CN104022987A - Interference elimination method in MIMO alternating relay system on basis of decoding forwarding - Google Patents

Abstract

The invention discloses an interference elimination method in an MIMO alternating relay system on the basis of decoding forwarding. The method mainly solves the problem of interference among relays of the MIMO system. The method comprises the realizing steps that (1), the system is set; (2), according to the set system, vectors of mixed signals sent to the relays from source nodes at different time slots and processed through a precoding matrix are constructed; (3), the relays receive the signals sent from the source nodes and decode the signals; (4), the precoding matrix is designed; (5) target nodes receive the vectors of interference-free signals sent by the relays. According to the method, the limitation to parity of the number of antennas of the system nodes can be avoided, the system complexity is reduced, and the maximum freedom degree of the system can be achieved.

Description

MIMO replaces the interference elimination method forwarding based on decoding in relay system

Technical field

The invention belongs to communication technical field, particularly a kind of relay well disturbs the method for eliminating, and can be used for multiple-input and multiple-output MIMO and replaces relay system.

Background technology

Multiple-input and multiple-output mimo system can improve spectrum efficiency.In addition, relaying auxiliary transmission has the ability of expanding the coverage area and space diversity being provided.Therefore the hybrid system, being made up of multiple-input and multiple-output MIMO and relaying attracts more people to carry out deep research.

In the hybrid system of multiple-input and multiple-output MIMO and relaying composition, relaying is divided into full duplex relaying and half-duplex relaying.Full duplex relaying can carry out transmitted signal simultaneously and receive signal, and half-duplex relay system is relaying at the same time can only transmitted signal or receive signal.Because full duplex relay system implements more difficultly, thereby half-duplex relaying obtains broader applications.But in half-duplex relay system, in the situation that signal to noise ratio is higher, the capacity loss of system is larger.A lot of schemes have all proposed to recover the method for capacity loss, and in these schemes, alternately trunking plan has attracted more people to study.It can make two relayings forward successively from transmitting terminal to receiving terminal.But for trunking plan alternately, an intrinsic shortcoming is to exist relay well to disturb IRI, it has reduced the performance of system to a great extent.

For multiple-input and multiple-output, MIMO replaces relay system, someone has proposed to disturb the scheme of alignment IA, at signal receiving end, interference signal is snapped to an interference space, just can be used for receiving useful signal with the orthogonal subspace of interference space like this.But thisly use the deficiency based on disturbing alignment IA scheme to be:

1. system can only obtain the degree of freedom of 3/4M, and wherein, M is the antenna number of each node in system and can only is even number;

2. must configure three repeaters and just can complete interference alignment IA, system arranges complexity.

Summary of the invention

The object of the invention is to the deficiency for above-mentioned prior art, propose a kind of MIMO and replaced the interference elimination method forwarding based on decoding in relay system, to avoid the parity restriction of the node antenna number to system, reduce system complexity, the maximum degree of freedom M of the system that reaches.

Realizing technical scheme of the present invention is: replace at MIMO the pre-coding matrix that source node in relay system sends mixed signal and designs respectively cascade at source node place and relaying place, make relay well disturb IRI to eliminate completely, to realize the maximum degree of freedom and the low-complexity of system.Implementation step comprises as follows:

1) system setting:

If mimo system comprises a source node S, a destination node D and two relaying R ₁, R ₂, they all configure M root antenna, M >=2, and the transmission means of relaying is half-duplex;

While making time slot be odd number, source node S and the second relaying R ₂transmitted signal, simultaneously destination node D and the first relaying R ₁receive signal;

While making time slot be even number, source node S and the first relaying R ₁transmitted signal, simultaneously destination node D and the second relaying R ₂receive signal;

2) build the mixed signal vector that source node S sends at different time-gap:

2.1), at first time slot, the signal indication that source node S is sent is s ₁:

s ₁＝x ₁＝[x ₁₁，?x ₁₂，?…?x _1i，?…?x _1M] ^T，

Wherein, x _1ii the signal component that source node S sends at first time slot, i=1,2 ... M, T representing matrix transposition;

