CN106027206A - search method and communication method for calculating and forwarding coding coefficient vector in bidirectional cooperative relay channel - Google Patents

Abstract

The invention discloses a search method and a communication method for calculating and forwarding a coding coefficient vector in a bidirectional cooperative relay channel. In view of the characteristics of a working mode of calculating a forwarding code in the bidirectional cooperative relay channel, the method is used for modeling a coding coefficient vector optimization problem of a relay node into an integer quadratic programming model with quadratic constraints; and aiming at the characteristics of the optimization problem, the original optimization problem is converted into a new relaxation planning problem that can be resolved more easily by improvement, convex relaxation, cutting plane generation and other steps, and the relaxation planning problem is resolved to effectively acquire the optimal solution of the original problem. The selection of the coding coefficient vector has important influence on network reachability indexes, and by adopting the method provided by the invention, an optimal coefficient vector combination under the current channel state can be effectively acquired, and a good optimization foundation is laid for the application of calculating, forwarding and coding.

Description

A kind of two-way cooperating relay channel calculation forwards code coefficient vector search method and leads to Letter method

Technical field

The invention belongs to wireless communication technology field, relate to the cooperative communication technology under physical-layer network coding, it is proposed that A kind of two-way cooperating relay channel calculation forwards code coefficient vector search method.

Background technology

In wireless communication system, launch signal and be transmitted in physical layer by electromagnetic wave.At wireless multi-source communications network In, the broadcast characteristic of transmitting node makes receiving node may receive in same time slot to come from multiple not source node Transmitting information, this will cause interfering between different transmission signal, thus affect whole network performance.Therefore, receiving terminal The problem that interferes between multiple reception signal that the most effectively processes is a significant challenge of wireless communication technology research.

In recent years, linear network encoding technology cable network apply in achieved with the achievement in research attracted people's attention.Net Network coding has stronger compatibility and information extraction ability, and this makes the interference problem solving between above-mentioned multiple user signals become For possible.Traditional network coding scheme mostly operates in MAC layer, in order to reduce existing wireless communications system hardware and software equipment With the corresponding amendment of agreement, general employing MAC layer resource and user scheduling algorithm reduce interference as far as possible.But sending multi-source data Time, traditional network coding method is the most inefficient.In the wireless network, the broadcast characteristic of transmitting node is the most effectively utilized Promote radio channel capacity and seem more important.

Calculating based on nested Lattice forwards network coding scheme, not only can solve via node under high order modulation The decoding problem at place, and can be close to AWGN bidirectional relay channel capacity.After the construction features of Lattice coding makes superposition Signal phasor remains a code word, and via node only need to decode the linear combination of each code word.Destination node is each by obtaining The linear combination information that via node forwards, gets final product the transmission information of efficient decoding source node.

Summary of the invention

Goal of the invention: in order to overcome the deficiencies in the prior art, the present invention provides a kind of two-way cooperating relay channel Calculating and forward code coefficient vector search method and communication means, the present invention can effectively obtain the optimum under current channel condition Coefficient vector combines, and provides good optimization basis for calculating the application forwarding coding.

Technical scheme: for achieving the above object, the technical solution used in the present invention is:

A kind of two-way cooperating relay channel calculation forwards code coefficient vector search method, at two-way cooperating relay channel meter Calculate and forward in encoding scheme, forward code coefficient Vector Optimization Problems to be modeled as with quadratic constraints calculating of via node Integer quadratic programming model.Forwarding code coefficient vector according to via node introduces auxiliary variable by this integer quadratic programming mould In type quadratic term rise to new higher dimensional space.By convexification and relaxation, the whole of new higher dimensional space will be risen to Number quadratic programming model is converted into new relaxation problem, solves this relaxation problem, if the optimal solution obtained meets integer requirement, It it is then the optimal solution of former problem.Otherwise, according to introducing auxiliary variable and the convexification of previous step and loose constraint conditional definition one Cutting plane constraint condition, increases in the constraint set of relaxation problem, and the feasible solution required to cut away a part to be unsatisfactory for, reducing can Row territory, then, solves new lax planning problem.Repeat above procedure, until obtaining integer optimal solution.

Specifically include following steps:

Step 1: the calculating at two-way cooperating relay channel forwards in coding strategy, obtains the coefficient vector a of via node Optimization aim letter.

