CN102571279B - Combined signal processing method for source end and relay end in bidirectional relay system - Google Patents

Abstract

A combined signal processing method for a source end and a relay end in a bidirectional relay system comprises the following steps: the source end emits a training sequence to a relay, and the relay performs backward channel estimation to obtain an estimation channel between the source end and the relay; the relay emits a training sequence to the source end, and the source end performs forward channel estimation to obtain an estimation channel between the relay and a user; the source end feed forward channel information back to the relay, the relay performs iterative computation to source end nonlinear pre-coding matrixes, relay linear pre-coding matrixes and source end receiving equalization matrixes; the relay feeds the information of the source end back to the source end; emitting signals are subjected to nonlinear pre-treatment through the source end and then emitted to the relay; the signals received by the relay are subjected to linear pre-treatment through the relay and then broadcasted to the source end; and the source end detects the received signals to obtain information required to be transmitted therebetween. According to the invention, the bidirectional relay information transmission mode is adopted to improve the channel capacity, and the nonlinear signal processing method is adopted at the source end, so that the bit error rate performance of the system is improved.

Description

The united signal processing method of source and relay in bidirectional relay system

Technical field

What the present invention relates to is a kind of method of wireless communication field, specifically the united signal processing method of source and relay in a kind of bidirectional relay system.

Background technology

Relaying technique in mobile communication system, effectively can expand the coverage of network, improve power system capacity, by 3GPP (3rd Generation Partnership Project, 3G (Third Generation) Moblie partner program) LTE-A (Long Term Evolation-Advanced, Long Term Evolution-senior) standard adopted.Current trunking scheme mainly contains DF (Decode-and-Forward, decoding forwards), AF (Amplify-and-Forward, amplification forwarding) etc.Wherein AF mode is widely used owing to realizing simple, low complex degree.The information that traditional relay system will realize between two users passes mutually, needs 4 time slots.And bidirectional relay system realizes information between two users and passes mutually and only need 2 time slots, therefore, adopt the bidirectional relay system power system capacity that can double.In addition, multiple antenna can be installed in source and relay, improves systematic function further in conjunction with MIMO (Multiple-Input Multiple-Output, multiple-input and multiple-output) technology.In order to give full play to the advantage such as diversity, spatial reuse of MIMO technology, need the signal processing method of further design source and relay.Linear signal processing method realizes simple due to it, is widely used, but in contrast to this, adopts Nonlinear harmonic oscillator method but can obtain better systematic function.

Through finding existing literature search, Ronghong Mo Yong Huat Chew, " MMSE-Based Joint Source and Relay Precoding Design for Amplify-and-Forward MIMO Relay Networks, " IEEE Trans.Wireless Commun., vol.8, no.9, pp.4668-4676, 2009 (" based on the associating source of MMSE criterion and relaying code Design in AFMIMO relay systems ", ), what this article adopted in source is linear signal processing mode, compared with Nonlinear harmonic oscillator mode, it has lower complexity, but also bring system system loss of energy.

Find through retrieval again, Fan-Shuo Tseng, Min-Yao Chang, Wen-Rong Wu, " Joint Tomlinson-Harashima Source and Linear Relay Precoders Design in Amplify-and-Forward MIMO Relay System via MMSE Criterion, " IEEE transaction on vehicular technology.vol.60, No.4, MAY 2011 (" based on the associating THP source precoding of MMSE criterion and relay Progressive linear precoder design in AF MIMO relay system ", IEEE vehicle technology periodical, 60th phase, 4th volume, 2011.03), this article considers AF MIMO relay scene, base station and relaying adopt THP precoding and linear predictive coding respectively, and it is minimum for optimization aim with the MSE of system.Realize transinformation between two users, if adopt traditional relay system to need 4 time slots, and if adopt bidirectional relay system, then needs 2 time slots, the throughput that thus can double.

Also find through retrieval, Rui Wang Meixia Tao, " Joint Source and Relay Precoding Designs for MIMO Two-wWay Relay Systems, " IEEE ICC, 2011 (" the associating source in MIMO bidirectional relay system and relaying Precoding Designs ", IEEE international communication conference, 2011), the signal transacting of this article co-design source and relaying, with the mean square error of minimized detection signal, but what this technology adopted in source is linear signal processing method, compared with employing Nonlinear harmonic oscillator method, realize simple, but bring the systemic loss of energy such as mean square error and bit error rate.

Summary of the invention

The object of the invention is to the above-mentioned deficiency overcoming prior art, the source in a kind of bidirectional relay system and relay signal processing method are provided.The present invention is according to MMSE (Minimum Mean Squared Error, least mean-square error) criterion, non-linear and the relay linear information processing method of co-design source, the method takes full advantage of the advantage of Nonlinear harmonic oscillator method, effectively can improve the bit error rate performance of system.

