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

Communication system and method using relay node

Technical Field

The present invention relates generally to wireless communication, and more particularly, to a communication node and a communication method using a multi-hop scheme and a Multiple Input Multiple Output (MIMO) scheme.

Background

In recent years, a system based on a combination of a multi-hop scheme and a MIMO (or multi-antenna) scheme (the system is referred to as a MIMO multi-hop system) continues to be spotlighted. In a multi-hop scheme, a signal is transmitted from a source node to a destination node (or target node) through one or more relay nodes located between the source and destination. This system has the advantage of extending the coverage area (theoretically, an unlimited signal transmission area) by relaying signals and the fast establishment of a wireless network. With the MIMO system, signals are transmitted and received using multiple transmit antennas and multiple receive antennas to improve communication capacity through efficient use of space.

The signal transmission in the MIMO multi-hop system is performed in the following steps. First, a signal S transmitted from a source node is received at a relay node. The received signal Y at the relay node is represented as:

Y＝HS+n (1)

wherein H denotes a channel matrix between a source and a relay node, S denotes a transmission signal vector, and n denotesNoise. Then, the transmission signal S is detected by a Zero Forcing (ZF) method. The method is to calculate a pseudo-inverse matrix W₁＝(H^HH)-¹H^HAnd multiplying the received signal by a pseudo-inverse matrix W₁And normalizing the coefficient to detect the transmission signal S. The process is represented as:

W₁Y＝S+W₁n (2)

pseudo-inverse matrix W₁The superscript H in (a) denotes the conjugate transpose.

The Norm (Norm) of an arbitrary matrix a can be defined as:

‖A‖＝(Tr(E[AA^H]))^1/2 (3)

where the symbol | represents the norm, the symbol Tr (·) represents the sum (i.e., the trace) of the diagonal elements of the matrix in parentheses, and the symbol E [ ·]The values in the brackets are shown averaged. Specifically, vector V is set to (V)₁，v₂，...，v_M)^TThe norm of (V) is represented as:

‖V‖＝[|v₁|²+|v₂|²+…+|v_M|²]^1/2 (3)′

where the superscript T denotes transpose. The pseudo-inverse matrix corresponds to the Moore-Penrose inverse matrix. In general, the Moore-Penrose inverse matrix B is defined as an m × n matrix, which holds for the n × m matrix a, BA ═ I. In the example shown, for the matrix H, W₁H ═ I holds.

Then, a pseudo-inverse matrix W is calculated₂＝(G^HG)^-1G^HWhere G denotes a channel matrix between the relay node and the destination node. Multiplying both sides of equation (2) by the pseudo-inverse W matrix simultaneously₂And a normalization coefficient E. This relationship is expressed as:

E(W₂W₁)Y＝EW₂(S+W₁n) (4)

wherein, <math> <mrow> <mi>E</mi> <mo>=</mo> <mn>1</mn> <mo>/</mo> <mrow> <mo>(</mo> <mo>|</mo> <mo>|</mo> <msub> <mi>W</mi> <mn>1</mn> </msub> <mo>|</mo> <mo>|</mo> <mo>|</mo> <mo>|</mo> <msub> <mi>W</mi> <mn>2</mn> </msub> <mo>|</mo> <mo>|</mo> <mo>)</mo> </mrow> <mo>*</mo> <msup> <mrow> <mo>(</mo> <msub> <mi>P</mi> <mi>s</mi> </msub> <mo>/</mo> <mrow> <mo>(</mo> <msub> <mi>P</mi> <mi>s</mi> </msub> <mo>+</mo> <msubsup> <mi>&sigma;</mi> <mi>n</mi> <mn>2</mn> </msubsup> <mo>)</mo> </mrow> <mo>)</mo> </mrow> <mrow> <mn>1</mn> <mo>/</mo> <mn>2</mn> </mrow> </msup> </mrow> </math> where true, Ps denotes transmission power, and σ²Is the noise variance.

The signal thus calculated is transmitted from the relay node to the destination node. Signal Y received at destination node_RExpressed as:

Y_R＝GEW₂W₁Y+n_R (5)

wherein n is_RRepresenting the noise component. Can be based on W₁And W₂The definition of (5) is rewritten as:

Y_R＝E(S+W₁n)+n_R (6)

in this way, the transmission signal S can be instantly obtained at the destination node. Such MIMO multihop systems are described, for example, in the following documents, "CapacityScalling Laws in MIMO Wireless networks", Allerton Conference communication, Control, and Computing, Monticello, IL., pp.378-389, Oct.2003.