Step 2: introduce auxiliary variable A according to the coefficient vector a of via node_ijExcellent by the coefficient vector a of via node The quadratic term changing object function rises to new higher dimensional space, obtains the optimization aim letter of the coefficient vector a of new via node Number.

Step 3: use PSD relaxation method that the non-convex constraint in the problem of step 2 Central Plains is carried out convexification and relax, by step 2 The optimization object function of the coefficient vector a of the new via node obtained is rewritten as relaxation problem.

Step 4: the relaxation problem in solution procedure 3, if gained optimal solution meets integer requirement, is then the optimum of former problem Solve, and jump procedure 6.Otherwise jump to step 4.

Step 5: according to introducing auxiliary variable and the convexification of previous step in the constraint set of the relaxation problem obtained in step 3 Cutting plane constraint condition is increased with loose constraint condition, the feasible solution required to cut away a part to be unsatisfactory for, reduces feasible zone, so After, solve new lax planning problem.

Step 6: return, exports optimal solution.

The optimization object function of the coefficient vector a of the via node obtained in described step 1 is as follows:

A=arg min (a^TGa)

s.t.||a||²≤1+P||h||²

a₁≠0,a₂≠0

Wherein, a represents the coefficient vector of via node, the integer quotient of namely network code linear combination,a₁Represent node S₁The integer quotient of network code, a₂Represent node S₂The integer quotient of network code, Representing matrix transposition,P represents about beam power, h=[h₁,h₂]^T,For node S₁、S₂In with Real valued channel gain between continuing, I representation unit matrix.

The optimization object function of the coefficient vector a of the new via node in described step 2 is:

A=arg min<G, A>

And

Wherein, I is the unit matrix of 2 × 2,It is the symmetrical matrix set of 2 × 2, b=1+P | | h | |².Symmetrical matrix A Can promote and be expressed as A=aa^T.<A, B>represents the Frobenius inner product of symmetrical matrix A Yu B, i.e. tr (A^TB).In the most former problem Quadratic term a^TGa is represented by < G, aa^T>。

The relaxation problem obtained in described step 3 is:

$a=\arg min<\tilde{G},\tilde{A}>$

$\begin{array}{cc}s.t.& <\widetilde{G_1},\tilde{A}>\&le;0\end{array}$

$v^T\tilde{A}v\&GreaterEqual;0$

Wherein, A >=0 represents that A is symmetric positive semidefinite matrix.

V represents any real number vector.

The Cutting plane constraint condition obtained in described step 5 is expressed as:

$v_k^T\tilde{A}{v}_k\&GreaterEqual;0.$

In, v_kRepresent respectivelyEigenvalue characteristic of correspondence vector,RepresentThe variable space in any point, K=1,2.

A kind of two-way cooperating relay channel communication method, comprises the following steps:

Step (1), forwards in encoding scheme in two-way cooperating relay channel calculation, source node by respective information from limited The nested Lattice code word of domain mapping to.

Step (2), the codeword information after source node will map simultaneously sends to via node.

Step (3), via node receives the composite signal from each source node, and the information received is decoded into The linear combination equation of Lattice code word.

Step (4), relay node broadcasts Lattice equation is to the first source node S₁With the second source node S₂, each source node will Lattice code maps back finite field, utilizes the information self stored to complete decoding.

Beneficial effect: the one two-way cooperating relay channel calculation that the present invention provides forwards code coefficient vector search method And communication means, compared to existing technology, have the advantages that

The present invention is directed to two-way cooperating relay channel calculate the feature forwarding coding work mode, by the volume of via node Code coefficient vector optimization problem is modeled as the integer quadratic programming model with quadratic constraints；For the feature of this optimization problem, By steps such as lifting, convex lax, generation cutting planes, former optimization problem is converted into the new lax planning problem being easier to solve, leads to Cross and solve lax planning problem effectively to obtain the optimal solution of former problem.Choosing network reachability of code coefficient vector Can index important；Through simulating, verifying, the method that the present invention proposes can effectively obtain under current channel condition Major clique number vector combines, and provides good optimization basis for calculating the application forwarding coding.

Accompanying drawing explanation

Fig. 1 is that in the present invention, bidirectional relay channel calculates forwarding strategy block diagram；

Fig. 2 is that in the present invention, code coefficient is vectorial on the impact up to computation rate；

Fig. 3 is that in the present invention, the computation rate of different coefficient vector search plans compares；

When Fig. 4 is P=10dB in the present invention, channel coefficients is on the impact up to computation rate；

When Fig. 5 is P=20dB in the present invention, channel coefficients is on the impact up to computation rate；

When Fig. 6 is P=30dB in the present invention, channel coefficients is on the impact up to computation rate；

Fig. 7 is the schematic diagram of two-way cooperating relay channel communication method.

