CN1744459A - Communication system and method using a relay node - Google Patents

Abstract

The invention provides a communication system and method using a relay node. A communication node relays a transmission signal transmitted from a desired source node to a target destination node among multiple source nodes and multiple destination nodes. The communication node includes a first unitary matrix estimation unit that estimates a first unitary matrix by performing singular value decomposition involving one or more channel matrices between the relay node and the source nodes other than the desired source node; a second unitary matrix estimation unit that estimates a second unitary matrix by performing singular value decomposition involving one or more channel matrices between the relay node and the destination nodes other than the target destination node; and a transmission unit configured to transmit a relaying signal generated by multiplying a received signal by the first and second unitary matrices toward the target destination node.

Description

Use the communication system and the method for via node

Technical field

Present invention relates in general to radio communication, more specifically, relate to the communication node and the communication means that use multi-hop scheme and multiple-input and multiple-output (MIMO) scheme.

Background technology

In recent years, the system's (this system is called as the MIMO multihop system) based on the combination of multi-hop scheme and MIMO (or many antennas) scheme continues to obtain to pay close attention to.In the multi-hop scheme, signal sends to destination node (or destination node) by the one or more via nodes between source and destination from source node.This system has the advantage of coming the quick foundation of extended coverage range (being unrestricted signal transmission region in theory) and wireless network by repeating signal.By mimo system, use multiple transmit antennas and multiple receive antenna sends and received signal, to be used for improving message capacity by effectively making of space.

The signal of carrying out in the MIMO multihop system with following step transmits.At first, receive the signal S that sends from source node at the via node place.Received signal Y at the via node place is expressed as:

Y＝HS+n (1)

Wherein H represents the channel matrix between source and the via node, and S represents transmission signal vector, and n represents noise.Then, come detection of transmitted signals S by ZF (ZF) method.This method is by calculating pseudo inverse matrix W ₁=(H ^HH)- ¹H ^H, and received signal be multiply by pseudo inverse matrix W ₁And normalization coefficient comes detection of transmitted signals S.This processing list is shown:

W ₁Y＝S+W ₁n (2)

Pseudo inverse matrix W ₁In subscript H represent to grip altogether transposition.

The norm (Norm) of matrix A can be defined as arbitrarily:

‖A‖＝(Tr(E[AA ^H])) ^1/2 (3)

Wherein symbol ‖ ‖ represents norm, the summation (that is, mark) of the diagonal element of the matrix in symbol Tr () the expression round parentheses, and symbol E[] represent the numerical value in the square brackets is averaged.Particularly, with vectorial V=(v ₁, v ₂..., v _M) ^TNorm ‖ V ‖ be expressed as:

‖V‖＝[|v ₁| ²+|v ₂| ²+…+|v _M| ²] ^1/2 (3)′

Wherein subscript T represents transposition.Above-mentioned pseudo inverse matrix is corresponding with the Moore-Penrose inverse matrix.Usually, Moore-Penrose inverse matrix B is defined as m * n matrix, it is for n * m matrix A, and BA=I sets up.In the example shown, for matrix H, W ₁H=I sets up.

Then, calculate pseudo inverse matrix W ₂=(G ^HG) ^-1G ^H, wherein G is illustrated in the channel matrix between via node and the destination node.The both sides of equation (2) be multiply by this pseudo inverse matrix W simultaneously ₂With normalization coefficient E.This relation table is shown:

E(W ₂W ₁)Y＝EW ₂(S+W ₁n) (4)

Wherein, $E=1/\left(\right|\left|{\mathrm{<figure-callout\; id="1"\; label="W"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">W</figure-callout>}}_1\right|\left|\right|\left|{\mathrm{<figure-callout\; id="2"\; label="W"\; filenames="A20051009384000511.png"\; state="\{\{state\}\}">W</figure-callout>}}_2\right|\left|\right)*{(P_s/(P_s+{\mathrm{\&sigma;}}_n^2))}^{1/2}$ Set up, Ps represents transmitted power, and σ ²It is noise variance.

The signal that calculates is thus sent to the destination node from via node.Signal Y in the reception of destination node _RBe expressed as:

Y _R＝GEW ₂W ₁Y+n _R (5)

N wherein _RThe expression noise component(s).Can be according to W ₁And W ₂Definition equation (5) is rewritten as:

Y _R＝E(S+W ₁n)+n _R (6)

Like this, can obtain to send signal S immediately at the destination node.This MIMO multihop system has for example been described in the document below, Rohit U.Nabar, et al., " CapacityScaling Laws in MIMO Wireless networks ", Allerton Conference onCommunication, Control, and Computing, Monticello, IL., pp.378-389, Oct.2003.

According to equation (6), be to be understood that the signal Y that is received _RComprise and the relevant factor 1/ (the ‖ W of transmission signal S ₁‖ ‖ W ₂‖).This factor ‖ W ₁‖ and ‖ W ₂‖ is absolutely necessary for the transmission power control of carrying out at via node.Yet, because W ₁And W ₂Be respectively the inverse matrix (it is subjected to the influence of noise amplitude) of channel matrix H and G, so signal quality will reduce inevitably.In addition, equation (6) comprises noise component(s) " n ", and this noise component(s) " n " is to introduce in the communication process from the source to the via node, thereby has had a strong impact on received signal.Therefore, along with the increase of jumping figure, because the Signal Degrade that noise causes will become remarkable.

In addition, must consider following wireless communication system, in this wireless communication system, signal is relayed to relevant destination node from the multiple source node simultaneously by via node.In this system, the signal that receives at the destination node not only comprises the influence of desirable source node, also comprises the influence of other source node.Following worry is arranged in this system: noise is exaggerated at via node, and seriously descends at the received signal quality of destination node.

Summary of the invention

The present invention is intended to overcome the problems referred to above, and an object of the present invention is to provide a kind of communication system, communication node and communication means, it is compared with conventional art in the transmission of the signal from the source node to the destination, can more effectively prevent to reduce at the received signal quality of destination node.

In one aspect of the invention, provide a kind of communication node, be used between multiple source node and a plurality of destinations node, the transmission signal relay that will send from desirable source node is to target destination node.This communication node comprises:

(a) the first unitary matrice estimation unit, it is configured to estimate first unitary matrice by one or more channel matrix between via node and the multiple source node except desirable source node is carried out singular value decomposition;

(b) the second unitary matrice estimation unit, it is configured to estimate second unitary matrice by one or more channel matrix between via node and a plurality of destinations node except this target destination node is carried out singular value decomposition; And

(c) transmitting element, it is configured to and will sends to target destination node by the repeating signal that the signal times that is received is generated with first and second unitary matrice.

In the communication system of using this via node, the destination node is according to the repeating signal detection of transmitted signals that is received.

In another aspect of this invention, provide a kind of communication node, the transmission signal relay that is used for the desirable source node from the multiple source node is sent is to the destination node.This communication node comprises:

(a) Matrix Estimation unit, it is configured to estimate the Moore-Penrose inverse matrix that derives according to a plurality of channel matrixes between via node and a plurality of node;

(b) the first unitary matrice estimation unit, it is configured to estimate first unitary matrice by one or more channel matrix between via node and the multiple source node except desirable source node is carried out singular value decomposition;

(c) the second unitary matrice estimation unit, it is configured to estimate second unitary matrice by one or more channel matrix between via node and a plurality of destinations node except the node of described destination is carried out singular value decomposition;

(d) repeating signal generation unit, it is configured to generate repeating signal by the signal times that will be received with two in the weighting matrix of definition Moore-Penrose inverse matrix, first unitary matrice, second unitary matrice; And

(e) transmitting element, it is configured to repeating signal is sent to the destination node.

By the communication node of arbitrary type, use multi-hop MIMO scheme from source node when the destination node sends signal, can prevent that the signal quality that receives at the destination node from descending.

Description of drawings

When read in conjunction with the accompanying drawings, according to following detailed, other purpose of the present invention, feature and advantage become more obvious.In the accompanying drawing:

Fig. 1 is the schematic diagram that the communication system that adopts MIMO scheme and multi-hop scheme is shown;

Fig. 2 is the schematic block diagram of via node;

Fig. 3 is the functional block diagram according to the repeating signal maker of first embodiment of the invention;

Fig. 4 is the flow chart that illustrates according to the operation of the communication system of first embodiment of the invention;

Fig. 5 is the functional block diagram according to the repeating signal maker of second embodiment of the invention;

Fig. 6 is the flow chart that the operation of the communication system of using repeating signal maker shown in Figure 5 is shown;

Fig. 7 A and Fig. 7 B are the curve chart of expression according to the simulation result of the present invention of third embodiment of the invention;

Fig. 8 illustrates the schematic diagram of wherein a plurality of nodes by the communication system of via node transmission and received signal;

Fig. 9 is the functional block diagram of traditional via node;

Figure 10 is the functional block diagram according to the via node of fourth embodiment of the invention;

Figure 11 is illustrated in the example of the arithmetical operation of via node execution;

Figure 12 is illustrated in another example of the arithmetical operation of via node execution;

Figure 13 is illustrated in another example of the arithmetical operation of via node execution;

Figure 14 is illustrated in another example of the arithmetical operation of via node execution;

Figure 15 represents compared with prior art, the curve chart of simulation result of the present invention.

Embodiment

Describe the present invention in detail below in conjunction with accompanying drawing.In specification and claim, " unitary matrice " needs not to be normal matrix (normal matrix), thereby line number and columns can differ from one another." unitary matrice " is the orthogonal matrix of each row (or row).Therefore, also comprise the diagonalizable N of the non-square matrix B * M non-square matrix that the diagonalizable normal matrix of square formation A, " unitary matrice " is comprised be used to make M * N.