From equation (6), it will be appreciated that the received signal Y_RComprising a factor 1/(| W) related to the transmitted signal S₁‖‖W₂|). The factor | W₁II and II W₂Is indispensable for transmission power control performed at the relay node. However, since W₁And W₂Which are the inverse of the channel matrices H and G, respectively, which are affected by the noise amplitude, the signal quality inevitably degrades. In addition, equation (6) includes a noise component "n" introduced during propagation from the source to the relay node, thereby seriously affecting the received signal. Therefore, as the number of hops increases, signal degradation due to noise will become significant.

Furthermore, a wireless communication system in which signals are simultaneously relayed from a plurality of source nodes to the relevant destination nodes via relay nodes must be considered. In such a system, the signal received at the destination node includes not only the effects of the desired source node, but also the effects of other source nodes. In such a system, there are the following concerns: the noise is amplified at the relay node and the received signal quality at the destination node is severely degraded.

Disclosure of Invention

The present invention is intended to overcome the above-described problems, and an object of the present invention is to provide a communication system, a communication node, and a communication method which can more effectively prevent a reduction in the quality of a received signal at a destination node in signal transmission from a source node to the destination, as compared with the conventional art.

In one aspect of the present invention, a communication node for relaying a transmission signal transmitted from a desired source node to a target destination node between a plurality of source nodes and a plurality of destination nodes is provided. The communication node includes:

(a) a first unitary matrix estimation unit configured to estimate a first unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and a plurality of source nodes other than the desired source node;

(b) a second unitary matrix estimation unit configured to estimate a second unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and a plurality of destination nodes other than the target destination node; and

(c) a transmitting unit configured to transmit a relay signal generated by multiplying the received signal by the first and second unitary matrices to a target destination node.

In a communication system using such a relay node, a destination node detects a transmission signal from a received relay signal.

In another aspect of the present invention, there is provided a communication node for relaying a transmission signal transmitted from a desired source node among a plurality of source nodes to a destination node. The communication node includes:

(a) a matrix estimation unit configured to estimate a Moore-Penrose inverse matrix derived from a plurality of channel matrices between the relay node and the plurality of nodes;

(b) a first unitary matrix estimation unit configured to estimate a first unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and a plurality of source nodes other than the desired source node;

(c) a second unitary matrix estimation unit configured to estimate a second unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and a plurality of destination nodes other than the destination node;

(d) a relay signal generation unit configured to generate a relay signal by multiplying the received signal by two of a weighting matrix defining a Moore-Penrose inverse matrix, a first unitary matrix, and a second unitary matrix; and

(e) a transmitting unit configured to transmit the relay signal to the destination node.

With either type of communication node, when a signal is transmitted from a source node to a destination node using a multi-hop MIMO scheme, it is possible to prevent the quality of the signal received at the destination node from degrading.

Drawings

Other objects, features and advantages of the present invention will become more apparent from the following detailed description when read in conjunction with the accompanying drawings. In the drawings:

fig. 1 is a schematic diagram showing a communication system employing a MIMO scheme and a multi-hop scheme;

fig. 2 is a schematic block diagram of a relay node;

fig. 3 is a functional block diagram of a relay signal generator according to a first embodiment of the present invention;

fig. 4 is a flowchart showing the operation of the communication system according to the first embodiment of the present invention;

fig. 5 is a functional block diagram of a relay signal generator according to a second embodiment of the present invention;

fig. 6 is a flowchart showing an operation of a communication system using the relay signal generator shown in fig. 5;

FIGS. 7A and 7B are graphs showing simulation results of the present invention according to a third embodiment of the present invention;

fig. 8 is a schematic diagram showing a communication system in which a plurality of nodes transmit and receive signals through a relay node;

fig. 9 is a functional block diagram of a conventional relay node;

fig. 10 is a functional block diagram of a relay node according to a fourth embodiment of the present invention;

fig. 11 shows an example of arithmetic operations performed at a relay node;

fig. 12 shows another example of arithmetic operations performed at a relay node;

fig. 13 shows another example of arithmetic operations performed at a relay node;

fig. 14 shows another example of arithmetic operations performed at a relay node;

fig. 15 is a graph showing simulation results of the present invention compared to the prior art.