The present invention is achieved by the following technical solutions, the present invention includes following steps:

The first step, source 1 is to repeat transmitted training sequence S ₁, relaying is according to the signal X received ₁carry out backward channel estimating, obtain the backward channel H of source 1 and relay well ₁, source 2 is to repeat transmitted training sequence S simultaneously ₂, relaying is according to the signal X received ₂carry out backward channel estimating, obtain the backward channel H of source 2 and relay well ₂;

Second step, relaying is simultaneously to source 1 and source 2 transmitting training sequence S ₃, source 1 is according to the signal X received ₃carry out forward channel estimation, obtain the forward channel G between relaying and source 1 ₁, source 2 is according to the signal X received ₄carry out forward channel estimation, obtain the backward channel G between relaying and source 2 ₂;

3rd step, source 1 and source 2 will estimate the forward channel information G obtained ₁and G ₂feed back to relaying.Relaying is according to the transmitting pre-coding matrix C of all channel information iterative computation sources 1 ₁, F ₁, the transmitting pre-coding matrix C of source 2 ₂, F ₂, relaying pre-coding matrix F _rmatrix W balanced with the reception of source 1 ₁, the balanced matrix W of reception of source 2 ₂;

4th step, the source 1 calculated is launched pre-coding matrix C by relaying ₁, F ₁with the balanced matrix W of reception ₁feed back to source 1, source 2 is launched pre-coding matrix C ₂, F ₂with the balanced matrix W of reception ₂feed back to source 2, simultaneously by channel information H ₁feed back to source 1, by H ₂feed back to source 2;

5th step, source 1 is to the s that transmits in advance ₁carry out preliminary treatment, obtain the x that transmits ₁, and by this signal x ₁be transmitted to relaying, source 2 is to the s that transmits in advance simultaneously ₂carry out preliminary treatment, obtain the x that transmits ₂, and by this signal x ₂be transmitted to relaying;

6th step, relaying is y to the received signal _rcarry out linear process, obtain signal and will be broadcast to source 1 and source 2;

7th step, source 1 y to the received signal ₁carry out check processing, obtain signal signal is obtained again through the modular arithmetic identical with step 5 meanwhile, source 2 y to the received signal ₂carry out check processing, obtain signal signal is obtained again through the modular arithmetic identical with step 5

Backward channel estimation process method in the described first step is:

$H_1=\sqrt{\frac{M_1}{{\mathrm{\ρ}}_{\mathrm{\τ}1}}}{X}_1{S}_1^{*}{(\frac{M_1}{{\mathrm{\ρ}}_{\mathrm{\τ}1}}{I}_{M_1}+S_1{S}_1^{*})}^{-1}$

$H_2=\sqrt{\frac{M_2}{{\mathrm{\ρ}}_{\mathrm{\τ}2}}}{X}_2{S}_2^{*}{(\frac{M_2}{{\mathrm{\ρ}}_{\mathrm{\τ}2}}{I}_{M_2}+S_2{S}_2^{*})}^{-1}$

Wherein: M ₁the antenna number of source 1, M ₂the antenna number of source 2, ρ _{τ 1}training sequence S ₁signal to noise ratio, ρ _{τ 2}training sequence S ₂signal to noise ratio, t _{τ 1}the length of the training sequence that source 1 is launched, t _{τ 2}the length of the training sequence that source 2 is launched, with be the signal that relay reception arrives, N is the antenna number of relaying, be respectively M ₁× M ₁and M ₂× M ₂unit matrix.

Forward channel estimating processing method in described second step is:

$G_1=\sqrt{\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}}{(\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}{I}_N+S_3{S}_3^{*})}^{-1}{S}_3{X}_3^{*}$

$G_2=\sqrt{\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}}{(\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}{I}_N+S_3{S}_3^{*})}^{-1}{S}_3{X}_4^{*}$

Wherein: N is the antenna number of relaying, ρ _{τ 3}training sequence S ₃signal to noise ratio, t _{τ 3}the length of the training sequence of repeat transmitted, with be respectively the signal that source 1 and source 2 receive, I _nfor the unit matrix of N × N.