In a preferred embodiment, by being that the product that comprises first triangular matrix is determined first unitary matrice with first channel matrix decomposition between source node and the via node, and by being that the product that comprises second triangular matrix is determined second unitary matrice with the second channel matrix decomposition between via node and the destination node.If i+j does not satisfy setting, then the matrix element of the capable j row of i is 0.

Employed communication node comprises among the embodiment: first device is used for the channel matrix H between source node and the via node is decomposed into the product that comprises the first triangular matrix E; Second device is used for the channel matrix G between via node and the destination node is decomposed into the product that comprises the second triangular matrix P; The transformation matrix generating apparatus is used for generating transformation matrix A based on first and second triangular matrixes; Multiplier, the signal and first unitary matrice, transformation matrix and second unitary matrice that are used for being received multiply each other, to generate repeating signal; And dispensing device, be used for repeating signal is sent to the destination node.If i+j does not satisfy setting, then the capable j column element of i of transformation matrix A is 0.

Because use unitary matrice and transformation matrix to generate repeating signal, so can when reducing the loss of signal and degraded signal quality, realize multi-hop communication.

In an example, estimate transformation matrix based on the associate matrix of first unitary matrice, commutative matrix and second unitary matrice.By this set, the destination node can in phase merge the repeating signal from a plurality of via nodes.Because the signal merge coefficient does not comprise imaginary component (phase component), thus during merging, signal do not need to delete some component, therefore, can be at destination node homophase merging coherently repeating signal.

In a preferred exemplary, use by the feedback channel of via node to source node, feeds back to source node with the information relevant with power level with the speed that sends signal from the destination node from the destination node.Obtain this information at the destination node according to channel estimation value.

In a preferred exemplary, provide a kind of and will be relayed to the method for destination node by via node from the transmission signal that source node sends.In the method, at the via node place, with first channel matrix decomposition between source node and the via node is the product that comprises first triangular matrix, and is the product that comprises second triangular matrix with the second channel matrix decomposition between via node and the destination node.Then, generate transformation matrix based on first and second triangular matrixes, if wherein i+j does not satisfy setting, then the capable j column element of i of transformation matrix is 0.Then, the signal times that will receive at via node is with first unitary matrice, transformation matrix and second unitary matrice.Then, multiplied signals is sent to the destination node from via node.

Preferably, based on first and second triangular matrixes and transformation matrix, generate the 3rd triangular matrix at the destination node.The destination node uses the 3rd triangular matrix detection of transmitted signals from the signal that is received then.

In another example, at the via node place, with first channel matrix decomposition between source node and the via node is the product that comprises first triangular matrix, is the product that comprises second triangular matrix with the second channel matrix decomposition between via node and the destination node.Then, the signal times that will receive at via node is with unitary matrice.Then.Use first triangular matrix from the signal that is received, to detect the transmission signal that sends from source node.Then.With detected transmission signal times with the transformation matrix and second unitary matrice.Resulting signal is sent to the destination node from via node.

In this example, use second triangular matrix, in destination node detection of transmitted signals from the described signal that obtains.

This method helps preventing effectively the noise storage of each jumping at each via node.Because the destination node does not need to carry out unitary transformation, so can reduce workload in the signal processing of destination node.

In another example, the communication node signal relay that will send from particular source node under the environment that carries out radio communication between multiple source node and the destination node is to target destination node.This communication node is estimated first unitary matrice based on one or more channel matrix between via node and one or more source node except desirable source node, take advantage of with first unitary matrice, second unitary matrice by the signal times that will be received to generate repeating signal, and this repeating signal is sent to the destination node.This first unitary matrice comprises by one or more channel matrix between via node and the source node except desirable source node is carried out the matrix that singular value decomposition obtains.

Take advantage of with first unitary matrice by the signal times that will be received, can will be separated from the transmission signal of desirable source node and signal component from other source nodes.In other words, interference can be removed, but interference can not be removed from desirable source node from other source nodes.On the contrary, can not cause noise component(s) to amplify,, and not be exaggerated so the noise component(s) in the received signal is remained lower because the signal that is received and first unitary matrice multiply each other.

Second unitary matrice comprises by one or more channel matrix between via node and the destination node except the node of target destination is carried out the matrix that singular value decomposition obtains.Make the destination node transmission signal of desirable source node can be separated with the signal component of other source node signal and multiplying each other of second unitary matrice.

In another example, communication node is also estimated transformation matrix, and this transformation matrix is the product of following matrix and one or more unitary matrice, and in this matrix, if capable number is not setting with row sum (i+j), then the matrix element at the capable j row of i is 0.In this case, communication node will be by sending to the destination node to the signal times that is received with the repeating signal that first unitary matrice, transformation matrix and second unitary matrice generate.

In another example, communication node is estimated transformation matrix, this transformation matrix is the product of diagonal matrix and unitary matrice, in this case, this communication node will be by sending to the destination node to the signal times that is received with the repeating signal that first unitary matrice, transformation matrix and second unitary matrice generate.

This set has can reduce at the destination node and separates advantage from the computing workload of the transmission signal of desirable source node.

This communication node can also and comprise that a plurality of channel matrixes between the multiple source node of desirable source node estimate weighting matrix based on via node.In this case, communication node will send to target destination node by the repeating signal that the signal times that is received is generated with weighting matrix and unitary matrice.Unitary matrice comprises by one or more channel matrix between via node and the destination node except the node of target destination is carried out the matrix that singular value decomposition obtains.

In another example, this communication node by the signal times that will be received with first unitary matrice, second unitary matrice with comprise that two in the weighting matrix of Moore-Penrose inverse matrix generate repeating signal.The quality that these two matrixes are based on channel status is selected.This set makes via node to select suitable trunking plan according to channel status, and can improve the quality of the received signal at node place, destination.

(embodiment 1)

Fig. 1 represents the schematic diagram of the general structure of communication system according to an embodiment of the invention.This communication system adopts multi-hop scheme and multiple-input and multiple-output (MIMO) scheme.This communication system comprises source node 12, and the individual via node 14-1 of destination node 16 and K (K 〉=1) is to 14-K.K via node is expressed as 14-k (1≤k≤K).Use the MIMO scheme to carry out communication between source node 12 and the via node 14-k and the communication between via node 14-k and the destination node 16.The signal of carrying out from source node 12 to destination node 16 by the multi-hop scheme transmits.For brevity, in this embodiment, each in this K via node can be jumped by one signal is relayed to destination node 16 from source node 12.Yet, can increase jumping figure.

Source node 12 sends the signal that can distinguish mutually from a plurality of antennas (for example, M antenna).In this M antenna each sends the signal that is associated independently under the MIMO scheme.From the signal limiting of M antenna transmission transmission signal vector S, each signal all is a component of a vector.

In K via node 14 each is carried out prearranged signal to the signal that is received and is handled all from source node 12 received signals, generating repeating signal, and this repeating signal is sent to destination node 16.This K via node 14 has identical 26S Proteasome Structure and Function, will describe its 26S Proteasome Structure and Function below.

Destination node 16 receives repeating signal from this K via node 14, and detects from the content of the transmission signal vector S of source node 12 transmissions.

Fig. 2 is the block diagram of via node 14-k.Via node 14-k has a plurality of antenna 22-1 to 22-M, receiving element 24, channel estimator 25, repeating signal maker 26 and transmitting element 28.Because source node 12 and destination node 16 also can be via nodes, so this structure not only can be applied to via node 14, but also can be applied to source node 12 and destination node 16.

In this embodiment, for simple and clear purpose, suppose that each in 14-K and the destination node 16 of source node 12, via node 14-1 all has and be used to send and M antenna of received signal.Yet these nodes can have the antenna of varying number, in addition, can also use the antenna of varying number in the transmission of signal and receiving course.

24 couples of signal Y that receive to the 22-M place at M antenna 22-1 of receiving element _kThe execution appropriate signals is handled.The sort signal processing comprises receiving front-end processing (for example frequency inverted and bandwidth constraints) and each antenna is weighted.The signal Y that is received _kBe expressed as by the vector of forming with M the corresponding M of an antenna component.Receiving element 24 is also analyzed received signal Y _kHead, to determine to send the destination node of signal to it.Do not reach the destination node if signal jumps to by one, then via node 14-k sends to another via node with this signal.

The channel matrix H that channel estimator 25 is estimated between source node 12 and the via node 14-k _kBy receiving each pilot channel that sends from source node 12, can obtain channel matrix H _kMatrix element.Similarly, the channel matrix G between channel estimator 25 estimation via node 14-k and the destination node 16 _kWhen needed, channel estimator 25 is also estimated channel status.For example can be by estimating wireless channel state according to the signal measurement SNR or the SIR that are received.Can use the rank of channel status in the following embodiments.

Repeating signal maker 26 is according to the signal Y that is received _kGenerate repeating signal X with channel estimation results _kRepeating signal X _kBy the vector of forming with M the corresponding M of an antenna component.Repeating signal maker 26 will be described in detail belows.

Transmitting element 28 is carried out signal processing, to pass through a plurality of antennas with repeating signal X _kSend to destination node 16.This signal processing comprises frequency inverted, bandwidth constraints, power amplification and each antenna is weighted.

Fig. 3 is the functional block diagram of repeating signal maker 26.Repeating signal maker 26 has QR resolving cell 32, weighted factor computing unit 34 and weighted units 36.

