum mean square error criterion:

(2.1) make uplink channel H _UL=[h ₁, h ₂], downlink channel H _DL=[h ₂, h ₁] ^T, h ₁And h ₂Be respectively the channel of source node, to uplink channel H to via node _ULCarry out singular value decomposition H _UL=U ∑ V ^H, U ∈ C ^{M * 2}, be A with relaying least mean-square error beam forming matrix notation _MMSE=U ^*B _MMSEU ^H, B _MMSE∈ C ^{2 * 2}

(2.2) the Receiver And Transmitter expression formula according to minimum mean square error criterion makes up relaying least mean-square error beam forming matrix:

$A_{\mathrm{MMSE}}=H_{\mathrm{DL}}^H{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}{a}_{\mathrm{<figure-callout\; id="0"\; label="MMSE"\; filenames="H2009102193421E0000031.png"\; state="\{\{state\}\}">MMSE</figure-callout>}}& 0\\ 0& b_{\mathrm{MMSE}}\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}{H}_{\mathrm{UL}}^H$

σ wherein _n ²And σ _a ²Be respectively noise power and signal power, a _MMSEAnd b _MMSEBe to make up two complex variables that the beam forming matrix need be found the solution, I is multiple unit matrix,

For known receiver expression formula based on least mean-square error is,

For known transmitter expression formula based on least mean-square error is,

(2.3) utilize the singular value decomposition formula

With To H _DL ^HAnd H _UL ^HDecompose respectively, the above-mentioned least mean-square error beam forming of substitution matrix is translated into:

$A_{\mathrm{MMSE}}=H^{*}{\mathrm{\&Sigma;V}}^T{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}0& a_{\mathrm{MMSE}}\\ b_{\mathrm{MMSE}}& 0\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}{\mathrm{V\&Sigma;U}}^H$

Order $B_{\mathrm{MMSE}}={\mathrm{\&Sigma;V}}^T{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}0& a_{\mathrm{MMSE}}\\ b_{\mathrm{MMSE}}& 0\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="n"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">n</figure-callout>}}^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\mathrm{V\&Sigma;},$

Make A _MMSE=U ^*B _MMSEU ^H

U in the formula ^*Be the conjugate matrices of U, U ^HBe the associate matrix of U, B _MMSEMatrix passes through source node S ₁And S ₂Achievable rate find the solution and obtain;

(2.4), utilize the optimization of source node achievable rate to find the solution B with reference to Fig. 3 _MMSEMatrix:

At first, introduce vectorial α=[α ₂₁, α ₁₂] ^T,

R _Snm=r ₂₁+ r ₁₂, r ₁₂Be source node S ₁By via node to source node S ₂The achievable rate of transmission, r ₂₁Be source node S ₂By via node to source node S ₁The achievable rate of transmission, for the vectorial α that determines, the source node achievable rate is optimized expression formula and is:

$\underset{R_{\mathrm{sum},B}}{\mathrm{Maximize}}{R}_{\mathrm{sum}}$

$\mathrm{Subject\; to}\frac{1}{2}{\log }_2(1+\frac{{\left|g_1^T{\mathrm{<figure-callout\; id="2"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_2\right|}^2{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_2}{{\left|\right|{B_{\mathrm{<figure-callout\; id="1"\; label="MMSE\; H\; g"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">MMSE</figure-callout>}}}^H{g}_{}^{}{}_1^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{21}{R}_{\mathrm{sum}}$

$\frac{1}{2}{\log }_2(1+\frac{{\left|g_2^T{\mathrm{<figure-callout\; id="1"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_1\right|}^2{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_1}{{\left|\right|{B_{\mathrm{<figure-callout\; id="2"\; label="MMSE\; H\; g"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">MMSE</figure-callout>}}}^H{g}_{}^{}{}_2^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{12}{R}_{\mathrm{sum}},$

${\left|\right|{\mathrm{<figure-callout\; id="1"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_1\left|\right|}^2{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_1+{\left|\right|{\mathrm{<figure-callout\; id="2"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_2\left|\right|}^2{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_2+\mathrm{tr}\left(B_{\mathrm{MMSE}}{B_{\mathrm{MMSE}}}^H\right)\&le;P_R$