Detailed Description

The present invention will be described in detail below with reference to the accompanying drawings. In the specification and claims, the "unitary matrix" does not have to be a normal matrix (normal matrix), and thus the number of rows and columns may be different from each other. A "unitary matrix" is a matrix in which rows (or columns) are orthogonal to each other. Therefore, a normal matrix that diagonalizes the square matrix a is also included, and the "unitary matrix" includes an N × M non-square matrix for diagonalizing the M × N non-square matrix B.

In a preferred embodiment, the first unitary matrix is determined by decomposing a first channel matrix between the source node and the relay node into a product comprising a first triangular matrix, and the second unitary matrix is determined by decomposing a second channel matrix between the relay node and the destination node into a product comprising a second triangular matrix. If i + j does not satisfy the prescribed value, the matrix element in the ith row and the jth column is 0.

The communication nodes used in the embodiments include: first means for decomposing a channel matrix H between a source node and a relay node into a product comprising a first triangular matrix E; second means for decomposing a channel matrix G between the relay node and the destination node into a product comprising a second triangular matrix P; transformation matrix generating means for generating a transformation matrix a based on the first and second triangular matrices; multiplying means for multiplying the received signal with a first unitary matrix, a transform matrix and a second unitary matrix to generate a relayed signal; and transmitting means for transmitting the relay signal to the destination node. If i + j does not satisfy the prescribed value, the i-th row and j-th column element of the transformation matrix A is 0.

Since the relay signal is generated using the unitary matrix and the transform matrix, multi-hop communication can be achieved while reducing signal loss and signal quality degradation.

In an example, a transform matrix is estimated based on a first unitary matrix, a commutative matrix, and a conjugate transpose of a second unitary matrix. With this arrangement, the destination node can combine the relay signals from the plurality of relay nodes in phase. Since the signal combining coefficients do not contain imaginary components (phase components), some components do not need to be deleted during signal combining, and therefore, the relay signals can be coherently combined in phase at the destination node.

In a preferred example, information relating to the rate and power level of the transmitted signal is fed back from the destination node to the source node using a feedback channel from the destination node to the source node through the relay node. The information is obtained at the destination node based on the channel estimate.

In a preferred example, there is provided a method of relaying a transmission signal transmitted from a source node to a destination node through a relay node. In the method, at the relay node, a first channel matrix between the source node and the relay node is decomposed into a product comprising a first triangular matrix, and a second channel matrix between the relay node and the destination node is decomposed into a product comprising a second triangular matrix. Then, a transformation matrix is generated based on the first and second triangular matrices, wherein if i + j does not satisfy a prescribed value, an i-th row and a j-th column element of the transformation matrix is 0. Then, the signal received at the relay node is multiplied by a first unitary matrix, a transformation matrix and a second unitary matrix. The multiplied signal is then transmitted from the relay node to the destination node.

Preferably, a third triangular matrix is generated at the destination node based on the first and second triangular matrices and the transformation matrix. The destination node then detects a transmission signal from the received signal using the third triangular matrix.

In another example, at the relay node, a first channel matrix between the source node and the relay node is decomposed into a product comprising a first triangular matrix, and a second channel matrix between the relay node and the destination node is decomposed into a product comprising a second triangular matrix. Then, the signal received at the relay node is multiplied by a unitary matrix. Then. A transmission signal transmitted from a source node is detected from received signals using a first triangular matrix. Then. The detected transmission signal is multiplied by a transform matrix and a second unitary matrix. The resulting signal is transmitted from the relay node to the destination node.

In this example, a transmission signal is detected from the resulting signal at the destination node using a second triangular matrix.

The method is beneficial to effectively preventing the noise accumulation of each hop at each relay node. Since the destination node does not need to perform the unitary transform, the workload of signal processing at the destination node can be reduced.

In another example, a communication node relays a signal transmitted from a particular source node to a target destination node in an environment where wireless communication is conducted between multiple source nodes and the destination node. The communication node estimates a first unitary matrix based on one or more channel matrices between the relay node and one or more source nodes other than the desired source node, generates a relay signal by multiplying the received signal by the first unitary matrix, multiplying by a second unitary matrix, and transmits the relay signal to the destination node. The first unitary matrix includes a matrix obtained by performing singular value decomposition on one or more channel matrices between the relay node and source nodes other than the desired source node.