When receiving and channel matrix H from channel estimator 25 _kAnd G _kDuring relevant information, QR resolving cell 32 is with channel matrix H _kBe decomposed into unitary matrice Q _kWith triangular matrix R _kThe form of product.As a result, determined to satisfy the unitary matrice Q of equation (7) _kWith triangular matrix R _k

H _k＝Q _kR _k (7)

Should be noted that triangular matrix R _kIn i capable first to (i-1) column element be 0 (2≤i≤M), be expressed as follows by equation (8):

QR resolving cell 32 is also with channel matrix G _kBe decomposed into triangular matrix P by equation (9) expression _k ^HWith unitary matrice O _k ^HThe form of product, wherein subscript H represents to grip altogether transposition.

G _k＝P _k ^HO _k ^H (9)

Should be noted that triangular matrices P _kIn i capable first to (i-1) column element be 0 (2≤i≤M), be expressed as follows by equation (10):

Because matrix P _kBe upper triangular matrix, so P _k ^HIt is lower triangular matrix.

According to channel matrix H _kAnd G _kAnd the QR breakdown, weighted factor computing unit 34 calculates received signal Y _kWeighted factor.The computational details of weighted factor is described below in conjunction with the operation of communication system.

Weighted units 36 is carried out the predetermined matrix computings, with received signal Y _kBe converted to repeating signal X _k

Fig. 4 represents the flow chart of the operation of communication system according to an embodiment of the invention.In this communication system, the transmission signal vector S that source node 12 will be made up of the set of M signal component from M antenna transmission to around via node.The via node that is positioned at preset range is from this source node 12 received signal S.This scope can be called 1 jumping scope.For convenience of description, suppose that K via node receives transmission signal S and carry out similar signal processing, to relay the signals to the destination node.Although (1≤k≤K), other via node is also carried out similar operation only to show k via node among Fig. 4.

At first, source node 12 and destination node 16 difference pilot signal transmitted L _kAnd Z _k, receive these pilot signals at via node 14-k place.In step 401, via node 14-k is according to pilot signal L _kAnd Z _kCarry out channel estimating, with the channel matrix H between estimation source node 12 and the via node 14-k, and the channel matrix G between via node 14-k and the destination node 16.

In step 402, the transmission signal that source node 12 will be expressed as the signal vector S that is made up of the set of M component from M antenna transmission to around via node.

In step 404, via node 14-k is from source node 12 received signals.The signal indication that is received is:

Y _k＝H _kS+n _k (11)

Wherein, H _kBe the channel matrix between source node 12 and k the via node, as mentioned above, n _kThe expression noise component(s).

In step 406, via node 14-k is in 32 pairs of channel matrix H of QR resolving cell _kAnd G _kCarry out QR and decompose (referring to Fig. 3).In this step, channel matrix H _kBe broken down into unitary matrice Q _KWith triangular matrix R _KProduct (H _k=Q _kR _k) form, and channel matrix G _kBe broken down into triangular matrix P _k ^HWith unitary matrice O _k ^HProduct (G _k=P _k ^HO _k ^H) form.

In step 408, at weighted factor computing unit 34 places according to triangular matrix P _kAnd R _kThe computational transformation matrix A _k(Fig. 3).(i+j ≠ M+1) is then at transformation matrix A if i+j is not equal to M+1 _kThe capable and j of i row in matrix element be 0.In this case, transformation matrix A _kRepresent by equation (12).

In other words, when these row and columns (contrary diagonal matrix) are set with backward, transformation matrix A _kIt is the matrix that becomes diagonal matrix.If i+j equals M+1, matrix element then ${\left(A_k\right)}_{i,M-\mathrm{<figure-callout\; id="1"\; label="i\; +"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">i</figure-callout>}+1}=a_i^k$ Be expressed as:

$a_i^k=\frac{{\left({P_k}^H\mathrm{\&Pi;}{R}_k\right)}_{i,M-i+1}^H}{\left|\right|{\left({P_k}^H\mathrm{\&Pi;}{R}_k\right)}_{i,M-i+1}^H\left|\right|}---\left(13\right)$

Wherein matrix ∏ represents commutative matrix, and it is represented by equation (14):

In step 410, generate repeating signal X _k, this repeating signal is represented by equation (15):

X _k＝E _kO _kA _kQ _k ^HY _k (15)

Coefficient E _kBe the scalar that limits by equation (16):

$E_k=\sqrt{\frac{\mathrm{PM}}{P\left[\mathrm{tr}\right\{\left(P_k^H{A}_k{R}_k\right){\left(P_k^H{A}_k{R}_k\right)}^H\left\}\right]+\mathrm{MN}{\mathrm{\&sigma;}}^2}}---\left(16\right)$

Wherein P is illustrated in total transmitted power at source node 12 places, and σ ²The expression noise level.

In step 412, with repeating signal X _kSend to destination node 16.

In step 414, at the signal of destination node 16 places reception from all via nodes, these via nodes carry out relaying to the signal from source node 12.The signal Y that will receive at destination node 16 places _RBe expressed as:

$\begin{array}{c}{Y}_R=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{G}_k{X}_k{+n}_R\\ =\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{T}_{k}S+n\end{array}---\left(17\right)$

N wherein _RRepresent noise component(s) with n.According to equation (7), (9) and (11), below relation is set up:

Q _k ^HY _k＝Q _k ^H(H _kS+n _k)

＝Q _k ^H(Q _kR _kS+n _k)

＝R _kS+Q _k ^Hn _k

In addition, according to above-mentioned relation and equation (9) and (16), following relation of plane is set up:

G _kX _k＝P _k ^HO _k ^H·E _kO _kA _kQ _k ^HY _k

＝E _kP _k ^HA _kQ _k ^HY _k

＝E _kP _k ^HA _kR _kS+E _kP _k ^HA _kQ _k ^Hn _k

=E _kT _kS+ (noise component(s))

Wherein, T _k=P _k ^HA _kR _k

Can be according to equation (8), (10) and (16) with matrix T _kBe expressed as equation (18):

Consider equation (13), be to be understood that the plain a of non-vanishing matrix element _i ^kEqual p _Ii(r _{M-i+1 M-i+1}) ^*/ | p _Ii(r _{M-i+1 M-i+1}) ^*|, wherein complex conjugate represented in asterisk.Therefore, Y _kS becomes first matrix to the M element that has by equation (19) expression.

In step 416, come detection of transmitted signals S according to equation (17) and (18).Use continuous interference inversion (to be used for eliminating continuously T _kNon-diagonal components) carry out input.Suppose and carry out continuous removing method in an ideal way,, utilize equation (20-1) to calculate the equivalent signal-to-noise ratio (λ m) that each sends stream according to channel estimation results at destination node 16 places.

${\mathrm{\&lambda;}}_m=\frac{P}{M}\frac{{\left(\underset{k=1}{\overset{k}{\mathrm{\&Sigma;}}}{\left(E_k{P_k}^H{A}_k{R}_k\right)}_{m,M-m+1}\right)}^2}{{\mathrm{\&sigma;}}_{\mathrm{<figure-callout\; id="2"\; label="r"\; filenames="A20051009384000511.png"\; state="\{\{state\}\}">r</figure-callout>}}^2\underset{k=1}{\overset{k}{\mathrm{\&Sigma;}}}{E}_k{\left|\right|{\left({P_k}^H{A}_k\right)}_m\left|\right|}^2+{\mathrm{\&sigma;}}_d^2}---(20-1)$

σ wherein _r ²And σ _d ²Be respectively noise component(s) n _kAnd n _RVariance, and P represents total transmitted power of source node 12.According to equation (20-1), as independent control flows S ₁..., S _MSpeed the time, represent message capacity C between source node 12 and the destination node 16 by equation (20-2).

$C=\underset{m=1}{\overset{\mathrm{<figure-callout\; id="1"\; label="M"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">M</figure-callout>}}{\mathrm{\&Sigma;}}}\frac{1}{2}{\log }_2(1+{\mathrm{\&lambda;}}_m)---(20-2)$

Can give source node 12 with the report information relevant by giving source node 12 from the feedback information of destination node with the speed of each stream.Can also control the power level of each stream independently.

Shown in equation (19), eliminate T _kNon-diagonal components, and the signal component S of the signal vector that will obtain from via node 14 ₁To S _MIn each multiply by an arithmetic number.The place makes up these matrix elements at the destination node.Because the coefficient that uses does not comprise imaginary component (phase component),, therefore, can realize the in-phase signal combination by maximum rate so these components need be eliminated hardly in the signal combination process in signal combination.In other words, can phase coherence the ground combination from the repeating signal of each via node 14.

Because mainly calculate scalar E according to the conversion of unitary matrice _kWith other coefficients,, can reduce the adverse effect that noise increases so compare with conventional art.From reducing the angle of the loss of signal, this set is favourable.Therefore, can solve the decline (this is problems of the prior art) of signal quality.

(embodiment 2)

Fig. 5 is the functional block diagram according to the repeating signal maker 26 that uses in via node 14 of second embodiment of the invention.Repeating signal maker 26 comprises QR resolving cell 32, weighted factor computing unit 34, first weighted units 36, signal detector 39 and second weighted units 62.In a second embodiment, destination node 16 can have 26S Proteasome Structure and Function shown in Figure 5, and perhaps alternatively, it can have 26S Proteasome Structure and Function shown in Figure 3.