G wherein ₁=U ^Hh ₁And g ₂=U ^Hh ₂Be respectively the equivalent channels of two source nodes, R to via node _SumBe the source node achievable rate;

Secondly, above-mentioned achievable rate optimization being converted into the expression formula that the repeat transmitted power optimization finds the solution is:

${\underset{B}{\mathrm{Minimize}}{p}_R=\left|\right|{\mathrm{<figure-callout\; id="1"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_1\left|\right|}^2{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_1+{\left|\right|{\mathrm{<figure-callout\; id="2"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_2\left|\right|}^2{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_2+\mathrm{tr}\left(B_{\mathrm{MMSE}}{B_{\mathrm{MMSE}}}^H\right)$

$\mathrm{subject\; to}\frac{1}{2}{\log }_2(1+\frac{{\left|g_1^T{\mathrm{<figure-callout\; id="2"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_2\right|}^2{\mathrm{<figure-callout\; id="2"\; label="p"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_2}{{\left|\right|{B_{\mathrm{<figure-callout\; id="1"\; label="MMSE\; H\; g"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">MMSE</figure-callout>}}}^H{g}_{}^{}{}_1^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{21}r,$

$\frac{1}{2}{\log }_2(1+\frac{{\left|g_2^T{\mathrm{<figure-callout\; id="1"\; label="B\; MMSE\; g"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">B</figure-callout>}}_{\mathrm{MMSE}}{g}_{}{}_1\right|}^2{\mathrm{<figure-callout\; id="1"\; label="p"\; filenames="H2009102193421E0000031.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">p</figure-callout>}}_1}{{\left|\right|{B_{\mathrm{<figure-callout\; id="2"\; label="MMSE\; H\; g"\; filenames="H2009102193421E0000011.png,H2009102193421E0000032.png"\; state="\{\{state\}\}">MMSE</figure-callout>}}}^H{g}_{}^{}{}_2^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{12}r$

Wherein r is a constant in the solution procedure;

At last, above-mentioned repeat transmitted power optimization expression formula is found the solution, step is as follows:

(a) set vectorial α, source node speed and the upper limit

And repeat transmitted power upper limit P _RBe known, the order

r _Min=0,

(b) according to the parametric solution repeat transmitted power optimization expression formula of setting, obtain optimal solution p _R ^*And the matrix B of this moment _MMSEIf,

Make r _Min=r, otherwise make r _Max=r;

(c) make δ _rBe a given constant, if r _Max-r _Min〉=δ _r, utilize

After upgrading the value of r, return step (b) and carry out circulation, continue to find the solution repeat transmitted power optimization expression formula, up to r _Max-r _Min≤ δ _r, loop ends finally obtains matrix B _MMSE

Step 3, the via node beam forming

The least mean-square error beam forming matrix that utilization obtains carries out the least mean-square error beam forming to the mixed signal that via node receives, and the mixed signal that via node receives is represented

z _R(n) expression additive white Gaussian noise, the via node beam forming is expressed as x _R(n)=A _MMSEy _R(n), n=1 ..., N, wherein A _MMSE∈ C ^{M * M}Be via node least mean-square error beam forming matrix.

Step 4, the via node broadcast message

The information x of via node after second time slot handled beam forming _R(n) be broadcast to two source nodes, via node is respectively h to the channel of source node ₁ ^TAnd h ₂ ^T, source node S ₁And S ₂The signal indication that receives is:

$y_1\left(n\right)=h_1^T{x}_R\left(R\right)+z_1\left(n\right)$

$=h_1^T{A}_{\mathrm{MMSE}}{h}_1\sqrt{p_1}{s}_1\left(n\right)+h_1^T{A}_{\mathrm{MMSE}}{h}_2\sqrt{p_2}{s}_2\left(n\right)+h_1^T{A}_{\mathrm{MMSE}}{z}_R+z_1\left(n\right),$

$y_2\left(n\right)=h_2^T{x}_R\left(R\right)+z_2\left(n\right)$

$=h_2^T{A}_{\mathrm{MMSE}}{h}_2\sqrt{p_2}{s}_2\left(n\right)+h_2^T{A}_{\mathrm{MMSE}}{h}_1\sqrt{p_1}{s}_1\left(n\right)+h_2^T{A}_{\mathrm{MMSE}}{z}_R\left(n\right)+z_2\left(n\right),$

Z wherein ₁(n) and z ₂(n) be respectively via node to source node S ₁And S ₂Additive white Gaussian noise, source node deducts from the information that receives separately respectively from disturbing mutually

With

After, obtain the information of the other side's source node by demodulation, thereby in two time slots, realize the exchange of information between the source node.