2.2), at second time slot, the mixed signal vector representation that source node S is sent is s ₂:

s ₂＝A ₁B ₁(x ₂+x ₁)，

Wherein, A ₁b ₁the cascade pre-coding matrix of source node in the time that time slot is even number, A ₁and B ₁be M × M and tie up matrix, x ₂=[x ₂₁, x ₂₂... x _2i... x _2M] ^t, x _2ii the useful signal component that source node S sends at the second time slot, i=1,2 ... M;

2.3), at the 3rd time slot, the mixed signal vector representation that source node S is sent is s ₃:

s ₃＝A ₂B ₂(x ₃+x ₂)，

Wherein, A ₂b ₂the cascade pre-coding matrix of source node in the time that time slot is odd number, A ₂and B ₂be M × M and tie up matrix, x ₃=[x ₃₁, x ₃₂... x _3i... x _3M] ^t, x _3ii the useful signal component that source node S sends at the 3rd time slot, i=1,2 ... M;

3) according to step 2) build source node S send mixed signal vector, obtain the first relaying R ₁the signal phasor y receiving at first time slot _r1,1for:

y _r1，1＝H ₁x ₁+n _r1，1，

Suppose the first relaying R ₁can be by x ₁be correctly decoded and at the second time slot by x ₁transmit the signal phasor s that second time slot the first relaying sends _r1,2be expressed as:

s _r1，2＝T ₁W ₁x ₁，

The second relaying R ₂the signal phasor y receiving at second time slot _r2,2for:

y _r2，2＝H ₂s ₂+F ₁s _r1，2+n _r2，2

＝H ₂A ₁B ₁x ₂+(H ₂A ₁B ₁+F ₁T ₁W ₁)x ₁+n _r2，2，

Wherein, n _r1,1at first time slot the first relaying R ₁the additive white Gaussian noise at place, H ₁that source node S is to the first relaying R ₁m × M dimension flat fading channel matrix, n _r2,2at second time slot the second relaying R ₂the additive white Gaussian noise at place, H ₂that source node S is to the second relaying R ₂m × M dimension flat fading channel matrix, F ₁from the first relaying R ₁to the second relaying R ₂m × M dimension flat fading channel matrix, T ₁w ₁first relaying R ₁the cascade pre-coding matrix at place, T ₁and W ₁be M × M and tie up matrix, H ₂a ₁b ₁x ₂at the second relaying R ₂the useful signal vector at place, (H ₂a ₁b ₁+ F ₁t ₁w ₁) x ₁at the second relaying R ₂the relay well at place disturbs;

4) the pre-coding matrix A of design source node in the time that time slot is even number ₁, B ₁, the first relaying R ₁the pre-coding matrix T at place ₁, W ₁, make the second relaying R ₂the relay well at place disturbs (H ₂a ₁b ₁+ F ₁t ₁w ₁) x ₁eliminate completely;

5) according to step 4) design, suppose the second relaying R ₂can be by x ₂be correctly decoded and at the 3rd time slot by x ₂transmit the signal phasor s that the 3rd time slot the second relaying sends _r2,3be expressed as:

s _r2，3＝T ₂W ₂x ₂，

The first relaying R ₁the signal phasor y receiving at the 3rd time slot _r1,3for:

y _r1，3＝H ₁s ₃+F ₂s _r2，3+n _r1，3

＝H ₁A ₂B ₂x ₃+(H ₁A ₂B ₂+F ₂T ₂W ₂)x ₂+n _r1，3

Wherein, n _r1,3at the 3rd time slot the first relaying R ₁the additive white Gaussian noise at place, F ₂from the second relaying R ₂to the first relaying R ₁m × M dimension flat fading channel matrix, T ₂w ₂the second relaying R ₂the pre-coding matrix of the cascade at place, T ₂and W ₂be M × M and tie up matrix, H ₁a ₂b ₂x ₃at the first relaying R ₁the useful signal at place, (H ₁a ₂b ₂+ F ₂t ₂w ₂) x ₂at the first relaying R ₁the relay well at place disturbs;