Detailed description of the invention

Below in conjunction with the accompanying drawings and specific embodiment, it is further elucidated with the present invention, it should be understood that these examples are merely to illustrate this Invention rather than limit the scope of the present invention, after having read the present invention, various to the present invention of those skilled in the art The amendment of the equivalent form of value all falls within the application claims limited range.

A kind of two-way cooperating relay channel calculation forwards code coefficient vector search method, as it is shown in figure 1, in two-way cooperation Trunk channel calculates and forwards in encoding scheme, the calculating of via node is forwarded code coefficient Vector Optimization Problems be modeled as with The integer quadratic programming model of quadratic constraints.Forwarding code coefficient vector according to via node introduces auxiliary variable by this integer In quadratic programming model quadratic term rise to new higher dimensional space.By convexification and relaxation, new height will be risen to The integer quadratic programming model of dimension space is converted into new relaxation problem, solves this relaxation problem, if the optimal solution obtained is full Foot integer requirement, then be the optimal solution of former problem.Otherwise, according to introducing auxiliary variable and the convexification of previous step and loose constraint bar Part definition one Cutting plane constraint condition, increase in the constraint set of relaxation problem, with cut away a part be unsatisfactory for require can Row solves, and reduces feasible zone, then, solves new lax planning problem.Repeat above procedure, until obtaining integer optimal solution.

As it is shown in figure 1, the reception signal that multiple access access phase is obtained by via nodeThe estimating of linear combination equation Evaluation is broadcasted to each source node.Each source node passes through decoderReception code word is mapped back finite field, deducts certainly The information of body storage, can recover required estimated information

Specifically include following steps:

Step 1: the calculating at two-way cooperating relay channel forwards in coding strategy, obtains the coefficient vector a of via node Optimization aim letter, the linear equation coefficient vector that via node recovers is different, and the most each node is the most different up to computation rate, for Maximization power system capacity, forwards code coefficient Vector Optimization Problems to be modeled as with quadratic constraints calculating of via node Integer quadratic programming model.Wherein, the optimization object function of the coefficient vector a of via node is as follows:

A=argmin (a^TGa)

Wherein, a represents the coefficient vector of via node, the integer quotient of namely network code linear combination,a₁Represent node S₁The integer quotient of network code, a₂Represent node S₁The integer quotient of network code, Representing matrix transposition,P represents about beam power, h=[h₁,h₂]^T,For node S₁、S₂In with Real valued channel gain between continuing, I representation unit matrix.

Step 2: introduce auxiliary variable A according to the coefficient vector a of via node_ijExcellent by the coefficient vector a of via node The quadratic term changing target function type (1) rises to new higher dimensional space,.Make A_ij=a_ia_j(1≤i, j≤2), symmetrical matrix A can Lifting is expressed as A=aa^T,<A, B>represents the Frobenius inner product of symmetrical matrix A Yu B, i.e. tr (A^TB).The most former problem (formula (1) the quadratic term a in)^TGa is represented by < G, aa^T>.Therefore the optimization aim letter of the coefficient vector a of new via node is obtained Number:

And

Wherein, I is the unit matrix of 2 × 2,It is the symmetrical matrix set of 2 × 2, b=1+P | | h | |².Symmetrical matrix A Can promote and be expressed as A=aa^T.<A, B>represents the Frobenius inner product of symmetrical matrix A Yu B, i.e. tr (A^TB).In the most former problem Quadratic term a^TGa is represented by < G, aa^T>。

From above formula (2), the minimization problem after lifting by original quadratic function and quadratic constraints be changed into about The linear integer optimization problem of (a, A), reduces and solves difficulty.

Step 3: use PSD relaxation method that the non-convex constraint in the problem of step 2 Central Plains is carried out convexification and relax, by step 2 The optimization object function of the coefficient vector a of the new via node obtained is rewritten as relaxation problem.

In former problem formula (1), quadratic constraints is convex constraint.But constraints A=aa introduced in lifting process^TNon- Convex, use suitable relaxation method to make optimization problem convexification.Former problem formula (1) can be written as following equivalents

Wherein,RepresentConvex closure network.