Useful result of the present invention illustrates by following further emulation:

Adopt the system model of a via node of two source nodes in the emulation, source node respectively disposes an antenna, four antennas of via node configuration, source node information adopt turbo coding and BPSK modulation to handle, and respectively the capacity and the error rate of system are carried out emulation.

The volumetric properties that Fig. 4 has provided the present invention and existing optimal beam forming method and two kinds of suboptimum beam forming methods compares, and from Fig. 4 as seen, the present invention can not bring considerable influence to the power system capacity performance under reduction beam forming matrix complex degree.

The bit error rate performance that Fig. 5 has provided the present invention and existing optimal beam forming method and two kinds of suboptimum beam forming methods compares, from Fig. 5 as seen, the present invention has improved the error rate of system performance preferably, has reduced the error of transmission bit, has guaranteed the correctness of transmission.

Claims (2)

1. In the two-way channel based on the information transferring method of least mean-square error beam forming, comprise the steps:

   (1) adopts the data message d of Turbo code to two source nodes ₁, d ₂Encode respectively, obtain coded sequence x ₁, x ₂, this coded sequence is modulated, and the signal s after will modulating ₁(n) and s ₂(n) with power p ₁And p ₂Be sent to via node simultaneously;

   (2) via node makes up the beam forming matrix based on minimum mean square error criterion:

   2a) make H _UL=[h ₁, h ₂], H _DL=[h ₂, h ₁] T, h ₁And h ₂Be respectively the channel of source node, to H to via node _ULCarry out singular value decomposition H _UL=U ∑ V ^H, U ∈ C ^{M * 2}, be A with relaying least mean-square error beam forming matrix notation _MMSE=U ^*B _MMSEU ^H, B _MMSE∈ C ^{2 * 2}

   2b) adopt minimum mean square error criterion to make up junction waves beam forming matrix:

   Receiver And Transmitter expression formula according to minimum mean square error criterion makes up relaying least mean-square error beam forming matrix:

   $A_{\mathrm{MMSE}}=H_{\mathrm{DL}}^H{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_n^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}{a}_{\mathrm{MMSE}}& 0\\ 0& b_{\mathrm{MMSE}}\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_n^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}{H}_{\mathrm{UL}}^H$

   σ wherein _n ²And σ _a ²Be respectively noise power and signal power, a _MMSEAnd b _MMSEBe to make up two complex variables that the beam forming matrix need be found the solution, I is multiple unit matrix,

   

   Be that the least mean-square error receiver is expressed formula, It is least mean-square error transmitter expression formula;

   Utilize the singular value decomposition formula

   

   With

   

   Respectively to H _DL ^HAnd H _UL ^HDecompose, obtain junction waves beam forming matrix and be:

   $A_{\mathrm{MMSE}}=U^{*}{\mathrm{\&Sigma;V}}^T{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_n^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}0& a_{\mathrm{MMSE}}\\ b_{\mathrm{MMSE}}& 0\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_n^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}{\mathrm{V\&Sigma;U}}^H$

   Order $B_{\mathrm{MMSE}}={\mathrm{\&Sigma;V}}^T{(H_{\mathrm{DL}}{H_{\mathrm{DL}}}^H+\frac{{\mathrm{\&sigma;}}_n^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\left[\begin{array}{cc}0& a_{\mathrm{MMSE}}\\ b_{\mathrm{MMSE}}& 0\end{array}\right]{(H_{\mathrm{UL}}^H{H}_{\mathrm{UL}}+\frac{{\mathrm{\&sigma;}}_n^2}{{\mathrm{\&sigma;}}_a^2}I)}^{+}\mathrm{V\&Sigma;},$

   Make A _MMSE=U ^*B _MMSEU ^H

   B in the formula _MMSEMatrix passes through source node S ₁And S ₂Achievable rate find the solution and obtain U ^*Be the conjugate matrices of U, U ^HAssociate matrix for U;