6) according to step 4) in the pre-coding matrix A of source node in the time that time slot is even number ₁, B ₁, the first relaying R ₁the pre-coding matrix T at place ₁, W ₁identical method for designing, the pre-coding matrix A of design source node in the time that time slot is odd number ₂, B ₂, the second relaying R ₂the pre-coding matrix T at place ₂, W ₂, make the first relaying R ₁the relay well at place disturbs (H ₁a ₂b ₂+ F ₂t ₂w ₂) x ₂eliminate completely;

7) relay well disturbs by step 4) and step 6) after elimination, destination node D is at the signal phasor y of second time slot _{d, 2}with the vector y at the 3rd time slot _{d, 3}for:

y _d，2＝G ₁s _r1，2+n _d，2＝G ₁T ₁W ₁x ₁+n _d，2，

y _d，3＝G ₂s _r2，3+n _d，3＝G ₂T ₂W ₂x ₂+n _d，3，

Wherein, n _{d, 2}the additive white Gaussian noise at second time slot destination node D place, G ₁the first relaying R ₁to M × M dimension flat fading channel matrix of destination node D, G ₁t ₁w ₁x ₁for receive from the first relaying R ₁the signal of non-relay interference, n _{d, 3}the additive white Gaussian noise at the 3rd time slot destination node D place, G ₂the second relaying R ₂to M × M dimension flat fading channel matrix of destination node D, G ₂t ₂w ₂x ₂for destination node receive from the second relaying R ₂the signal of non-relay interference.

Compared with prior art, tool has the following advantages in the present invention:

1) improved the degree of freedom.Existing based on disturbing in the scheme of alignment IA, by interference signal being snapped to the degree of freedom that can only reach the 3/4M of system in an interference space, the present invention is by making system reach maximum degree of freedom M at source node and relaying place by design pre-coding matrix;

2) system arranges simply.Existing based on disturbing in the scheme of alignment IA, must configure three repeaters and just can complete interference and align, the present invention only need configure two relayings and can complete interference and eliminate completely;

3) condition restriction is little.The existing scheme based on disturbing alignment IA can only be operated under the condition that node antenna number is even number, and the present invention, to the not requirement of the parity of system node antenna number, makes the condition of work of system not be subject to the restriction of node antenna number.

Brief description of the drawings

The MIMO that Fig. 1 the present invention uses replaces relay system schematic diagram;

Fig. 2 is realization flow figure of the present invention;

Fig. 3 is under the condition to the total node antenna number of fixed system, the degree of freedom comparison diagram obtaining by the present invention and existing interference alignment IA method system respectively;

Fig. 4 is under the condition to the each node antenna number of fixed system, the degree of freedom comparison diagram obtaining by the present invention and existing interference alignment IA method system respectively.

Embodiment

Referring to accompanying drawing, technical scheme of the present invention and effect are described in further detail.

According to Fig. 2, performing step of the present invention is as follows:

The setting of step 1. system:

With reference to Fig. 1, the mimo system of setting of the present invention comprises a source node S, a destination node D and two relaying R ₁, R ₂, they all configure M root antenna, M >=2, and the transmission means of relaying is half-duplex;

While making time slot be odd number, source node S and the second relaying R ₂transmitted signal, simultaneously destination node D and the first relaying R ₁receive signal;

While making time slot be even number, source node S and the first relaying R ₁transmitted signal, simultaneously destination node D and the second relaying R ₂receive signal;

Step 2. builds the mixed signal vector that source node S sends at different time-gap:

2.1), at first time slot, the signal indication that source node S is sent is s ₁:

s ₁＝x ₁＝[x ₁₁，?x ₁₂，?…?x _1i，?…?x _1M] ^T，

Wherein, x _1ii the signal component that source node S sends at first time slot, i=1,2 ... M, T representing matrix transposition;

2.2), at second time slot, the mixed signal vector representation that source node S is sent is s ₂:

s ₂＝A ₁B ₁(x ₂+x ₁)，

Wherein, A ₁b ₁the cascade pre-coding matrix of source node in the time that time slot is even number, A ₁and B ₁be M × M and tie up matrix; x ₂=[x ₂₁, x ₂₂... x _2i... x _2M] ^t, x _2ii the useful signal component that source node S sends at the second time slot, i=1,2 ... M;