Do not consider A=aa^TWithTwo constraintss, then problem constraint can be expressed as follows:

A is symmetric positive semidefinite matrix, by A=aa to make A >=0 represent^TA-aa can be obtained^T>=0, therefore can adopt on this basis With PSD relaxation method, the non-convex constraint in former problem is carried out convexification to relax.

ForMeet:

Therefore, linear positive semidefinite inequalityIt is applicable toPSD constraint definition is as follows:

$PSD=\{(a,A):\tilde{A}\&GreaterEqual;0\}$

After increasing PSD loose constraint,

Order

$\tilde{G}=\left(\begin{array}{cc}0& 0\\ 0& G\end{array}\right),\widetilde{G_1}=\left(\begin{array}{cc}-b& 0\\ 0& I\end{array}\right)$

Target function type (2) after then relaxing is represented by:

$a=\arg min<\tilde{G},\tilde{A}>$

$\begin{array}{cc}s.t.& <\widetilde{G_1},\tilde{A}>\&le;0\\ & \tilde{A}\&GreaterEqual;0\end{array}---\left(3\right)$

Due toFor positive semidefinite matrix, it may be assumed that

PSD constraint in formula (3) can be replaced with following form (relaxation problem):

$a=\arg min<\tilde{G},\tilde{A}>$

$\begin{array}{cc}s.t.& <\widetilde{G_1},\tilde{A}>\&le;0\end{array}$

$v^T\tilde{A}v\&GreaterEqual;0$

Wherein, v represents any real number vector.

Step 4: the relaxation problem in solution procedure 3, if gained optimal solution meets integer requirement, is then the optimum of former problem Solve, and jump procedure 6.Otherwise jump to step 4.

Step 5: according to introducing auxiliary variable and the convexification of previous step in the constraint set of the relaxation problem obtained in step 3 Cutting plane constraint condition is increased with loose constraint condition, the feasible solution required to cut away a part to be unsatisfactory for, reduces feasible zone, so After, solve new lax planning problem.

OrderRepresentThe variable space in any point.Can pass throughFeature decomposition differentiateWhether position In the constraint of PSD circular cone (PSD cone).Make λ_kAnd v_kRepresent respectivelyEigenvalue and characteristic of correspondence vector, k =1,2.For without loss of generality, it is assumed that λ₁≤λ₂, λ_t<0≤λ_t+1, t ∈ 0,1,2.If t=0, illustrate that all eigenvalues are non- It is negative,For positive semidefinite matrix；If t ≠ 0, thenK=1 ..., t, it is impossible to meetFor positive definite The requirement of matrix.

Therefore, can increase in original constraint set using following formula as Cutting plane constraint condition, solve new lax planning Problem.

$v_k^T\tilde{A}{v}_k\&GreaterEqual;0---\left(4\right)$

Wherein, v_kRepresent respectivelyEigenvalue characteristic of correspondence vector,RepresentThe variable space in any one Point, k=1,2.

Step 6: return, exports optimal solution.

A kind of two-way cooperating relay channel communication method, as it is shown in fig. 7, comprises following steps:

Step (1), forwards in encoding scheme in two-way cooperating relay channel calculation, source node by respective information from limited The nested Lattice code word of domain mapping to.

Step (2), the codeword information after source node will map simultaneously sends to via node.

Step (3), via node receives the composite signal from each source node, and the information received is decoded into The linear combination equation of Lattice code word.

Step (4), relay node broadcasts Lattice equation is to the first source node S₁With the second source node S₂, each source node will Lattice code maps back finite field, utilizes the information self stored to complete decoding.

Fig. 2 is that in the present invention, code coefficient is vectorial on the impact up to computation rate.Code word linear combination coefficient vector Search problem is the key issue calculating and forwarding coding strategy.Under conditions of channel coefficients is fixing, different code coefficients to Amount will produce material impact to network achievable rate index.Assume two-way cooperative relay network each source node transmission rate pair Claim, i.e. R₁=R₂.Signaling channel coefficient vector h=[1, h₂]^T, h₂∈ [0,1], power constraint P are 10dB.Different code word linear combinations Coefficient vector on the impact of information rate as shown in Figure 2.It is respectively compared a=[0,1] herein^T, a=[-1 ,-1]^T, a=[1, 2]^TIn the case of three kinds of coefficient vectors up to computation rate situation.As can be seen from Figure, up to computation rate and code coefficient Choosing of vector is closely related.When code coefficient vector matches with current channel coefficient, network is up to computation rate ability Realize maximizing.Work as h₂When=1, a=[-1 ,-1]^TOptimal with the matching status of channel coefficients, therefore it is maximum up to computation rate. Therefore, in order to ensure the maximization of system-computed speed, need code coefficient vector is made optimal choice.