When receiving and channel matrix H from channel estimator 25 _kAnd G _kDuring relevant information, QR resolving cell 32 is with channel matrix H _kBe decomposed into unitary matrice Q _kWith triangular matrix R _kProduct (H _k=Q _kR _k) form.QR resolving cell 32 is also with channel matrix G _kBe decomposed into triangular matrix P _k ^HWith unitary matrice O _k ^HProduct (G _k=P _k ^HO _k ^H) form.

According to channel matrix H _kAnd G _kAnd the information relevant with the QR breakdown, weighted factor computing unit 34 is Y to received signal _kCalculate weighted factor.

First weighted units 36 is with received signal Y _kWith the weighted factor Q that estimates by weighted factor computing unit 34 _k ^HMultiply each other, to extract each component of received signal.

Signal detector 39 detects the transmission signal S that sends from source node 12 according to weighting received signal and the information relevant with triangular matrix from weighted units 36 outputs _k=(S _K1..., S _KM).

Second weighted units 62 is with the transmission signal S that is detected _kWith the weighted factor A that calculates by weighted factor computing unit 34 _kO _k ^HMultiply each other, and output repeating signal A _kO _k ^HS _kEach component.

Fig. 6 is the flow chart of expression according to the operation of the communication system of second embodiment of the invention.

At first, source node 12 and destination node 16 difference pilot signal transmitted L _kAnd Z _k, receive this pilot signal at via node 14-k place.In step 701, via node 14-k is according to pilot signal L _kAnd Z _kCarry out channel estimating, with the channel matrix H between estimation source node 12 and the via node 14-k, and the channel matrix G between via node 14-k and the destination node 16.

In step 702, the transmission signal that source node 12 will be expressed as the signal vector S that is made up of the set of M component from M antenna transmission to around via node.

In step 704, via node 14-k is from source node 12 received signals.The signal indication that is received is:

Y _k＝H _kS+n _k

In step 706, to channel matrix H _kAnd G _kCarrying out QR decomposes.With channel matrix H _kBe decomposed into unitary matrice Q _kWith triangular matrix R _kProduct (H _k=Q _kR _k) form, and with channel matrix G _kBe decomposed into triangular matrix P _k ^HWith unitary matrice O _k ^HProduct (G _k=P _k ^HO _k ^H) form.

In step 708, by with received signal Y _kWith unitary matrice Q ^HMultiply each other and carry out unitary transformation.Will be through the received signal Z of unitary transformation _kBe expressed as:

Z _k＝Q _k ^HY _k

＝R _kS+Q _k ^Hn _k

Because matrix R _kBe upper triangular matrix, so, then descend relation of plane to set up if ignore noise.

Z _k1＝r ₁₁S ₁+r ₁₂S ₂+…+r _1MS _M

Z _k2＝r ₂₂S ₁₂+…+r _2MS _M

…

Z _kM-1＝r _M-1?M-1S _M-1+r _M-1?MS _M

Z _kM＝r _MMS _M

In step 710, according to received signal detection of transmitted signals S through unitary transformation.At first, pay close attention to M received signal component Z _KM, according to known Z _KMAnd r _MMDetection of transmitted signals component S _MPay close attention to (M-1) individual received signal component Z then _KM-1, according to known r _{M-1 M-1}, r _MMAnd S _MDetection of transmitted signals component S _M-1By similar mode, come the continuous detecting transmitting signal components.

In step 712, by the transmission signal S that will be detected _kWith A _kO _k ^HMultiply each other and carry out further conversion, wherein matrix A _kBe the diagonal matrix of following expression:

A _k＝diag(P _k ^H)

In step 714, with figure signal O _k ^HS _kSend to destination node 16 as repeating signal.

In step 716, at the signal of destination node 16 places reception from all relevant via node 14 relayings.Received signal Y _RBe expressed as:

$\begin{array}{c}{Y}_R=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{G}_k{A}_k{O_k}^{H}S+n\\ =\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{P}_k\mathrm{diag}\left({P_k}^H\right)S+n\\ =\mathrm{DS}+n\end{array}---\left(21\right)$

Wherein n represents noise component(s).Equation (21) has utilized can be with channel matrix G _kBe decomposed into G _k=P _kO _kThe fact of form.Because P _kBe triangular matrix, so K matrix P _kSummation (or combination) also be triangular matrix.Combined result is expressed as matrix D (having element dij).Can decompose to determine and triangular matrix P by carry out QR at destination node 16 _kWith unitary matrice O _kRelevant information perhaps alternatively, can be gathered these information from each via node 14.If ignored noise component(s), then equation (21) is expanded into following form.

Y _R1＝d ₁₁S ₁+d ₁₂S ₂+…+d _1MS _M

Y _R2＝d ₂₂S ₂+…+d _2MS _M

…

Y _RM-1＝d _M-1?M-1S _M-1+d _M-1?MS _M

Y _RM＝d _MMS _M

In step 718, at the via node 14 detection of transmitted signals S of place.At first, pay close attention to M received signal component Y _RM, according to known Z _RMAnd d _MMDetection of transmitted signals component S _MThen, pay close attention to (M-1) individual received signal component Y _RM-1, according to known d _{M-1 M-1}, d _{M-1 M}And S _MCome detection of transmitted signals component S _M-1By similar mode, the continuous detecting transmitting signal components.

In a second embodiment, destination node 16 is not to carry out unitary transformation in the step 716 of Fig. 6.

(embodiment 3)

Fig. 7 A and Fig. 7 B are the curve chart of expression according to the simulation result of the signal transmission of the embodiment of the invention.Trunnion axis is represented power noise than (PNR), and vertical axis is represented capacity.In Fig. 7 A, the quantity of transmitting antenna and the quantity of reception antenna are respectively four, and two via nodes (K=2) are between source node and destination node, and jump in the communication range one.The curve representation of one theory is as the one theory of the capacity of PNR function, and the capacity of the curve representation of prior art when using the ZF method that signal is carried out relaying.Obtain the curve of embodiment 1 by the method for implementing first embodiment.In Fig. 7 B, the quantity of transmitting antenna and the quantity of reception antenna are respectively four, and four via nodes (K=4) are between source node and destination node, and jump in the communication range one.According to the curve chart of Fig. 7 A and Fig. 7 B, be appreciated that when transmitted power increases power system capacity increases, and aspect the realization capacity sufficient, the method for embodiment 1 is better than traditional method.

(embodiment 4)

In the 4th embodiment,, between multiple source node and a plurality of destinations node, carry out relaying to sending signal by one or more via node.

Fig. 8 is the schematic diagram according to the wireless communication system of fourth embodiment of the invention.This system comprises: L source node (802-1 is to 802-L), each source node have M antenna; K via node (804-1 is to 804-K), each via node have N antenna; With L destination node (806-1 is to 806-L), each destination node has M antenna.Satisfied N 〉=the L*M that concerns of Integer N, M and L.In this example, in order to simplify, all source nodes and destination node all have M antenna, and all via nodes have N antenna.Certainly, these nodes can have the antenna of varying number, as long as the antenna amount of source node is equal to or less than the antenna amount of destination node.

As above described in conjunction with Fig. 1, at source node 802-l with have channel status between the via node 804-k of N antenna by N * M channel matrix H with M antenna _{L, k}Expression.Similarly, at via node 804-k with have channel status between the destination node 806-l of M antenna by M * N channel matrix G _{K, l}(reduced representation becomes G _Kl) expression.

By via node receive and relaying from the transmission signal of multiple source node.By the signal of the destination node reception that signal was addressed to that sends from source node from a plurality of via nodes, and recovery is from the signal of this source node.Therefore, except being subjected to influence, also be subjected to the influence (interference) of the signal that sends from other source node at signal that the destination node receives from the transmission signal of desirable source node.The destination node must detect desirable transmission signal by removing to disturb.

Before the signal processing of describing the 4th embodiment, the general signal to legacy communications system (for example, describing in people's such as above-mentioned Rohit U.Nabar article) is described earlier.

Fig. 9 is the functional block diagram of one of traditional via node (k via node).This via node have L receiving filter 902-1 that the quantity L with source node is provided with accordingly to 902-L, L transmitting filter 904-1 to 904-L, and signal merge cells 906.

Received signal Y with the via node place _kDistribute to L receiving filter 902-1 to 902-L.Because received signal Y _kComprise signal from L source node, so it is expressed as:

$y_k=\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{H}_{l,k}{s}_l+n---\left(30\right)$

(N * 1 matrix)

Wherein, S _lBe M the signal component (S that have from the transmission of the 1st source node _L1, S _L2..., S _LM) the transmission signal phasor, and n _kBe illustrated in the noise component(s) of introducing between k via node and the multiple source node.The dimension of received signal is N * 1.