2.3), at the 3rd time slot, the mixed signal vector representation that source node S is sent is s ₃:

s ₃＝A ₂B ₂(x ₃+x ₂)，

Wherein, A ₂b ₂the cascade pre-coding matrix of source node in the time that time slot is odd number, A ₂and B ₂be M × M and tie up matrix, x ₃=[x ₃₁, x ₃₂... x _3i... x _3M] ^t, x _3ii the useful signal component that source node S sends at the 3rd time slot, i=1,2 ... M;

The mixed signal vector that the source node S that step 3. builds according to step 2 sends, obtains the first relaying R ₁the signal phasor y receiving at first time slot _r1,1for:

y _r1，1＝H ₁x ₁+n _r1，1，

Suppose the first relaying R ₁can be by x ₁be correctly decoded and at the second time slot by x ₁transmit the signal phasor s that second time slot the first relaying sends _r1,2be expressed as:

s _r1，2＝T ₁W ₁x ₁，

The second relaying R ₂the signal phasor y receiving at second time slot _r2,2for:

y _r2，2＝H ₂s ₂+F ₁s _r1，2+n _r2，2

＝H ₂A ₁B ₁x ₂+(H ₂A ₁B ₁+F ₁T ₁W ₁)x ₁+n _r2，2，

Wherein, n _r1,1at first time slot the first relaying R ₁the additive white Gaussian noise at place, H ₁that source node S is to the first relaying R ₁m × M dimension flat fading channel matrix, n _r2,2at second time slot the second relaying R ₂the additive white Gaussian noise at place, H ₂that source node S is to the second relaying R ₂m × M dimension flat fading channel matrix, F ₁from the first relaying R ₁to the second relaying R ₂m × M dimension flat fading channel matrix, T ₁w ₁first relaying R ₁the cascade pre-coding matrix at place, T ₁and W ₁be M × M and tie up matrix, H ₂a ₁b ₁x ₂at the second relaying R ₂the useful signal vector at place, (H ₂a ₁b ₁+ F ₁t ₁w ₁) x ₁at the second relaying R ₂the relay well at place disturbs;

Step 4. designs the pre-coding matrix A of source node in the time that time slot is even number ₁, B ₁, the first relaying R ₁the pre-coding matrix T at place ₁, W ₁, make the second relaying R ₂the relay well at place disturbs (H ₂a ₁b ₁+ F ₁t ₁w ₁) x ₁eliminate completely:

4.1) design matrix A ₁, T ₁, make H ₂a ₁and F ₁t ₁be aligned in H ₂and F ₁common factor SPACE V ₁upper,

V ₁＝H ₂A ₁＝F ₁T ₁，

Above formula is deformed into:

V ₁-H ₂A ₁＝0，

V ₁-F ₁T ₁＝0，

Wherein A ₁the pre-coding matrix that source node is even number at time slot, T ₁the first relaying R ₁the pre-coding matrix at place, H ₂that source node S is to the second relaying R ₂m × M dimension flat fading channel matrix, F ₁from the first relaying R ₁to the second relaying R ₂m × M dimension flat fading channel matrix, V ₁that M × M ties up matrix;

4.2) equal the character of this square formation of unit matrix premultiplication according to square formation in matrix theory, can obtain:

V ₁=I _mv ₁, wherein I _mthat M × M ties up unit matrix;

4.3) by 4.2) V that obtains ₁=I _mv ₁substitution 4.1) in V ₁-H ₂a ₁=0 and V ₁-F ₁t ₁=0, obtain following two equations:

I _MV ₁-H ₂A ₁＝0，

I _MV ₁-F ₁T ₁＝0，

4.4) by 4.3) two equation matrix notations obtaining are as follows:

$\left[\begin{array}{ccc}{I}_M& -H_2& 0\\ I_M& 0& -F_1\end{array}\right]\left[\begin{array}{c}{V}_1\\ A_1\\ T_1\end{array}\right]=0,$

4.5) order $\left[\begin{array}{ccc}{I}_M& -H_2& 0\\ I_M& 0& -F_1\end{array}\right]=U_1,\left[\begin{array}{c}{V}_1\\ A_1\\ T_1\end{array}\right]=X_1,$ By 4.4) in matrix equation be expressed as:

U ₁X ₁＝0

Wherein, U ₁the matrix and the order that are 2M × 3M dimension are 2M, X ₁the matrix of 3M × M dimension, matrix X ₁for matrix U ₁kernel;

4.6), in order to eliminate interference, make the relay well distracter H at the second relaying place ₂a ₁b ₁+ F ₁t ₁w ₁equal zero, that is: H ₂a ₁b ₁+ F ₁t ₁w ₁=0,

4.7) by 4.1) the equation V that obtains ₁=H ₂a ₁=F ₁t ₁be updated to 4.6) equation in, can obtain:

V ₁B ₁+V ₁W ₁＝0

Wherein, W ₁=α ₁i _m, α ₁represent the first relaying R ₁power confinement factor,

B ₁=-β ₁i _m, β ₁represent the power confinement factor of source node S, I _mthat M × M ties up unit matrix;

4.8) for making 4.7) in equation V ₁b ₁+ V ₁w ₁=0 sets up, and must meet B ₁=-W ₁:

If the β 4.7) ₁=α ₁, meet B ₁=-W ₁if, 4.7) in β ₁≠ α ₁, make W ₁=γ ₁i _m, B ₁=-γ ₁i _m, wherein, γ ₁=min{ α ₁, β ₁.

Step 5., according to the design of step 4, is supposed the second relaying R ₂can be by x ₂be correctly decoded and at the 3rd time slot by x ₂transmit the signal phasor s that the 3rd time slot the second relaying sends _r2,3be expressed as:

s _r2，3＝T ₂W ₂x ₂，

The first relaying R ₁the signal phasor y receiving at the 3rd time slot _r1,3for:

y _r1，3＝H ₁s ₃+F ₂s _r2，3+n _r1，3

＝H ₁A ₂B ₂x ₃+(H ₁A ₂B ₂+F ₂T ₂W ₂)x ₂+n _r1，3

Wherein, n _r1,3at the 3rd time slot the first relaying R ₁the additive white Gaussian noise at place, F ₂from the second relaying R ₂to the first relaying R ₁m × M dimension flat fading channel matrix, T ₂w ₂the second relaying R ₂the pre-coding matrix of the cascade at place, T ₂and W ₂be M × M and tie up matrix, H ₁a ₂b ₂x ₃at the first relaying R ₁the useful signal at place, (H ₁a ₂b ₂+ F ₂t ₂w ₂) x ₂at the first relaying R ₁the relay well at place disturbs;

Step 6. according to step 4 in the pre-coding matrix A of source node in the time that time slot is even number ₁, B ₁, the first relaying R ₁the pre-coding matrix T at place ₁, W ₁identical method for designing, the pre-coding matrix A of design source node in the time that time slot is odd number ₂, B ₂, the second relaying R ₂the pre-coding matrix T at place ₂, W ₂, make the first relaying R ₁the relay well at place disturbs (H ₁a ₂b ₂+ F ₂t ₂w ₂) x ₂eliminate completely;

After step 7. relay well disturbs and eliminates completely by step 4 and step 6, destination node D is at the signal phasor y of second time slot _{d, 2}with the vector y at the 3rd time slot _{d, 3}for:

y _d，2＝G ₁s _r1，2+n _d，2＝G ₁T ₁W ₁x ₁+n _d，2，

y _d，3＝G ₂s _r2，3+n _d，3＝G ₂T ₂W ₂x ₂+n _d，3，

Wherein, n _{d, 2}the additive white Gaussian noise at second time slot destination node D place, G ₁the first relaying R ₁to M × M dimension flat fading channel matrix of destination node D, G ₁t ₁w ₁x ₁for receive from the first relaying R ₁the signal of non-relay interference, n _{d, 3}the additive white Gaussian noise at the 3rd time slot destination node D place, G ₂the second relaying R ₂to M × M dimension flat fading channel matrix of destination node D, G ₂t ₂w ₂x ₂for destination node receive from the second relaying R ₂the signal of non-relay interference.