Fig. 3 is the computation rate comparable situation of different coefficient vector search plans in the present invention.Use Monte-Carlo side Method, carries out 1000 random experiments, and acquired results is carried out statistical average.Node S₁、S₂Real valued channel gain with relay well h₁,h₂Separate and obeyDistribution, the excursion of power constraint P is 0dB to 30dB.Figure 3 show institute herein Extracting method, up to the computation rate upper bound, poor search algorithm and without coefficient vector search several different schemes of PNC under up to calculate speed Rate comparable situation.Calculate being represented by up to the computation rate upper bound of forwarding strategy Poor search algorithm, by searching all integer combinations of code coefficient vector in restriction range thoroughly, meets the solution that current computation rate is maximum It is the optimal coefficient vector under current channel condition.LRCP algorithmic notation proposed coefficient vector searching algorithm.Nothing Code coefficient vector search PNC schema definition is that via node does not carries out coefficient vector search and directly decodes and to each source node Forwarding information x₁+x₂.It can be seen that different coding strategy gained is up to the obvious difference of computation rate.Owing to searching calculation thoroughly The solution that method obtains is optimal coefficient vector, therefore closest to the theoretical upper bound.LRCP algorithm acquired results is consistent with poor search algorithm, table Understand that algorithm presented here can effectively obtain the optimal coefficient vector combination of network.Without coefficient vector search PNC scheme due to The change for channel coefficients does not carries out the corresponding adjustment of coefficient vector, and therefore performance is worst, deposits under the conditions of high s/n ratio In bigger performance gap.

Fig. 4,5,6 sets forth power constraint P when being respectively 10dB, 20dB, 30dB, and channel coefficients changes up to meter Calculate the impact of speed.Signaling channel coefficient vector h=[1, h₂]^T, h₂∈ [0,1], first passes through institute's extracting method herein and obtains major clique Number vector, then calculate under optimal coefficient vector up to computation rate.As can be seen from Figure, under Low SNR, Institute's extracting method, poor search algorithm and the performance without coefficient vector search PNC scheme are closer to herein, and all close to up to calculating speed The upper bound of rate.But along with the increase of signal to noise ratio, the performance of institute's extracting method and poor search algorithm will be significantly better than without coefficient vector herein Search PNC scheme.It should be noted that without code coefficient vector search PNC scheme acquiescence code coefficient vector be a=[1, 1]^T, work as h₂When >=0.8, the matching degree of the program and channel coefficients improves, therefore its performance and this paper institute's extracting method and search thoroughly Algorithm is suitable.Meanwhile, h is worked as₂When=0.5, institute's extracting method can reach the theoretical upper bound of computation rate herein.

Above-mentioned simulation result shows, what code coefficient was vectorial chooses network reachability energy index important, this Literary composition institute extracting method can effectively obtain the optimal coefficient vector combination under current channel condition, provides for calculating the application forwarding coding Good optimization basis.

The above is only the preferred embodiment of the present invention, it should be pointed out that: for the ordinary skill people of the art For Yuan, under the premise without departing from the principles of the invention, it is also possible to make some improvements and modifications, these improvements and modifications also should It is considered as protection scope of the present invention.

Claims (7)

1. a two-way cooperating relay channel calculation forwards code coefficient vector search method, it is characterised in that: in two-way cooperation Trunk channel calculates and forwards in encoding scheme, the calculating of via node is forwarded code coefficient Vector Optimization Problems be modeled as with The integer quadratic programming model of quadratic constraints；Forwarding code coefficient vector according to via node introduces auxiliary variable by this integer In quadratic programming model quadratic term rise to new higher dimensional space；By convexification and relaxation, new height will be risen to The integer quadratic programming model of dimension space is converted into new relaxation problem, solves this relaxation problem, if the optimal solution obtained is full Foot integer requirement, then be the optimal solution of former problem；Otherwise, according to introducing auxiliary variable and the convexification of previous step and loose constraint bar Part definition one Cutting plane constraint condition, increase in the constraint set of relaxation problem, with cut away a part be unsatisfactory for require can Row solves, and reduces feasible zone, then, solves new lax planning problem；Repeat above procedure, until obtaining integer optimal solution.