Iterative calculation method in described 3rd step is:

Iteration one: suppose C ₁, F ₁, C ₂, F ₂, F _rknown, calculate the balanced matrix W of receiving terminal of source 1 ₁matrix W balanced with the receiving terminal of source 2 ₂, method is as follows:

W ₁＝C ₂F ₂ ^H(G ₁F _rH ₂) ^H(G ₁F _rH ₂F ₂F ₂ ^H(G ₁F _rH ₂) ^H+R _n1) ^-1

W ₂＝C ₁F ₁ ^H(G ₂F _rH ₁) ^H(G ₂F _rH ₁F ₁F ₁ ^H(G ₂F _rH ₁) ^H+R _n2) ^-1

R _n1＝G ₁F _rF _r ^HG ₁ ^H+I _N

R _n2＝G ₂F _rF _r ^HG ₂ ^H+I _N

I _nrepresent the unit matrix of N × N, upper table () ^hrepresent conjugate transpose;

Iteration two: suppose Fr, W1, W2 are known, calculate the transmitting pre-coding matrix C of source 1 ₁, F ₁, the transmitting pre-coding matrix C of source 2 ₂, F ₂, method is as follows:

First calculate the transmitting pre-coding matrix C of source 1 ₁, F ₁:

First to (G ₂f _rh ₁) ^h(G ₂f _rf _r ^hg ₂ ^h+ I _n) ^-1(G ₂f _rh ₁) do Eigenvalues Decomposition, obtain:

(G ₂f _rh ₁) ^h(G ₂f _rf _r ^hg ₂ ^h+ I _n) ^-1(G ₂f _rh ₁)=V Λ V ^h, wherein Λ is diagonal matrix, and V is unitary matrix;

Then carry out water injection power distribution to the elements in a main diagonal in diagonal matrix Λ, obtain power division matrix Ω, its elements in a main diagonal is wherein (y) ⁺=max (0, y), σ _sfor transmit signal energy, Λ _ifor i-th diagonal entry of Λ, u is for making Ω _imeet power constraints constant, wherein M ₁for the antenna number of source 1, P _s1for the transmitting power of source 1;

Order $D_1={({\mathrm{\Ω}}^H\mathrm{\Λ\Ω}+{\mathrm{\σ}}_s^{-2}{I}_{M_1})}^{-1/2}$

To D ₁do geometric mean decomposition (GMD), obtain:

D ₁=QRF, wherein R is the upper triangular matrix that diagonal entry is all equal, Q and F is unitary matrix; Finally obtain transmitting terminal pre-coding matrix C ₁, F ₁, be respectively:

$C_1=\mathrm{diag}\left\{R_{(k,k)}^H\right\}{\left(R^H\right)}^{-1}$

F ₁＝VΩF ^H

Wherein diag{A _{(k, k)}represent the diagonal matrix be made up of all diagonal entries of A;

Adopt similar method, the transmitting pre-coding matrix C of source 2 can be obtained ₂, F ₂;

Iteration three: suppose C ₁, F ₁, C ₂, F ₂, W ₁, W ₂known, calculate relaying pre-coding matrix F _r, method is as follows:

Use Lagrangian Arithmetic, can by F _rbe expressed as:

$F_r=\mathrm{mat}\left\{{[R_{x2}\⊗R_{r1}+R_{x1}\⊗R_{r2}+\mathrm{\λ}{R}_x\⊗I_n]}^{-1}\mathrm{vec}\left(R_r\right)\right\}$

Wherein R _x1=H ₁f ₁f ₁ ^hh ₁ ^h+ I _n

R _x2＝H ₂F ₂F ₂ ^HH ₂ ^H+I _N

R _r1＝G ₁ ^HW ₁ ^HW ₁G ₁

R _r2＝G ₂ ^HW ₂ ^HW ₂G ₂

R _r＝G ₁ ^HW ₁ ^HC ₂F ₂ ^HH ₂ ^H+G ₂ ^HW ₂ ^HC ₁F ₁ ^HH ₁ ^H

R _x＝H ₁F ₁F ₁ ^HH ₁ ^H+H ₂F ₂F ₂ ^HH ₂ ^H+I _N

λ is Lagrange multiplier, $\mathrm{\λ}\∈(0,\sqrt{\mathrm{trace}\left(R_r{\left(R_x\right)}^{-T}{R}_x{\left(R_x\right)}^{-T}{R_r}^H\right)}/P_r),$ P _rfor repeat transmitted power;

for Kronecker computing, vec () is matrixing vector operation, and mat () is the inverse operation of vec (), and trace () is track taking computing, () ^tfor transpose operation;

Dichotomy is used to search in its span λ, until repeat transmitted power equals P _r, that is:

trace{F _rR _xF _r ^H}＝P _r

Thus, optimum λ and relaying pre-coding matrix F can be obtained _r.