Due to the first relaying R ₁the pre-coding matrix T at place ₁only and source node S to the second relaying R ₂flat fading channel matrix H ₂, the first relaying R ₁to the second relaying R ₂flat fading channel matrix F ₁relevant, and H ₂, F ₁with the first relaying R ₁to the flat fading channel matrix G of destination node D ₁linear independence, so G ₁t ₁w ₁order be also M, the second relaying R ₂the pre-coding matrix T at place ₂only and source node S to the second relaying R ₁flat fading channel matrix H ₁, the second relaying R ₂to the first relaying R ₁flat fading channel matrix F ₂relevant, and H ₁, F ₂with the second relaying R ₂to the flat fading channel matrix G of destination node D ₂linear independence, so G ₂t ₂w ₂order be also M, system reaches maximum degree of freedom M.

By the source node of above-mentioned design and the pre-encoder matrix of relaying, the interference of relay well is eliminated completely.Eliminate in order to describe more clearly the present invention the process that relay well disturbs, the each node of system be illustrated at data symbol and the pre-coding matrix thereof of different time-gap transmission with table 1:

Data symbol and pre-coding matrix thereof that the each node of table 1 system sends at different time-gap

Time slot

1

2

3

4

5

…

The data symbol that source node sends

x 1

x 2+x 1

x 3+x 2

x 4+x 3

x 5+x 4

…

The first relaying R 1The data symbol sending

?

x 1

?

x 3

?

…

The second relaying R 2The data symbol sending

?

?

x 2

?

x 4

…

The pre-coding matrix of source node

I

A 1B 1

A 2B 2

A 1B 1

A 2B 2

…

The first relaying R 1Pre-coding matrix

?

T 1W 1

?

T 1W 1

?

…

The second relaying R 2Pre-coding matrix

?

?

T 2W 2

?

T 2W 2

…

Effect of the present invention can further illustrate by following simulation result:

1. simulated conditions: all node antenna number of initialization system are respectively 10,20,30,40,50 and the antenna number of each node be respectively 2,4,6,8,10.

2. emulation content:

Emulation 1. use the present invention and existing interference alignment IA method are respectively 10,20 in all antenna number, and 30,40,50 o'clock, the degree of freedom that system is obtained was carried out emulation, and result as shown in Figure 3.

As can be seen from Figure 3: under the identical condition of the antenna number of all nodes, the degree of freedom that the present invention reaches is far above the degree of freedom based on disturbing alignment IA method to reach.

Emulation 2. use the present invention and existing interference alignment IA method are respectively 2,4 in each node antenna number, and 6,8,10 o'clock, the degree of freedom that system is obtained was carried out emulation, and result as shown in Figure 4.

As can be seen from Figure 4: under the identical condition of each node antenna number, the degree of freedom that the present invention reaches is far above the degree of freedom based on disturbing alignment IA method to reach.

Claims (2)

1. MIMO replaces the interference elimination method forwarding based on decoding in relay system, comprises the steps:

1) system setting:

If mimo system comprises a source node S, a destination node D and two relaying R ₁, R ₂, they all configure M root antenna, M >=2, and the transmission means of relaying is half-duplex;

While making time slot be odd number, source node S and the second relaying R ₂transmitted signal, simultaneously destination node D and the first relaying R ₁receive signal;

While making time slot be even number, source node S and the first relaying R ₁transmitted signal, simultaneously destination node D and the second relaying R ₂receive signal;

2) build the mixed signal vector that source node S sends at different time-gap:

2.1), at first time slot, the signal indication that source node S is sent is s ₁:

s ₁＝x ₁＝[x ₁₁，?x ₁₂，?…x _1i，?…x _1M] ^T，

Wherein, x _1ii the signal component that source node S sends at first time slot, i=1,2 ... M, T representing matrix transposition;

2.2), at second time slot, the mixed signal vector representation that source node S is sent is s ₂:

s ₂＝A ₁B ₁(x ₂+x ₁)，

Wherein, A ₁b ₁the cascade pre-coding matrix of source node in the time that time slot is even number, A ₁and B ₁be M × M and tie up matrix; x ₂=[x ₂₁, x ₂₂... x _2i... x _2M] ^t, x _2ii the useful signal component that source node S sends at the second time slot, i=1,2 ... M;