Two-way cooperating relay channel calculation the most according to claim 1 forwards code coefficient vector search method, its feature It is, comprises the following steps:

Step 1: the calculating at two-way cooperating relay channel forwards in coding strategy, the coefficient vector a's of acquisition via node is excellent Change target letter；

Step 2: introduce auxiliary variable A according to the coefficient vector a of via node_ijOptimization mesh by the coefficient vector a of via node The quadratic term of scalar functions rises to new higher dimensional space, obtains the optimization object function of the coefficient vector a of new via node；

Step 3: use PSD relaxation method that the non-convex constraint in the problem of step 2 Central Plains is carried out convexification and relax, step 2 is obtained The optimization object function of coefficient vector a of new via node be rewritten as relaxation problem；

Step 4: the relaxation problem in solution procedure 3, if gained optimal solution meets integer requirement, is then the optimal solution of former problem, And jump procedure 6；Otherwise jump to step 4；

Step 5: according to introducing auxiliary variable and the convexification of previous step and pine in the constraint set of the relaxation problem obtained in step 3 Relaxation constraints increases Cutting plane constraint condition, the feasible solution required to cut away a part to be unsatisfactory for, and reduces feasible zone, then, Solve new lax planning problem；

Step 6: return, exports optimal solution.

Two-way cooperating relay channel calculation the most according to claim 1 forwards code coefficient vector search method, its feature It is: the optimization object function of the coefficient vector a of the via node obtained in described step 1 is as follows:

A=arg min (a^TGa)

s.t. ‖a‖²≤1+P‖h‖²

a₁≠0,a₂≠0

Wherein, a represents the network code integer quotient vector of via node,a₁Represent node S₁Network is compiled The integer quotient of code, a₂Represent node S₂The integer quotient of network code,Representing matrix transposition,P table Show about beam power, h=[h₁,h₂]^T,For node S₁、S₂With the real valued channel gain of relay well, I representation unit matrix.

Two-way cooperating relay channel calculation the most according to claim 1 forwards code coefficient vector search method, its feature It is: the optimization object function of the coefficient vector a of the new via node in described step 2 is:

A=arg min ＜ G, A ＞

And

Wherein, I is the unit matrix of 2 × 2,It is the symmetrical matrix set of 2 × 2, b=1+P ‖ h ‖²；Symmetrical matrix A can promote It is expressed as A=aa^T；＜ A, B ＞ represents the Frobenius inner product of symmetrical matrix A Yu B, i.e. tr (A^TB)；

Quadratic term a in the most former problem^TGa is represented by ＜ G, aa^T>。

Two-way cooperating relay channel calculation the most according to claim 1 forwards code coefficient vector search method, its feature It is: the relaxation problem obtained in described step 3 is:

$a=\mathrm{argmin}<\tilde{G},\tilde{A}>$

$\begin{array}{cc}s.t.& <{\tilde{G}}_1,\tilde{A}>\&le;0\end{array}$

$v^T\tilde{A}v\&GreaterEqual;0$

Wherein,Expression A is symmetric positive semidefinite matrix；

V represents any real number vector.

Two-way cooperating relay channel calculation the most according to claim 1 forwards code coefficient vector search method, its feature It is: the Cutting plane constraint condition obtained in described step 5 is expressed as:

$v_k^T\tilde{A}{v}_k\&GreaterEqual;0;$

Wherein, v_kRepresent respectivelyEigenvalue characteristic of correspondence vector,RepresentThe variable space in any point, k =1,2.

7. one kind forwards code coefficient vector search based on the arbitrary described two-way cooperating relay channel calculation of claim 1 to 6 The two-way cooperating relay channel communication method of method, it is characterised in that comprise the following steps:

Step (1), in two-way cooperating relay channel calculation forwards encoding scheme, respective information is reflected by source node from finite field It is mapped to a nested Lattice code word；；

Step (2), the codeword information after source node will map simultaneously sends to via node；

Step (3), via node receives the composite signal from each source node, and the information received is decoded into The linear combination equation of Lattice code word；

Step (4), relay node broadcasts Lattice equation is to the first source node S₁With the second source node S₂, each source node will Lattice code maps back finite field, utilizes the information self stored to complete decoding.