The received signal y that the 1st receiving filter 902-1 will be represented by M component of a vector _kMultiply by weighting matrix w ^b _KlThis weighting matrix w ^b _KlBe matrix and the satisfied relation of M * N:

${[W_{k,1}^{\mathrm{bT}}...W_{k,l}^{\mathrm{bT}}...W_{k,L}^{\mathrm{bT}}]}^T={(H_k^H\&CenterDot;H_k)}^{-1}\&CenterDot;H_k^H---(31-1)$

(ML * N matrix)

This relational expression is represented the matrix of ML * N.H _kBe the matrix that comprises a plurality of channel matrixes, and be defined as:

H _k＝[H _lk，…，H _Lk] (31-2)

According to equation (31-1) and (31-2) be appreciated that w ^b _KlAnd H _LkMutually orthogonal.Shown in equation (32), utilize this orthogonality, receiving filter 902-1 is with received signal y _kMultiply by weighting matrix w ^b _KlSo that the received signal vector is converted to y ' _Kl

$\begin{array}{c}{y^{\&prime;}}_{k,l}=W_{k,l}^b{y}_k\\ =s_l+W_{k,l}^b{n}_k\end{array}---\left(32\right)$

(M * 1 matrix)

Then, transmitting filter 904-l will carry out the received signal y ' of conversion _KlMultiply by another weighting matrix w ^f _KlThis weighting matrix w ^f _KlBe the matrix of N * M, and satisfy relation:

$[W_{k,1}^f...W_{k,l}^f...W_{\mathrm{kL}}^f]={G_k^H\left(G_k{\&CenterDot;G}_k^H\right)}^{-1}---\left(33\right)$

(N * ML matrix)

This relational expression is represented the matrix of N * ML.G _kBe the matrix that comprises a plurality of channel matrixes, and be defined as:

G _k＝[G _lk，…，G _Lk] (34)

Will be through multiplied signals w ^f _Kl* y ' _KlOffer signal merge cells 906.This signal merge cells 906 merges from transmitting filter 904-1 to the output signal of 904-L with generation repeating signal x _kThis repeating signal x _kBe expressed as:

$x_k=E_k{W}_{k,l}^f\&CenterDot;y_{k,l}^{\&prime;}---\left(35\right)$

(N * 1 matrix)

Wherein, E _kBe to be used for the transmitted power of via node is carried out normalized scalar.With this repeating signal x _kSend to the destination node.

In the middle of the node of a plurality of destinations, l destination node 806-l receives the signal from K via node, and each signal all reflects from l source node transmission signal that send and that be addressed to l destination node.Therefore, the signal r that receives at l destination node ₁Be represented as:

$\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{G}_{k,l}{x}_1\\ =\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}(E_k{s}_l+E_k{W}_{k,l}^b{n}_k)+z_l\end{array}---\left(36\right)$

(M * 1 matrix)

Wherein, z _lBe illustrated in the noise component(s) of introducing between a plurality of via nodes and the l destination node.Utilize channel matrix G _KlWith weighting matrix W ^f _KlBetween orthogonality relation, estimate in equation (36) the received signal r of definition ₁

In equation (36), received signal r ₁Each component depend on desirable transmission signal s linearly _lThe respective signal component.Therefore, can directly detect desirable transmission signal s according to received signal _l, and need not to carry out the Signal Separation of the complexity of in the MIMO scheme, carrying out usually.

Yet, by this method, by weighted factor w ^b _KlAmplify noise n _k, therefore, be the problem that merits attention in the decline of the received signal quality of destination node.Can pass through coefficient E _kBe provided with to such an extent that lessly reduce the contribution that weighting matrix amplifies noise.Yet, because coefficient E _kAlso be used for desirable signal s _lSo, along with coefficient E _kReduce, desirable signal component also diminishes.By traditional technology, the input precision at destination node place can reduce.

Figure 10 is the functional block diagram of expression according to the via node of fourth embodiment of the invention.This via node is one of via node shown in Fig. 8 (k via node 804-k).Other via node also has identical 26S Proteasome Structure and Function.This via node 804-k have L receiving filter 1002-1 to 1002-L, L receiving filter estimator 1004-1 to 1004-L, L medial filter 1006-1 to 1006-L, L medial filter estimator 1008-1 to 1008-L, L transmitting filter 1010-1 to 1010-L, L transmitting filter estimator 1012-1 to 1012-L and signal merge cells 1014.

Figure 11 is illustrated in the arithmetical operation of carrying out among l receiving filter estimator 1004-l, l medial filter estimator 1008-l and the l transmitting filter estimator 1012-l.

As shown in figure 10, will be at the signal y of k via node reception _kDistribute to L receiving filter 1002-1 to 1002-L.Because received signal y _kComprise signal, so it is represented by above-mentioned equation (30) from L source node.

L (the received signal y that the receiving filter 1002-l of 1≤l≤L) will be represented by M component of a vector _kMultiply by the first unitary matrice U _KlThis first unitary matrice has the dimension (N 〉=LM), and estimate this first unitary matrice by receiving filter estimator 1004-l of the capable and N-M (L-1) of N row.

In the middle of L channel matrix between (k) the via node 804-k that is paid close attention to and L the source node, l receiving filter estimator 1004-l thinks matrix H ^(l) _kComprise L-1 channel matrix between the 1st source node and via node 804-k,, it is represented as:

H ^(l) _k＝[H _l，k，…，H _l-1，k，H _l+1，k，…，H _L，k] (37)

Should be noted that different with equation (31-2), matrix H ^(l) _kDo not comprise channel matrix H _LkTherefore, H ^(l) _kHas the dimension that N is capable and M (L-1) is listed as.Shown in equation (38), by to matrix H ^(l) _kCarry out singular value decomposition and obtain the above-mentioned first unitary matrice U _Kl

(N * M (L-1) matrix)

In equation (38), Λ ^(l) _{K, l}..., Λ ^(l) _{K, L-1}In each be M * M diagonal matrix, and their diagonal components is H ^(l) _kSingular value.Matrix [U ^(l) _{K, l}..., U ^(l) _{K, L-1}] have the dimension that N is capable and M (L-1) is listed as, and comprise by matrix H ^(l) _kThe base vector of the signal space of definition.Similarly, [V ^(l) _{K, l}..., V ^(l) _{K, L-1}] ^TComprise by matrix H ^(l) _kThe base vector of signal space of definition, and represent by the square formation of M (L-1) * M (L-1).U _{K, l}Be to have first unitary matrice that N is capable and N-M (L-1) is listed as.This matrix is corresponding with the base vector of the kernel of above-mentioned signal space.

Shown in equation (39), the 1st receiving filter 1002-1 is with received signal y _KlMultiply by the first unitary matrice U ^H _Kl, to convert received signal vector to y ' _{K, l}

$\begin{array}{c}{y}_{k,l}^{\&prime;}=U_{k,l}^H{y}_k\\ =U_{k,l}^H{H}_{l,k}{s}_l+U_{k,l}^H{n}_k\end{array}---\left(39\right)$

Because the first unitary matrice U _KlWith by H ^(l) _kThe base vector of kernel of the signal space of definition is corresponding, so when received signal be multiply by first unitary matrice, can be with from the transmission signal of l source node and from the transmission Signal Separation of other source node.Should be noted that differently, do not have to remove between the signal component that sends from the l source node and disturb with equation (32).On the contrary, in this stage, prevented noise component(s) n _kAmplification.

L medial filter 1006-l will carry out the received signal y ' of conversion _KlMultiply by transformation matrix Φ _Kl, this transformation matrix Φ _KlGenerate by medial filter estimator 1008-l.Yet, in this embodiment, transformation matrix Φ _KlBe unit matrix, therefore, medial filter 1006-l and medial filter estimator 1008-l do not carry out concrete processing.Certainly, following described in another embodiment, medial filter estimator 1008-l can generate the matrix different with unit matrix.

L transmitting filter 1010-1 will carry out the received signal y ' of conversion _KlMultiply by the second unitary matrice A _KlThis second unitary matrice has the dimension (N 〉=LM), and generated by transmitting filter estimator 1012-l of the capable and N-M (L-1) of N row.

In L the channel matrix between (k) the via node 804-k that is paid close attention to and L destination node, l transmitting filter estimator 1012-l thinks matrix G ^(l) _kComprise L-1 channel matrix between l destination node and via node 804-k, this matrix is represented as:

${G^{\left(l\right)}}_k=[{\mathrm{<figure-callout\; id="1"\; label="G"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">G</figure-callout>}}_{1,k}^H,...,G_{l-1,k}^H,G_{l+1,k}^H...G_{L,k}^H]$

Should be noted that different with equation (34), matrix G ^(l) _kDo not comprise channel matrix G _LkTherefore, G ^(l) _kHas the dimension that N is capable and M (L-1) is listed as.Shown in equation (40), by to matrix G ^(l) _kCarry out singular value decomposition and obtain the above-mentioned second unitary matrice A _Kl

(N * M (L-1) matrix)

In equation (40), Ω ^(l) _{K, l}..., Ω ^(l) _{K, L-1}In each all be M * M diagonal matrix, and its diagonal element is G ^(l) _kSingular value.Matrix [A ^(l) _{K, l}..., A ^(l) _{K, L-1}] have the dimension that N is capable and M (L-1) is listed as, and comprise by matrix G ^(l) _kThe base vector of the signal space of definition.Similarly, [B ^(l) _{K, l}..., B ^(l) _{K, L-1}] ^TComprise by matrix G ^(l) _kThe base vector of signal space of definition, and represent by the square formation of M (L-1) * M (L-1).A _{K, l}It is second unitary matrice with dimension of the capable and N-M (L-1) of N row.This matrix is corresponding with the base vector of the kernel of above-mentioned signal space.

L transmitting filter 1010-1 is with signal y ' _KlMultiply by the second unitary matrice A _KlSignal A after will multiplying each other _KlY ' _{K, l}Offer signal merge cells 1014.Signal merge cells 1014 merges the signal of the output from transmitting filter 1010-1 to 1010-L to generate repeating signal x _kThis repeating signal x _kBe represented as:

$\begin{array}{c}{x}_k=E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{k,l}{U}_{k,l}^H\&CenterDot;y_k\\ =E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{k,l}{U}_{\mathrm{kl}}^H{s}_l+E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{U}_{\mathrm{kl}}^H{n}_k\end{array}$

(N * 1 matrix)

E wherein _kBe to be used for the transmitted power of via node 804-k is carried out normalized scalar.With this repeating signal x _kSend to the destination node.