Preprocess method in described 5th step is:

x ₁＝F ₁C ₁ ^-1(s ₁+e ₁)

x ₂＝F ₂C ₂ ^-1(s ₂+e ₂)

Wherein s ₁=[s ₁1 ..., s _1N1] ^t, s ₂=[s ₂₁..., s _2N2] ^tbe through the signal after m-QAM modulation, N ₁for the fluxion that transmits of source 1, N ₂for the fluxion that transmits of source 2;

E ₁can ask with the following method:

$a_{1k}=s_{1k}-\underset{l=1}{\overset{k-1}{\mathrm{\Σ}}}{B}_1(\mathrm{bk},l)a_{1l}$

e ₁＝a ₁-s ₁

Wherein for under round, B ₁=C ₁-I _n1, e ₁=[e ₁₁..., e _1N1] ^t;

E ₂can ask with the following method:

$a_{2k}=s_{2k}-\underset{l=1}{\overset{k-1}{\mathrm{\Σ}}}{B}_2(k,l)a_{2l}$

e ₂＝a ₂-s ₂

Wherein for under round, B ₂=C ₂-I _n2, e ₂=[e ₂₁..., e _2N1] ^t.

Linear processing methods in described 6th step is:

Check processing method in described 7th step is:

Source 1:

Source 2:

Compared with prior art, this beneficial effect of the invention is the message transmission mode that have employed bi-directional relaying, can improve channel capacity greatly, and have employed nonlinear signal processing method in source, effectively improves the bit error rate performance of system.

Accompanying drawing explanation

Fig. 1 is that the bit error rate performance of one embodiment of the invention compares schematic diagram.

Embodiment

Below in conjunction with accompanying drawing, method of the present invention is further described: the present embodiment is implemented under premised on technical solution of the present invention, give detailed embodiment and concrete operating process, but protection scope of the present invention is not limited to following embodiment.

The antenna number M of source 1 and source 2 in the present embodiment ₁=4, M ₂=4, the antenna number of relaying is N=4, and treat that the symbol of transmission is mutually the QPSK modulation symbol of stochastic generation, backward and forward channel is Ruili (Rayleigh) flat fading, the reception noise of relaying and two receiving terminals is the white complex gaussian noise of zero mean unit variance the signal to noise ratio of backward channel is with the signal to noise ratio of forward channel is and SNR ₁=SNR ₂=SNR _r=8 20dB, the transmitting power of source 1 is the transmitting power of source 2 is the transmitting power of relaying is

The present embodiment comprises the following steps:

The first step, source 1 is to repeat transmitted training sequence S ₁, relaying is according to the signal X received ₁carry out backward channel estimating, obtain the backward channel H of source 1 and relay well ₁.Source 2 is to repeat transmitted training sequence S simultaneously ₂, relaying is according to the signal X received ₂carry out backward channel estimating, obtain the backward channel H of source 2 and relay well ₂.

Described backward channel estimation process, is:

$H_1=\sqrt{\frac{M_1}{{\mathrm{\ρ}}_{\mathrm{\τ}1}}}{X}_1{S}_1^{*}{(\frac{M_1}{{\mathrm{\ρ}}_{\mathrm{\τ}1}}{I}_{M_1}+S_1{S}_1^{*})}^{-1}$

$H_2=\sqrt{\frac{M_2}{{\mathrm{\ρ}}_{\mathrm{\τ}2}}}{X}_2{S}_2^{*}{(\frac{M_2}{{\mathrm{\ρ}}_{\mathrm{\τ}2}}{I}_{M_2}+S_2{S}_2^{*})}^{-1}$

Wherein: M ₁the antenna number of source 1, M ₂the antenna number of source 2, ρ _{τ 1}training sequence S ₁signal to noise ratio, ρ _{τ 2}training sequence S ₂signal to noise ratio, t _{τ 1}the length of the training sequence that source 1 is launched, t _{τ 2}the length of the training sequence that source 2 is launched, with be the signal that relay reception arrives, N is the antenna number of relaying, be respectively M ₁× M ₁and M ₂× M ₂unit matrix.

Training sequence length T in the present embodiment _{τ 1}=T _{τ 2}=4, the signal to noise ratio of training sequence is ρ _{τ 1}=ρ _{τ 2}={ 49,199,999}.

Second step, relaying is simultaneously to source 1 and source 2 transmitting training sequence S ₃, source 1 is according to the signal X received ₃carry out forward channel estimation, obtain the forward channel G between relaying and source 1 ₁, source 2 is according to the signal X received ₄carry out forward channel estimation, obtain the backward channel G between relaying and source 2 ₂.