2.3), at the 3rd time slot, the mixed signal vector representation that source node S is sent is s ₃:

s ₃＝A ₂B ₂(x ₃+x ₂)，

Wherein, A ₂b ₂the cascade pre-coding matrix of source node in the time that time slot is odd number, A ₂and B ₂be M × M and tie up matrix, x ₃=[x ₃₁, x ₃₂... x _3i... x _3M] ^t, x _3ii the useful signal component that source node S sends at the 3rd time slot, i=1,2 ... M;

3) according to step 2) build source node S send mixed signal vector, obtain the first relaying R ₁the signal phasor y receiving at first time slot _r1,1for:

y _r1，1＝H ₁x ₁+n _r1，1，

Suppose the first relaying R ₁can be by x ₁be correctly decoded and at the second time slot by x ₁transmit the signal phasor s that second time slot the first relaying sends _r1,2be expressed as:

s _r1，2＝T ₁W ₁x ₁，

The second relaying R ₂the signal phasor y receiving at second time slot _r2,2for:

y _r2，2＝H ₂s ₂+F ₁s _r1，2+n _r2，2

＝H ₂A ₁B ₁x ₂+(H ₂A ₁B ₁+F ₁T ₁W ₁)x ₁+n _r2，2

Wherein, n _r1,1at first time slot the first relaying R ₁the additive white Gaussian noise at place, H ₁that source node S is to the first relaying R ₁m × M dimension flat fading channel matrix, n _r2,2at second time slot the second relaying R ₂the additive white Gaussian noise at place, H ₂that source node S is to the second relaying R ₂m × M dimension flat fading channel matrix, F ₁from the first relaying R ₁to the second relaying R ₂m × M dimension flat fading channel matrix, T ₁w ₁first relaying R ₁the cascade pre-coding matrix at place, T ₁and W ₁be M × M and tie up matrix, H ₂a ₁b ₁x ₂at the second relaying R ₂the useful signal vector at place, (H ₂a ₁b ₁+ F ₁t ₁w ₁) x ₁at the second relaying R ₂the relay well at place disturbs;

4) the pre-coding matrix A of design source node in the time that time slot is even number ₁, B ₁, the first relaying R ₁the pre-coding matrix T at place ₁, W ₁, make the second relaying R ₂the relay well at place disturbs (H ₂a ₁b ₁+ F ₁t ₁w ₁) x ₁eliminate;

5) according to step 4) design, suppose the second relaying R ₂can be by x ₂be correctly decoded and at the 3rd time slot by x ₂transmit the signal phasor s that the 3rd time slot the second relaying sends _r2,3be expressed as:

s _r2，3＝T ₂W ₂x ₂，

The first relaying R ₁the signal phasor y receiving at the 3rd time slot _r1,3for:

y _r1，3＝H ₁s ₃+F ₂s _r2，3+n _r1，3

＝H ₁A ₂B ₂x ₃+(H ₁A ₂B ₂+F ₂T ₂W ₂)x ₂+n _r1，3

Wherein, n _r1,3at the 3rd time slot the first relaying R ₁the additive white Gaussian noise at place, F ₂from the second relaying R ₂to the first relaying R ₁m × M dimension flat fading channel matrix, T ₂w ₂the second relaying R ₂the pre-coding matrix of the cascade at place, T ₂and W ₂be M × M and tie up matrix, H ₁a ₂b ₂x ₃at the first relaying R ₁the useful signal at place, (H ₁a ₂b ₂+ F ₂t ₂w ₂) x ₂at the first relaying R ₁the relay well at place disturbs;

6) according to step 4) in the pre-coding matrix A of source node in the time that time slot is even number ₁, B ₁, the first relaying R ₁the pre-coding matrix T at place ₁, W ₁identical method for designing, the pre-coding matrix A of design source node in the time that time slot is odd number ₂, B ₂, the second relaying R ₂the pre-coding matrix T at place ₂, W ₂, make the first relaying R ₁the relay well at place disturbs (H ₁a ₂b ₂+ F ₂t ₂w ₂) x ₂eliminate completely;