In the middle of the node of a plurality of destinations, the l destination node 806-l that is addressed to from the transmission signal of l source node 802-l receives K repeating signal from K via node.The signal r that will receive at l destination node 806-l _lBe expressed as:

$\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{G}_{k,l}{x}_k\\ =\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{G}_{k,l}{A}_{k,l}{U}_{k,l}^H{H}_{l,k}{s}_l+\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{G}_{k,l}{A}_{k,l}{U}_{k,l}^H{n}_k{+z}_l\end{array}---\left(41\right)$

(M * 1 matrix)

Z wherein _lIt is the noise component(s) of between a plurality of via nodes and l destination node, introducing.Utilize the following fact to come estimate equation (41): if l ≠ l ', then channel matrix G _KlWith the second unitary matrice A _{Kl '}Mutually orthogonal.If l=l ' is then by G _KlA _KlThe matrix of expression is the common matrix beyond the unit matrix.

As equation (41) clearly shown in, at received signal r _lIn, will be from the transmission signal s of desirable source node _lWith transmission signal s from other source node _l(l ≠ l ') separate.In other words, fully reduced interference between the source node; Yet, also exist to remain in from the interference between a plurality of signal components in the transmission signal of desirable source node.This be because, usually, by G _KlA _KlU ^H _KlH _LkThe matrix of expression is not a diagonal matrix.Therefore, the destination node must be carried out by the normal signal of carrying out with the MIMO scheme and separate, to detect desirable signal s from received signal _lCompare with conventional art, it is complicated a little that this input itself may become.

Yet this method has can prevent noise n at the via node place _kThe advantage of amplification.In equation (41), G is and noise n _kThe matrix that must introduce in the middle of those matrixes that multiply each other.Because matrix A _KlAnd U _KlBe unitary matrice, so these matrixes do not amplify noise.Therefore, do not need to adopt the same little coefficient E with equation (36) _k, in equation (36), noise is by weighting matrix W ^b _KlAmplify.This means to eliminate or to reduce by present embodiment and reduce as the input precision of the problem of being paid close attention in the conventional art.

(embodiment 5)

Figure 12 is illustrated in another example of the arithmetical operation of carrying out among l receiving filter estimator 1004-l, l medial filter estimator 1008-l, the l transmitting filter estimator 1012-l.Identical among the computing of carrying out by receiving filter 1002-1, receiving filter estimator 1004-l, transmitting filter 1010-l and transmitting filter estimator 1012-l and the 4th embodiment.

In the 5th embodiment, l medial filter 1006-1 will be from the signal y ' of receiving filter 1004-1 output _KlMultiply by transformation matrix Φ _Kl, to generate signal Phi _KlY ' _KlThis transformation matrix Φ _KlCalculate by l medial filter estimator 1008-1.

Shown in equation (50), this medial filter estimator 1008-1 is to matrix U ^H _KlH _LkCarrying out QR decomposes.

U ^H _klH _lk＝Q _1klR _1kl (50)

Wherein, Q _1klBe unitary matrice with dimension of N-M (L-1) row and M row, and R _1klIt is the upper right triangular matrix of M * M.Similarly, shown in equation (51), medial filter estimator 1008-1 is to matrix (G ^H _KlA _Lk) ^HCarrying out QR decomposes.

(G ^H _klA _lk) ^H＝Q _2klR _2kl (51)

Wherein, Q _2klBe unitary matrice with dimension of N-M (L-1) row and M row, and R _2klIt is the upper right triangular matrix of M * M.Medial filter estimator 1008-1 also uses the triangular matrix that satisfies equation (50) and (51) to come estimated matrix Θ _KlThis matrix Θ _KlBe represented as:

Wherein, matrix П is defined as:

Medial filter estimator 1008-1 uses these estimated matrix to estimate transformation matrix Φ _Kl, it is defined as:

Φ _kl＝Q _2klΘ _klQ ^H _1kl (53)

This matrix Φ _KlBe (N-M (L-1)) * matrix of (N-M (L-1)).

L medial filter 1006-1 is with signal Phi _KlY ' _KlOutput to transmitting filter 1010-1.Transmitting filter 1010-l multiply by input signal the matrix A of describing in the 4th embodiment _Kl, and the signal after will multiplying each other outputs to signal merge cells 1014.This signal merge cells 1014 will be from the signal plus of L transmitting filter 1010-1 to 1010-L, and output repeating signal x _kThis repeating signal x _kBe represented as:

$\begin{array}{c}{X}_k=E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{\mathrm{\&Phi;}}_{\mathrm{kl}}{y}_{\mathrm{kl}}^{\&prime;}\\ =E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}\left({\mathrm{<figure-callout\; id="2"\; label="Q"\; filenames="A20051009384000511.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{2\mathrm{kl}}{\mathrm{\&Theta;}}_{\mathrm{kl}}{\mathrm{<figure-callout\; id="1"\; label="Q"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{1\mathrm{kl}}^H\right)\left(U_{\mathrm{kl}}^H{H}_{\mathrm{lk}}{s}_l{+U}_{\mathrm{kl}}^H{n}_k\right)\\ =E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{\mathrm{<figure-callout\; id="2"\; label="Q"\; filenames="A20051009384000511.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{2\mathrm{kl}}{\mathrm{\&Theta;}}_{\mathrm{kl}}\left({\mathrm{<figure-callout\; id="1"\; label="R"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">R</figure-callout>}}_{1\mathrm{kl}}{s}_l{+\mathrm{<figure-callout\; id="1"\; label="Q"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{1\mathrm{kl}}^H{U}_{\mathrm{kl}}^H{n}_k\right)\\ =E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{\mathrm{<figure-callout\; id="2"\; label="Q"\; filenames="A20051009384000511.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{2\mathrm{kl}}{\mathrm{\&Theta;}}_{\mathrm{kl}}{\mathrm{<figure-callout\; id="1"\; label="R"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">R</figure-callout>}}_{1\mathrm{kl}}{s}_l+E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{\mathrm{\&Phi;}}_{\mathrm{kl}}{U}_{\mathrm{kl}}^H{n}_k\end{array}---\left(54\right)$

Wherein, E _kBe to be used for the transmitted power of via node 804-k is carried out normalized scalar.Estimating repeating signal x _kProcess in, used equation (39) and (50).With the estimated repeating signal x that goes out _kSend to the destination node.

In the middle of the node of a plurality of destinations, the l destination node 806-l that is addressed to from the transmission signal of l source node 802-l receives the repeating signal from K via node.Therefore, the signal r that receives at l destination node 806-l _lBe represented as:

$\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{G}_{k,l}{x}_k\\ =\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{R}_{2k,l}^H{{\mathrm{\&Theta;}}_{k,l}R}_{1k,l}{s}_l+\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{G}_{k,l}{A}_{k,l}{\mathrm{\&Phi;}}_{k,l}{U}_{k,l}^H{n}_k{+z}_l\end{array}---\left(55\right)$

(M * 1 matrix)

Z wherein _lIt is the noise component(s) of between a plurality of via nodes and l destination node, introducing.Utilize the following fact to come estimate equation (55): if l ≠ l ', then channel matrix G _KlWith the second unitary matrice A _{Kl '}Mutually orthogonal.The fact of also having utilized equation when l=l ' (51) to set up.

Shown in equation (55) is clear, at received signal r _lIn, will be from the transmission signal s of desirable source node _lWith transmission signal s from other source node _{L '}(l ≠ l ') separate.In other words, fully reduced interference between the multiple source node; Yet, also exist to remain in from the interference between a plurality of signal components in the transmission signal of desirable source node.Therefore, the destination node must be carried out the normal signal of carrying out with the MIMO scheme usually and separate, to detect desirable signal s from received signal _l

Incidentally, as what understand, be used to send signal s according to the definition of each matrix _lMatrix Q ^H _2kΘ _KlR _1klIt is the bottom right triangular matrix.Therefore, if determined only to depend on one of signal component of upper right triangular matrix element (for example, s _LM), then can one by one determine to send signal s continuously _lSignal component.Therefore, compare, can reduce the computing workload of signal operation with the 4th embodiment.

In addition, the matrix element of arranging to the lower left corner from the upper right corner of bottom right triangular matrix (that is, row number and row sum (i+j) equal those elements of stated number (row number add 1)) is to transmission signal s _lInfluence bigger than other matrix element.This matrix element is an arithmetic number, and does not comprise imaginary component.Therefore, with combined contribution ∑ E from L via node _kR ^H _2kΘ _KlR _1klMajor part, and can improve power signal-to-noise ratio at the destination node.In addition, because matrix A _KlAnd U _KlBe unitary matrice, so noise n _kDo not amplified by these matrixes.Therefore, can further improve precision at destination node detection signal.

Figure 15 represents compared with prior art, the curve chart of the simulation result of the 5th embodiment.This curve chart shows as the traversal capacity of power noise than the function of (PNR).Under the situation of the quantity K=2 of via node and K=8, the method for the 5th embodiment and the method for prior art are carried out emulation.The quantity of the quantity of source node and destination node also is two.The antenna amount of source node and destination node is four (4) individual, and the antenna amount of via node is eight (8) individual.Usually, and when PNR increases (, when signal power level increases), capacity increases.When the quantity of relaying node increased, capacity increased.As clear illustrating in curve chart, the technology of the 5th embodiment is better than prior art, shows as under the situation of the via node of equal number, and capacity improves about 5bps/Hz.