Described forward channel estimates process, is:

$G_1=\sqrt{\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}}{(\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}{I}_N+S_3{S}_3^{*})}^{-1}{S}_3{X}_3^{*}$

$G_2=\sqrt{\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}}{(\frac{N}{{\mathrm{\ρ}}_{\mathrm{\τ}3}}{I}_N+S_3{S}_3^{*})}^{-1}{S}_3{X}_4^{*}$

Wherein: N is the antenna number of relaying, ρ _{τ 3}training sequence S ₃signal to noise ratio, t _{τ 3}the length of the training sequence of repeat transmitted, with be respectively the signal that source 1 and source 2 receive, I _nfor the unit matrix of N × N.

Training sequence length T in the present embodiment _{τ 3}=4, the signal to noise ratio of training sequence is ρ _{τ 3}={ 49,199,999}.

3rd step, source 1 and source 2 will estimate the forward channel information G obtained ₁and G ₂feed back to relaying.Relaying is according to the transmitting pre-coding matrix C of all channel information iterative computation sources 1 ₁, F ₁, the transmitting pre-coding matrix C of source 2 ₂, F ₂, relaying pre-coding matrix F _rmatrix W balanced with the reception of source 1 ₁, the balanced matrix W of reception of source 2 ₂.

Iteration one: suppose C ₁, F ₁, C ₂, F ₂, F _rknown, calculate the balanced matrix W of receiving terminal of source 1 ₁matrix W balanced with the receiving terminal of source 2 ₂, method is as follows:

W ₁＝C ₂F ₂ ^H(G ₁F _rH ₂) ^H(G ₁F _rH ₂F ₂F ₂ ^H(G ₁F _rH ₂) ^H+R _n1) ^-1

W ₂＝C ₁F ₁ ^H(G ₂F _rH ₁) ^H(G ₂F _rH ₁F ₁F ₁ ^H(G ₂F _rH ₁) ^H+R _n2) ^-1

R _n1＝G ₁F _rF _r ^HG ₁ ^H+I _N

R _n2＝G ₂F _rF _r ^HG ₂ ^H+I _N

I _nrepresent the unit matrix of N × N, upper table () ^hrepresent conjugate transpose.

Iteration two: suppose Fr, W1, W2 are known, calculate the transmitting pre-coding matrix C of source 1 ₁, F ₁, the transmitting pre-coding matrix C of source 2 ₂, F ₂, method is as follows:

First calculate the transmitting pre-coding matrix C of source 1 ₁, F ₁:

First to (G ₂f _rh ₁) ^h(G ₂f _rf _r ^hg ₂ ^h+ I _n) ^-1(G ₂f _rh ₁) do Eigenvalues Decomposition, obtain:

(G ₂f _rh ₁) ^h(G ₂f _rf _r ^hg ₂ ^h+ I _n) ^-1(G ₂f _rh ₁)=V Λ V ^h, wherein Λ is diagonal matrix, and V is unitary matrix.

Then carry out water injection power distribution to the elements in a main diagonal in diagonal matrix Λ, obtain power division matrix Ω, its elements in a main diagonal is wherein (y) ⁺=max (0, y), σ _sfor transmit signal energy, Λ _ifor i-th diagonal entry of Λ, u is for making Ω _imeet power constraints constant, wherein M ₁for the antenna number of source 1, P _s1for the transmitting power of source 1.

Order $D_1={({\mathrm{\Ω}}^H\mathrm{\Λ\Ω}+{\mathrm{\σ}}_s^{-2}{I}_{M_1})}^{-1/2}$

To D ₁do geometric mean decomposition (GMD), obtain:

D ₁=QRF, wherein R is the upper triangular matrix that diagonal entry is all equal, Q and F is unitary matrix.

Finally obtain transmitting terminal pre-coding matrix C ₁, F ₁, be respectively:

$C_1=\mathrm{diag}\left\{R_{(k,k)}^H\right\}{\left(R^H\right)}^{-1}$

F ₁＝VΩF ^H

Wherein diag{A _{(k, k)}represent the diagonal matrix be made up of all diagonal entries of A.

Adopt similar method, the transmitting pre-coding matrix C of source 2 can be obtained ₂, F ₂.