7) relay well disturbs by step 4), step 6) eliminate after, destination node D is at the signal phasor y of second time slot _{d, 2}with the vector y at the 3rd time slot _{d, 3}for:

y _d，2＝G ₁s _r1，2+n _d，2＝G ₁T ₁W ₁x ₁+n _d，2，

y _d，3＝G ₂s _r2，3+n _d，3＝G ₂T ₂W ₂x ₂+n _d，3，

Wherein, n _{d, 2}the additive white Gaussian noise at second time slot destination node D place, G ₁the first relaying R ₁to M × M dimension flat fading channel matrix of destination node D, G ₁t ₁w ₁x ₁for receive from the first relaying R ₁the signal of non-relay interference, n _{d, 3}the additive white Gaussian noise at the 3rd time slot destination node D place, G ₂the second relaying R ₂to M × M dimension flat fading channel matrix of destination node D, G ₂t ₂w ₂x ₂for destination node receive from the second relaying R ₂the signal of non-relay interference.

2. method according to claim 1, wherein step 4) described design pre-encoder matrix, comprise the steps to carry out:

4a) design matrix A ₁, T ₁, make H ₂a ₁and F ₁t ₁be aligned in H ₂and F ₁common factor SPACE V ₁upper,

V ₁＝H ₂A ₁＝F ₁T ₁，

Above formula is deformed into:

V ₁-H ₂A ₁＝0，

V ₁-F ₁T ₁＝0，

Wherein A ₁the pre-coding matrix that source node is even number at time slot, T ₁the first relaying R ₁the pre-coding matrix at place, H ₂that source node S is to the second relaying R ₂m × M dimension flat fading channel matrix, F ₁from the first relaying R ₁to the second relaying R ₂m × M dimension flat fading channel matrix, V ₁that M × M ties up matrix;

4b) equal the character of this square formation of unit matrix premultiplication according to square formation in matrix theory, can obtain:

V ₁=I _mv ₁, wherein I _mthat M × M ties up unit matrix;

4c) by 4b) V that obtains ₁=I _mv ₁substitution 4a) in V ₁-H ₂a ₁=0 and V ₁-F ₁t ₁=0, obtain following two equations:

I _MV ₁-H ₂A ₁＝0，

I _MV ₁-F ₁T ₁＝0，

4d) by 4c) two equation matrix notations obtaining are as follows:

$\left[\begin{array}{ccc}{I}_M& -H_2& 0\\ I_M& 0& -F_1\end{array}\right]\left[\begin{array}{c}{V}_1\\ A_1\\ T_1\end{array}\right]=0,$

4e) order $\left[\begin{array}{ccc}{I}_M& -H_2& 0\\ I_M& 0& -F_1\end{array}\right]=U_1,\left[\begin{array}{c}{V}_1\\ A_1\\ T_1\end{array}\right]=X_1,$ By 4d) in matrix equation be expressed as:

U ₁X ₁＝0

Wherein, U ₁the matrix and the order that are 2M × 3M dimension are 2M, X ₁the matrix of 3M × M dimension, matrix X ₁for matrix U ₁kernel;

4f) in order to eliminate interference, make the relay well distracter H at the second relaying place ₂a ₁b ₁+ F ₁t ₁w ₁equal zero, that is: H ₂a ₁b ₁+ F ₁t ₁w ₁=0,

4g) by 4a) the equation V that obtains ₁=H ₂a ₁=F ₁t ₁be updated to 4f) equation in, can obtain:

V ₁B ₁+V ₁W ₁＝0

Wherein, W ₁=α ₁i _m, α ₁represent the first relaying R ₁power confinement factor,

B ₁=-β ₁i _m, β ₁represent the power confinement factor of source node S, I _mthat M × M ties up unit matrix;

4h) for making 4g) in equation V ₁b ₁+ V ₁w ₁=0 sets up, and must meet B ₁=-W ₁:

If the β 4g) ₁=α ₁, meet B ₁=-W ₁if, 4g) in β ₁≠ α ₁, make W ₁=γ ₁i _m, B ₁=-γ ₁i _m, wherein, γ ₁=min{ α ₁, β ₁.