(embodiment 6)

Figure 13 is illustrated in another example of the arithmetical operation of carrying out among l receiving filter estimator 1004-l, l medial filter estimator 1008-l, the l transmitting filter estimator 1012-l.The computing of carrying out by receiving filter 1002-1, receiving filter estimator 1004-l identical with described in the 4th embodiment.The computing of being carried out by transmitting filter 1010-1 and transmitting filter estimator 1012-l is identical with known technology.

In the 6th embodiment, l medial filter 1006-1 will be from the signal y ' of receiving filter 1004-1 output _KlMultiply by transformation matrix Φ _Kl, to generate signal Phi _KlY ' _KlThis transformation matrix Φ _KlCalculate by l medial filter estimator 1008-1.

Shown in equation (60), this medial filter estimator 1008-1 is to matrix U ^H _KlH _LkCarrying out QR decomposes.

U ^H _klH _lk＝Q _1klR _1kl (60)

Wherein, Q _1klBe unitary matrice with dimension of N-M (L-1) row and M row, R _1klIt is the upper right triangular matrix of M * M.Medial filter estimator 1008-1 uses the triangular matrix that satisfies equation (60) to come estimated matrix Θ _Kl, it is represented as:

Medial filter estimator 1008-1 finally estimates transformation matrix Φ based on above-mentioned matrix _Kl, it is represented as:

Φ _kl＝Θ _klQ ^H _1kl (61)

Φ wherein _KlIt is the matrix of M * (N-M (L-1)).

L medial filter 1006-1 is with signal Phi _KlY ' _KlOutput to transmitting filter 1010-l.The computing of being carried out by transmitting filter 1010-1 and transmitting filter estimator 1012-l is identical with routine techniques, therefore, determines L matrix A _Kl(l=1 ..., L) (w ^f _Kl=A _Kl) to satisfy:

[A _kl，…，A _lL]＝G ^H _k(G _kG _k ^H) ^-1，

G wherein _kBe defined as:

G _k＝[G ^H _kl，…，G ^H _kL]。

Transmitting filter 1010-1 multiply by matrix A with input signal _Kl, and the signal behind the output multiplication.

The output of transmitting filter 1010-l links to each other with the input of signal merge cells 1014.This signal merge cells 1014 generates repeating signal x _kThis repeating signal x _kBe represented as:

$\begin{array}{c}{X}_k=E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{\mathrm{\&Phi;}}_{\mathrm{kl}}{y}_{\mathrm{kl}}^{\&prime;}\\ =E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}\left({\mathrm{\&Theta;}}_{\mathrm{kl}}{\mathrm{<figure-callout\; id="1"\; label="Q"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">Q</figure-callout>}}_{1\mathrm{kl}}^H\right)\left(U_{\mathrm{kl}}^H{H}_{\mathrm{lk}}{s}_l{+U}_{\mathrm{kl}}^H{n}_k\right)\\ =E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{\mathrm{\&Theta;}}_{\mathrm{kl}}{\mathrm{<figure-callout\; id="1"\; label="R"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">R</figure-callout>}}_{1\mathrm{kl}}{s}_l+E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{\mathrm{kl}}{\mathrm{\&Phi;}}_{\mathrm{kl}}{U}_{\mathrm{kl}}^H{n}_k\end{array}---\left(62\right)$

Wherein, E _kBe to be used for the transmitted power of via node 804-k is carried out normalized scalar.Estimating repeating signal x _kProcess in, used equation (39) and (60).With the estimated repeating signal x that goes out _kSend to the destination node.

In the middle of the node of a plurality of destinations, the l destination node 806-l that is addressed to from the transmission signal of l source node 802-l receives the repeating signal from K via node.Therefore, the signal r that receives at l destination node 806-l _lBe represented as:

$\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{G}_{k,l}{x}_k\\ =\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{\mathrm{\&Theta;}}_{k,l}{R}_{1k,l}{s}_l+\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{\mathrm{\&Phi;}}_{k,l}{U}_{k,l}^H{n}_k{+z}_l\end{array}---\left(63\right)$

(M * 1 matrix)

Wherein, z _lIt is the noise component(s) of between a plurality of via nodes and l destination node, introducing.Utilize the following fact to come estimate equation (63): channel matrix G _KlWith unitary matrice A _{Kl '}Mutually orthogonal.

Shown in equation (63) is clear, at received signal r _lIn, will be from the transmission signal s of desirable source node _lWith transmission signal s from other source node _{L '}(l ≠ l ') separates, thereby fully reduces the interference between the multiple source node.Yet, also exist to remain in from the interference between a plurality of signal components in the transmission signal of desirable source node.Therefore, the destination node must be carried out the normal signal of carrying out with the MIMO scheme usually and separate, to detect desirable signal s from received signal _l

Incidentally, as what understand, be used to send signal s according to the definition of each matrix _lMatrix Θ _KlR _1klIt is upper right triangular matrix.Therefore, if determined only to depend on one of signal component of bottom right matrix element (for example, s _LM), then can one by one determine to send signal s continuously _lSignal component.Therefore, compare, can reduce the computing workload of signal operation with the 4th embodiment.

In addition, matrix Θ _KlR _1klDiagonal element to sending signal s _lInfluence bigger than other matrix element.This matrix element is an arithmetic number, and does not comprise imaginary component.Therefore, homophase merges the contribution ∑ E from L via node _kΘ _KlR _1klMajor part, and can improve power signal-to-noise ratio at the destination node.In addition, because matrix A _KlAnd U _KlBe unitary matrice, so noise n _kDo not increased by these matrixes.Therefore, can further improve input precision at the destination node.

(embodiment 7)

In embodiment 4,5 and 6, used and passed through received signal y _kMultiply by unitary matrice U ^H _KlAnd the signal y ' that obtains _Kl, this unitary matrice U ^H _KlEstimate by singular value decomposition.Yet the signal processing of describing in embodiment 4,5 and 6 can be applied to by with received signal y _kMultiply by weighting matrix W ^b _KlAnd the signal y ' that obtains _Kl, as (W in the routine techniques ^b _KlY _k=s _l+ W ^b _Kln _k).

In this case, transmitting filter 1010-l can export signal y ' _Kl(=W ^b _Kly _k) multiply by at the unitary matrice A described in the 4th embodiment _KlAnd the signal that generates.

Alternatively, transmitting filter 1010-l can export signal y ' _Kl(=W ^b _Kly _k) multiply by the matrix Φ described in the 5th embodiment _Kl(=Q _2klΘ _KlQ ^H _Kl) and unitary matrice A _KlAnd the signal that generates.

Also alternatively, transmitting filter 1010-l can export signal y ' _Kl(=W ^b _Kly _k) multiply by the matrix Φ described in the 6th embodiment _Kl(equal Θ _KlQ ^H _1kl) and the 4th embodiment described in unitary matrice A _KlAnd the signal that generates.

(embodiment 8)

In the 8th embodiment, as among the 7th embodiment, with W ^b _KlBe applied to receiving filter.As shown in figure 14, the receiving filter of via node, medial filter and transmitting filter are carried out arithmetical operation to generate signal.Via node uses the method described in first embodiment to generate as the second unitary matrice A among the 4th embodiment _KlShown in equation (40),, a plurality of channel matrixes obtain this second unitary matrice A by being carried out singular value decomposition _Kl

Then, to matrix (G _KlA _Kl) ^HCarrying out QR decomposes.

(G _klA _kl) ^H＝Q _2klR _2kl

Wherein, Q _2klBe the matrix of (N-M (L-1)) * M, its column vector mutually orthogonal (in this application, being called unitary matrice), and R _2klBe the matrix of M * M, and be upper right triangular matrix.

Use this triangular matrix to estimate diagonal matrix Θ _KlThis diagonal matrix Θ _KlBe defined as:

Based on diagonal matrix Θ _KlWith unitary matrice Q _2klEstimate the transformation matrix Φ of M * M _Kl, shown in equation (65).

Φ _kl＝Q _2klΘ _kl (65)

This via node is also estimated weighting matrix W ^b _Kl, it is defined by equation (66).

${[W_{k,1}^{\mathrm{bT}}...W_{k,l}^{\mathrm{bT}}...W_{k,L}^{\mathrm{bT}}]}^T={(H_k^H\&CenterDot;H_k)}^{-1}\&CenterDot;H_k^H---\left(66\right)$

(ML * N matrix)

Equation (66) is identical with the equation that has illustrated in the 4th embodiment (31-1).

Use unitary matrice, transformation matrix and weighting matrix, generate repeating signal x by equation (67) _k, and send it to the destination node.

$x_k=E_k\underset{l=1}{\overset{L}{\mathrm{\&Sigma;}}}{A}_{k,l}{\mathrm{\&Phi;}}_{k,l}^b{W}_{k,l}^b\&CenterDot;y_k---\left(67\right)$

(N * 1 matrix)

The signal r that receives at target destination node (for convenience, being referred to as l destination node) ₁Be represented as:

$\begin{array}{c}{\mathrm{<figure-callout\; id="1"\; label="r"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">r</figure-callout>}}_1=\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{G}_{k,l}{x}_k\\ =\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{R}_{2k,l}^H{\mathrm{\&Theta;}}_{k,l}{s}_l+\underset{k=1}{\overset{K}{\mathrm{\&Sigma;}}}{E}_k{R}_{2k,l}^H{\mathrm{\&Phi;}}_{k,l}{W}_{k,l}^b{n}_k{+z}_l\end{array}---\left(68\right)$

(M * 1 matrix)

Wherein, first of the right side (R ^H _2klΘ _Kl) be the lower-left triangular matrix, and its diagonal element is an arithmetic number.Therefore, when K repeating signal merging at the destination node from K via node, in phase merge diagonal element.As a result, power noise ratio can be improved, and accurately detection of transmitted signals s of continuous interference null method can be used at the destination node _l

(embodiment 9)

In the 9th embodiment, via node 14 comes detection signal according to channel status based on method or ZF (ZF) method of using unitary matrice.When using unitary matrice, estimate unitary matrice by above-mentioned singular value decomposition.When using ZF, calculate the ZF weighted factor by the Moore-Penrose inverse matrix.