Iteration three: suppose C ₁, F ₁, C ₂, F ₂, W ₁, W ₂known, calculate relaying pre-coding matrix F _r, method is as follows:

Use Lagrangian Arithmetic, can by F _rbe expressed as:

$F_r=\mathrm{mat}\left\{{[R_{x2}\⊗R_{r1}+R_{x1}\⊗R_{r2}+\mathrm{\λ}{R}_x\⊗I_n]}^{-1}\mathrm{vec}\left(R_r\right)\right\}$

Wherein R _x1=H ₁f ₁f ₁ ^hh ₁ ^h+ I _n

R _x2＝H ₂F ₂F ₂ ^HH ₂ ^H+I _N

R _r1＝G ₁ ^HW ₁ ^HW ₁G ₁

R _r2＝G ₂ ^HW ₂ ^HW ₂G ₂

R _r＝G ₁ ^HW ₁ ^HC ₂F ₂ ^HH ₂ ^H+G ₂ ^HW ₂ ^HC ₁F ₁ ^HH ₁ ^H

R _x＝H ₁F ₁F ₁ ^HH ₁ ^H+H ₂F ₂F ₂ ^HH ₂ ^H+I _N

λ is Lagrange multiplier, $\mathrm{\λ}\∈(0,\sqrt{\mathrm{trace}\left(R_r{\left(R_x\right)}^{-T}{R}_x{\left(R_x\right)}^{-T}{R_r}^H\right)}/P_r),$ P _rfor repeat transmitted power.

for Kronecker computing, vec () is matrixing vector operation, and mat () is the inverse operation of vec (), and trace () is track taking computing, () ^tfor transpose operation.

Dichotomy is used to search in its span λ, until repeat transmitted power equals P _r, namely

trace{F _rR _xF _r ^H}＝P _r

Thus, optimum λ and relaying pre-coding matrix F can be obtained _r.

4th step, the source 1 calculated is launched pre-coding matrix C by relaying ₁, F ₁with the balanced matrix W of reception ₁feed back to source 1, source 2 is launched pre-coding matrix C ₂, F ₂with the balanced matrix W of reception ₂feed back to source 2.Simultaneously by channel information H ₁feed back to source 1, by H ₂feed back to source 2.

5th step, source 1 is to the s that transmits in advance ₁carry out preliminary treatment, obtain the x that transmits ₁, and by this signal x ₁be transmitted to relaying, source 2 is to the s that transmits in advance simultaneously ₂carry out preliminary treatment, obtain the x that transmits ₂, and by this signal x ₂be transmitted to relaying.

Described preliminary treatment is:

x ₁＝F ₁C ₁ ^-1(s ₁+e ₁)

x ₂＝F ₂C ₂ ^-1(s ₂+e ₂)

Wherein s ₁=[s ₁₁..., s _1N1] ^t, s ₂=[s ₂₁..., s _2N2] ^tbe through the signal after m-QAM modulation, N1 is the fluxion that transmits of source 1, and N2 is the fluxion that transmits of source 2.

E ₁can ask with the following method:

$a_{1k}=s_{1k}-\underset{l=1}{\overset{k-1}{\mathrm{\Σ}}}{B}_1(\mathrm{bk},l)a_{1l}$

e ₁＝a ₁-s ₁

Wherein for under round, B ₁=C ₁-I _n1, e ₁=[e ₁₁..., e _1N1] ^t.

E ₂can ask with the following method:

$a_{2k}=s_{2k}-\underset{l=1}{\overset{k-1}{\mathrm{\Σ}}}{B}_2(k,l)a_{2l}$

e ₂＝a ₂-s ₂

Wherein for under round, B ₂=C ₂-I _n2, e ₂=[e ₂₁..., e _2N1] ^t.

6th step, relaying is y to the received signal _rcarry out linear process, obtain signal and will be broadcast to source 1 and source 2.

Described linear process is:

7th step, source 1 y to the received signal ₁carry out check processing, obtain signal signal is obtained again through the modular arithmetic identical with step 5 meanwhile, source 2 y to the received signal ₂carry out check processing, obtain signal signal is obtained again through the modular arithmetic identical with step 5

Described check processing is:

Source 1:

Source 2:

Fig. 1 is that the bit error rate performance of the present embodiment compares schematic diagram, the wherein antenna number M of source 1 and source 2 ₁=4, M ₂=4, the antenna number of relaying is N=4, backward channel SNRs SNR ₁=SNR ₂, forward channel signal to noise ratio is SNR _rand SNR ₁=SNR ₂=SNR _r.H ₁, H ₂, G ₁and G ₂each element all to generate according to CN (0,1) distribution is independent, stochastic generation 10000 secondary channels realize altogether, during every secondary channel realizes, pass 1000 QPSK symbols all mutually.The present embodiment with

The following three kinds of processing methods existed in prior art are done one and are compared:

1. do receiving terminal equilibrium treatment;

2. to do the Combined Treatment of receiving terminal and relaying;

3. combine transmitting terminal, relaying, receiving terminal process, but transmitting terminal adopts linear process mode.Document (Rui Wang Meixia Tao, " Joint Source and Relay Precoding Designs for MIMO Two-wWay Relay Systems; " IEEE ICC, 2011 (" the associating source in MIMO bidirectional relay system and relaying Precoding Designs ", IEEE international communication conference, 2011)).