   (3) utilize the least mean-square error beam forming matrix that obtains, the mixed signal that via node receives is carried out the least mean-square error beam forming, be expressed as x _R(n)=A _MMSEy _R(n), n=1 ..., N, wherein y _R(n) mixed information of two source nodes that receive for via node;

   (4) via node at second time slot with beam forming information x _R(n) be broadcast to two source nodes, these two source nodes deduct from the information that receives separately respectively after disturbing mutually, obtain the information of the other side's source node by demodulation, to realize the exchange of information between the source node in two time slots.
2. 2. in the two-way channel according to claim 1 based on the information transferring method of least mean-square error beam forming, step 2b wherein) described matrix B _MMSEAchievable rate by source node is found the solution, and carries out according to following steps:

   The first step is introduced vectorial α=[α ₂₁, α ₁₂] ^T, wherein

   

   R _Sum=r ₂₁+ r ₁₂, r ₁₂Be source node S ₁By via node to source node S ₂The achievable rate of transmission, r ₂₁Be source node S ₂By via node to source node S ₁The achievable rate of transmission,

   For the vectorial α that determines, the optimization of source node achievable rate is expressed as:

   $\begin{array}{cc}\underset{R_{\mathrm{sum},B}}{\mathrm{Maximize}}& R_{\mathrm{sum}}\end{array}$

   $\begin{array}{ccc}\mathrm{Subject}& \mathrm{to}& \frac{1}{2}{\log }_2(1+\frac{{\left|g_1^T{B}_{\mathrm{MMSE}}{g}_2\right|}^2{p}_2}{{\left|\right|{B_{\mathrm{MMSE}}}^H{g}_1^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{21}{R}_{\mathrm{sum}}\end{array},$

   $\frac{1}{2}{\log }_2(1+\frac{{\left|g_2^T{B}_{\mathrm{MMSE}}{g}_1\right|}^2{p}_1}{{\left|\right|{B_{\mathrm{MMSE}}}^H{g}_2^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{12}{R}_{\mathrm{sum}}$

   ${\left|\right|B_{\mathrm{MMSE}}{g}_1\left|\right|}^2{p}_1+{\left|\right|B_{\mathrm{MMSE}}{g}_2\left|\right|}^2{p}_2+\mathrm{tr}\left(B_{\mathrm{MMSE}}{B_{\mathrm{MMSE}}}^H\right)\&le;P_R$

   G wherein ₁=U ^Hh ₁, g ₂=U ^Hh ₂, R _SumBe the source node achievable rate;

   In second step, above-mentioned achievable rate optimization be converted into the expression formula that the repeat transmitted power optimization finds the solution be:

   p _R＝||B _MMSEg ₁|| ²p ₁+||B _MMSEg ₂|| ²p ₂+tr(B _MMSEB _MMSE ^H)

   subject?to $\frac{1}{2}{\log }_2(1+\frac{{\left|g_1^T{B}_{\mathrm{MMSE}}{g}_2\right|}^2{p}_2}{{\left|\right|{B_{\mathrm{MMSE}}}^H{g}_1^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{21}r,$

   $\frac{1}{2}{\log }_2(1+\frac{{\left|g_2^T{B}_{\mathrm{MMSE}}{g}_1\right|}^2{p}_1}{{\left|\right|{B_{\mathrm{MMSE}}}^H{g}_2^{*}\left|\right|}^2+1})\&GreaterEqual;{\mathrm{\&alpha;}}_{12}r$

   R is a constant in the solution procedure:

   In the 3rd step, the repeat transmitted power optimization in second step is found the solution expression formula finds the solution:

   (a) set vectorial α, source node speed and the upper limit

   

   And repeat transmitted power upper limit P _RBe known, the order r _Min=0,

   (b) according to the parametric solution repeat transmitted power optimization expression formula of setting, obtain optimal solution p _R ^*And the matrix B of this moment _MMSEIf, Make r _Min=r, otherwise make r _Max=r;

   (c) make δ _rBe a given constant, if r _Max-r _Min〉=δ _r, utilize

   

   After upgrading the value of r, return step (b) and carry out circulation, continue to find the solution repeat transmitted power optimization expression formula, up to r _Max-r _Min≤ δ _r, loop ends finally obtains matrix B _MMSE

Images