Generate repeating signal x according to channel status _k, and this repeating signal sent to the destination node.The destination node detects the transmission signal that sends from source node in above-mentioned mode.Can determine the quality of channel status at the via node place by channel estimator (Fig. 3 or Fig. 5).Alternatively, can determine the quality of channel status based on the ratio (for example SIR or SNR) of the power level of the power level of desirable ripple and undesirable ripple.

For example, via node is estimated the channel status SNR between source node and this via node _HAnd the channel status SNR between this via node and the destination node _G

If SNR _H＞＞SNR _G, then the channel status between source node and the via node is very good.Therefore, even use ZF between source node and via node, noise amplifies also enough little, and it can be ignored.On the other hand, because the influence that noise amplifies between via node and destination node increase, so between via node and destination node, use the method (similar) of using unitary matrice to Figure 14.

On the other hand, if SNR _H＜＜SNR _G, then carry out opposite processing (shown in Figure 13).

Can to those medial filters shown in Figure 14, suitably select medial filter from Figure 12.By at the quality adaptation ground change trunking plan of via node 14, can improve the quality of reception characteristic of destination node according to channel status.

Present patent application is based on No.2004-306171 that submits in the Japanese patent application No.2004-252879 that submits in August in 2004 31, on October 20th, 2004 and the No.2005-248823 that submitted on August 30th, 2005, and requires its priority.Be incorporated herein by reference in its entirety.

Claims (13)

1, a kind of wireless communication system is used for sending the transmission signal from the desirable source node of multiple source node to target destination node by via node,

Wherein said via node comprises:

The first unitary matrice estimation unit, it is configured to estimate first unitary matrice by one or more channel matrix between described via node and the described multiple source node except desirable source node is carried out singular value decomposition;

The second unitary matrice estimation unit, it is configured to estimate second unitary matrice by one or more channel matrix between described via node and the described a plurality of destinations node except the node of target destination is carried out singular value decomposition; And

Transmitting element, it is configured to and will sends to described target destination node by received signal being multiply by the repeating signal that described first and second unitary matrice generate;

Wherein, described destination node detects the transmission signal that sends from desirable source node from the repeating signal that receives.

2, a kind of communication node is used between multiple source node and a plurality of destinations node, and the transmission signal relay that will send from desirable source node is to target destination node, and this communication node comprises:

The first unitary matrice estimation unit, it is configured to estimate first unitary matrice by one or more channel matrix between described via node and the described multiple source node except desirable source node is carried out singular value decomposition;

The second unitary matrice estimation unit, it is configured to estimate second unitary matrice by one or more channel matrix between described via node and the described a plurality of destinations node except the node of target destination is carried out singular value decomposition; And

Transmitting element, it is configured to and will sends to described target destination node by received signal being multiply by the repeating signal that described first and second unitary matrice generate.

3, communication node according to claim 2 also comprises:

The transformation matrix estimation unit, it is configured to estimate transformation matrix, and this transformation matrix is made of the product of following matrix and one or more unitary matrice, and in this matrix, if i+j does not satisfy setting, then the matrix element of the capable j row of i is zero;

Wherein said transmitting element will send to described destination node by received signal being multiply by the repeating signal that described first unitary matrice, described transformation matrix and described second unitary matrice generate.

4, communication node according to claim 2 also comprises:

The transformation matrix estimation unit, it is configured to estimate transformation matrix, this transformation matrix is made of the product of an a pair of angular moment battle array and a unitary matrice, this unitary matrice be according between described source node and the described via node or the channel matrix between described via node and the described destination node derive;

Wherein said transmitting element will send to described destination node by received signal being multiply by the repeating signal that described first unitary matrice, described transformation matrix and described second unitary matrice generate.

5, a kind of communication means is used for the transmission signal that the desirable source node from the multiple source node sends is relayed to the destination node by via node, and this method may further comprise the steps:

At described via node place, estimate first unitary matrice by one or more channel matrix between described via node and the described multiple source node except desirable source node is carried out singular value decomposition, and estimate second unitary matrice by one or more channel matrix between described via node and a plurality of destinations node except the node of described destination is carried out singular value decomposition;

Will be at described via node place send to described destination node by the repeating signal that received signal be multiply by described first and second unitary matrice and generate; And

At node place, described destination, from received repeating signal, detect the transmission signal that sends from desirable source node.

6, a kind of wireless communication system is used for sending the transmission signal from the desirable source node of multiple source node to target destination node by via node, and it comprises:

The Matrix Estimation unit, it is configured to estimate the Moore-Penrose inverse matrix that derives according to a plurality of channel matrixes between described via node and a plurality of node;

The repeating signal generation unit, it is configured to generate repeating signal by the weighting matrix and the unitary matrice that received signal be multiply by the described Moore-Penrose inverse matrix of definition, and this unitary matrice is to obtain by one or more channel matrix between described via node and the described a plurality of nodes except defining node is carried out singular value decomposition; And

Transmitting element, it is configured to described repeating signal is sent to described destination node;

Wherein, described destination node detects described transmission signal from received repeating signal.

7, a kind of communication node, the transmission signal relay that is used for the desirable source node from the multiple source node is sent is to the destination node, and it comprises:

The Matrix Estimation unit, it is configured to estimate the Moore-Penrose inverse matrix that derives according to a plurality of channel matrixes between described via node and a plurality of node;

The repeating signal generation unit, it is configured to generate repeating signal by the weighting matrix and first unitary matrice that received signal be multiply by the described Moore-Penrose inverse matrix of definition, and this first unitary matrice is to obtain by one or more channel matrix between described via node and the described a plurality of nodes except defining node is carried out singular value decomposition; And

Transmitting element, it is configured to described repeating signal is sent to described destination node.

8, communication node according to claim 7 also comprises:

The transformation matrix estimation unit, it is configured to estimate transformation matrix, and this transformation matrix is made of the product of following matrix and one or more unitary matrice, and in this matrix, if i+j does not satisfy setting, then the matrix element of the capable j row of i is zero;

Wherein said transmitting element will send to described destination node by received signal being multiply by the repeating signal that described transformation matrix and described first unitary matrice generate.

9, communication node according to claim 7 also comprises:

The transformation matrix estimation unit, it is configured to estimate transformation matrix, this transformation matrix is made of the product of an a pair of angular moment battle array and a unitary matrice, this unitary matrice be according between described source node and the described via node or the channel matrix between described via node and the described destination node derive;

Wherein said transmitting element will send to described destination node by received signal being multiply by the repeating signal that described transformation matrix and described first unitary matrice generate.

10, a kind of communication means is used for the transmission signal that the desirable source node from the multiple source node sends is relayed to the destination node by via node, and this method may further comprise the steps:

At described via node place, estimate the Moore-Penrose inverse matrix that derives according to a plurality of channel matrixes between described via node and a plurality of node;

Generate repeating signal by weighting matrix and the unitary matrice that received signal be multiply by the described Moore-Penrose inverse matrix of definition, this unitary matrice is to obtain by one or more channel matrix between described via node and the described a plurality of nodes except defining node is carried out singular value decomposition;

Described repeating signal is sent to described destination node; And

At node place, described destination, from received repeating signal, detect described transmission signal.

11, a kind of communication node, the transmission signal relay that is used for the desirable source node from the multiple source node is sent is to the destination node, and it comprises:

The Matrix Estimation unit, it is configured to estimate the Moore-Penrose inverse matrix that derives according to a plurality of channel matrixes between described via node and a plurality of node;

The first unitary matrice estimation unit, it is configured to estimate first unitary matrice by one or more channel matrix between described via node and the described multiple source node except desirable source node is carried out singular value decomposition;

The second unitary matrice estimation unit, it is configured to estimate second unitary matrice by one or more channel matrix between described via node and the described a plurality of destinations node except the node of described destination is carried out singular value decomposition; And

The repeating signal generation unit, it is configured to generate repeating signal by two in the weighting matrix, described first unitary matrice and described second unitary matrice that received signal be multiply by the described Moore-Penrose inverse matrix of definition; And

Transmitting element, it is configured to described repeating signal is sent to described destination node.

12, communication node according to claim 11 is wherein selected described two matrixes in the described matrix based on the quality of channel status.

13, a kind of communication means is used for the transmission signal that the desirable source node from the multiple source node sends is relayed to the destination node by via node, may further comprise the steps:

At described via node place, estimate the Moore-Penrose inverse matrix that derives according to a plurality of channel matrixes between described via node and a plurality of node;

At described via node place, estimate first unitary matrice by the one or more channel matrixes between described via node and the described multiple source node except desirable source node are carried out singular value decomposition, and estimate second unitary matrice by the one or more channel matrixes between described via node and the described a plurality of destinations node except the node of described destination are carried out singular value decomposition;

Generate repeating signal by two in the weighting matrix, described first unitary matrice and described second unitary matrice that received signal be multiply by the described Moore-Penrose inverse matrix of definition; And

Described repeating signal is sent to described destination node.

Images