As can be seen from Figure 1, in low signal-to-noise ratio situation, the bit error rate performance of the present embodiment and transmitting terminal adopt linear signal processing mode to be more or less the same.In high s/n ratio situation, the present embodiment effectively improves the bit error rate performance of system, and along with signal to noise ratio increase, this performance advantage is also more obvious.

Claims (7)

1. the united signal processing method of source and relay in bidirectional relay system, is characterized in that, comprise the following steps:

The first step, source 1 is to repeat transmitted training sequence S ₁, relaying is according to the signal X received ₁carry out backward channel estimating, obtain the backward channel H of source 1 and relay well ₁, source 2 is to repeat transmitted training sequence S simultaneously ₂, relaying is according to the signal X received ₂carry out backward channel estimating, obtain the backward channel H of source 2 and relay well ₂;

Second step, relaying is simultaneously to source 1 and source 2 transmitting training sequence S ₃, source 1 is according to the signal X received ₃carry out forward channel estimation, obtain the forward channel G between relaying and source 1 ₁, source 2 is according to the signal X received ₄carry out forward channel estimation, obtain the backward channel G between relaying and source 2 ₂;

3rd step, source 1 and source 2 will estimate the forward channel information G obtained ₁and G ₂feed back to relaying, relaying is according to the transmitting nonlinear precoding Matrix C of all channel information iterative computation sources 1 ₁, F ₁, the transmitting nonlinear precoding Matrix C of source 2 ₂, F ₂, relaying Linear precoding matrix F _rmatrix W balanced with the reception of source 1 ₁, the balanced matrix W of reception of source 2 ₂;

4th step, the source 1 calculated is launched nonlinear precoding Matrix C by relaying ₁, F ₁with the balanced matrix W of reception ₁feed back to source 1, source 2 is launched nonlinear precoding Matrix C ₂, F ₂with the balanced matrix W of reception ₂feed back to source 2; Simultaneously by channel information H ₁feed back to source 1, by H ₂feed back to source 2;

5th step, source 1 is to the s that transmits in advance ₁carry out preliminary treatment, obtain the x that transmits ₁, and by this signal x ₁be transmitted to relaying, source 2 is to the s that transmits in advance simultaneously ₂carry out preliminary treatment, obtain the x that transmits ₂, and by this signal x ₂be transmitted to relaying;

6th step, relaying is y to the received signal _rcarry out linear process, obtain signal y _r, and by y _rsend to source 1 and source 2;

7th step, source 1 y to the received signal ₁carry out check processing, obtain signal a ₁, then obtain signal s through the modular arithmetic identical with step 5 ₁, meanwhile, source 2 y to the received signal ₂carry out check processing, obtain signal a ₂, then obtain signal s through the modular arithmetic identical with step 5 ₂.

2. the united signal processing method of source and relay in bidirectional relay system according to claim 1, it is characterized in that, the backward channel estimation process in the first step is:

Wherein: M ₁the antenna number of source 1, M ₂the antenna number of source 2, ρ _{τ 1}training sequence S ₁signal to noise ratio, ρ _{τ 2}training sequence S ₂signal to noise ratio, t _{τ 1}the length of the training sequence that source 1 is launched, t _{τ 2}the length of the training sequence that source 2 is launched, with be the signal that relay reception arrives, N is the antenna number of relaying, be respectively M ₁× M ₁and M ₂× M ₂unit matrix.

3. the united signal processing method of source and relay in bidirectional relay system according to claim 1, it is characterized in that, the forward channel estimating processing method in described second step is:

Wherein: N is the antenna number of relaying, ρ _{τ 3}training sequence S ₃signal to noise ratio, t _{τ 3}the length of the training sequence of repeat transmitted, with be respectively the signal that source 1 and source 2 receive, I _nfor the unit matrix of N × N.

4. the united signal processing method of source and relay in bidirectional relay system according to claim 1, it is characterized in that, the iterative operation method in described 3rd step is:

Iteration one: suppose C ₁, F ₁, C ₂, F ₂, F _rknown, calculate the balanced matrix W of receiving terminal of source 1 ₁matrix W balanced with the receiving terminal of source 2 ₂, method is as follows:

W ₁＝C ₂F ₂ ^H(G ₁F _rH ₂) ^H(G ₁F _rH ₂F ₂F ₂ ^H(G ₁F _rH ₂) ^H+R _n1) ^-1

W ₂＝C ₁F ₁ ^H(G ₂F _rH ₁) ^H(G ₂F _rH ₁F ₁F ₁ ^H(G ₂F _rH ₁) ^H+R _n2) ^-1

R _n1＝G ₁F _rF _r ^HG ₁ ^H+I _N

R _n2＝G ₂F _rF _r ^HG ₂ ^H+I _N

I _nrepresent the unit matrix of N × N, upper table () ^hrepresent conjugate transpose;

Iteration two: suppose Fr, W1, W2 are known, calculate the transmitting nonlinear precoding Matrix C of source 1 ₁, F ₁, the transmitting nonlinear precoding Matrix C of source 2 ₂, F ₂, method is as follows:

First calculate the transmitting nonlinear precoding Matrix C of source 1 ₁, F ₁:

First to (G ₂f _rh ₁) ^h(G ₂f _rf _r ^hg ₂ ^h+ I _n) ^-1(G ₂f _rh ₁) do Eigenvalues Decomposition, obtain:

(G ₂f _rh ₁) ^h(G ₂f _rf _r ^hg ₂ ^h+ I _n) ^-1(G ₂f _rh ₁)=V Λ V ^h, wherein Λ is diagonal matrix, and V is unitary matrix;

Then carry out water injection power distribution to the elements in a main diagonal in diagonal matrix Λ, obtain power division matrix Ω, its elements in a main diagonal is wherein (y) ⁺=max (0, y), σ _sfor transmit signal energy, Λ _ifor i-th diagonal entry of Λ, u is for making Ω _imeet power constraints constant, wherein M ₁for the antenna number of source 1, P _s1for the transmitting power of source 1;

Order

To D ₁do geometric mean decomposition (GMD), obtain:

D ₁=QRF, wherein R is the upper triangular matrix that diagonal entry is all equal, Q and F is unitary matrix;

Finally obtain launching nonlinear precoding Matrix C ₁, F ₁, be respectively:

F ₁＝VΩF ^H

Wherein diag{A _{(k, k)}represent the diagonal matrix be made up of all diagonal entries of A;

Adopt similar method, the transmitting nonlinear precoding Matrix C of source 2 can be obtained ₂, F ₂;

Iteration three: suppose C ₁, F ₁, C ₂, F ₂, W ₁, W ₂known, calculate relaying pre-coding matrix F _r, method is as follows:

Use Lagrangian Arithmetic, can by F _rbe expressed as:

Wherein R _x1=H ₁f ₁f ₁ ^hh ₁ ^h+ I _n

R _x2＝H ₂F ₂F ₂ ^HH ₂ ^H+I _N

R _r1＝G ₁ ^HW ₁ ^HW ₁G ₁

R _r2＝G ₂ ^HW ₂ ^HW ₂G ₂

R _r＝G ₁ ^HW ₁ ^HC ₂F ₂ ^HH ₂ ^H+G ₂ ^HW ₂ ^HC ₁F ₁ ^HH ₁ ^H

R _x＝H ₁F ₁F ₁ ^HH ₁ ^H+H ₂F ₂F ₂ ^HH ₂ ^H+I _N

λ is Lagrange multiplier, p _rfor repeat transmitted power;

for Kronecker computing, vec () is matrixing vector operation, and mat () is the inverse operation of vec (), and trace () is track taking computing, () ^tfor transpose operation;

Dichotomy is used to search in its span λ, until repeat transmitted power equals P _r, namely

trace{F _rR _xF _r ^H}＝P _r

Thus, optimum λ and relaying pre-coding matrix F can be obtained _r.

5. the united signal processing method of source and relay in bidirectional relay system according to claim 1, it is characterized in that, the preprocess method in described 5th step is:

Wherein s ₁=[s ₁₁..., s _1N1] ^t, s ₂=[s ₂₁..., s _2N2] ^tbe through the signal after m-QAM modulation, N1 is the fluxion that transmits of source 1, and N2 is the fluxion that transmits of source 2;

E ₁can ask with the following method:

e ₁＝a ₁-s ₁

Wherein for under round, B ₁=C ₁-I _n1, e ₁=[e ₁₁..., e _1N1] ^t;

E ₂can ask with the following method:

e ₂＝a ₂-s ₂

Wherein for under round, B ₂=C ₂-I _n2, e ₂=[e ₂₁..., e _2N1] ^t.

6. the united signal processing method of source and relay in bidirectional relay system according to claim 1, it is characterized in that, the linear processing methods in described 6th step is:

y _r＝F _ry _r。

7. the united signal processing method of source and relay in bidirectional relay system according to claim 1, it is characterized in that, the check processing method in described 7th step is:

Source 1:a ₁=W ₁(y ₁-G ₁f _rh ₁x ₁)

Source 2:a ₂=W ₂(y ₂-G ₂f _rh ₂x ₂).