The transmitted signal from the desired source node may be separated from the signal components from other source nodes by multiplying the received signal by a first unitary matrix. In other words, interference from other source nodes can be removed, but interference from a desired source node cannot be removed. Conversely, since the multiplication of the received signal by the first unitary matrix does not cause the noise component to be amplified, the noise component in the received signal is kept low without being amplified.

The second unitary matrix includes a matrix obtained by performing singular value decomposition on one or more channel matrices between the relay node and destination nodes other than the target destination node. Multiplying the signal with the second unitary matrix enables the destination node to separate the desired source node's transmitted signal from the signal components of other source nodes.

In another example, the communication node further estimates a transform matrix that is a product of a matrix in which a matrix element at an ith row and a jth column is 0 if a sum (i + j) of a row number and a column number is not a prescribed value and one or more unitary matrices. In this case, the communication node transmits a relay signal generated by multiplying the received signal by the first unitary matrix, the transform matrix, and the second unitary matrix to the destination node.

In another example, the communication node estimates a transform matrix that is a product of a diagonal matrix and a unitary matrix, in which case the communication node transmits a relay signal generated by multiplying the received signal by a first unitary matrix, the transform matrix, and a second unitary matrix to the destination node.

This arrangement has the advantage that the computational effort for separating the transmitted signal from the desired source node at the destination node can be reduced.

The communication node may also estimate the weighting matrix based on a plurality of channel matrices between the relay node and a plurality of source nodes including the desired source node. In this case, the communication node transmits a relay signal generated by multiplying the received signal by the weighting matrix and the unitary matrix to the target destination node. The unitary matrix includes a matrix obtained by performing singular value decomposition on one or more channel matrices between the relay node and destination nodes other than the target destination node.

In another example, the communication node generates the relayed signal by multiplying the received signal by two of a first unitary matrix, a second unitary matrix, and a weighting matrix comprising a Moore-Penrose inverse matrix. The two matrices are selected based on the quality of the channel state. This arrangement enables the relay node to select an appropriate relay scheme according to the channel status, and can improve the quality of the received signal at the destination node.

(example 1)

Fig. 1 is a diagram showing the overall configuration of a communication system according to an embodiment of the present invention. The communication system employs a multi-hop scheme and a Multiple Input Multiple Output (MIMO) scheme. The communication system includes a source node 12, a destination node 16, and K (K ≧ 1) relay nodes 14-1 through 14-K. The kth relay node is denoted as 14-K (1. ltoreq. K. ltoreq. K). Communication between the source node 12 and the relay node 14-k and communication between the relay node 14-k and the destination node 16 are performed using a MIMO scheme. The signal transmission from the source node 12 to the destination node 16 is performed by a multi-hop scheme. For simplicity, in this embodiment, each of the K relay nodes may relay a signal from the source node 12 to the destination node 16 over one hop. However, the number of hops may be increased.

The source node 12 transmits mutually distinguishable signals from multiple antennas (e.g., M antennas). Each of the M antennas independently transmits an associated signal under the MIMO scheme. The signals transmitted from the M antennas define a transmitted signal vector S, each signal being a vector component.

Each of the K relay nodes 14 receives a signal from the source node 12, performs predetermined signal processing on the received signal to generate a relay signal, and transmits the relay signal to the destination node 16. The K relay nodes 14 have the same structure and function, which will be described below.

The destination node 16 receives the relay signals from the K relay nodes 14 and detects the contents of the transmission signal vector S transmitted from the source node 12.

Fig. 2 is a block diagram of a relay node 14-k. The relay node 14-k has a plurality of antennas 22-1 to 22-M, a receiving unit 24, a channel estimator 25, a relay signal generator 26, and a transmitting unit 28. Since the source node 12 and the destination node 16 may also be relay nodes, this structure may be applied not only to the relay node 14 but also to the source node 12 and the destination node 16.

In this embodiment, for purposes of simplicity, it is assumed that each of the source node 12, the relay nodes 14-1 through 14-K, and the destination node 16 has M antennas for transmitting and receiving signals. However, the nodes may have different numbers of antennas, and in addition, different numbers of antennas may be used in the transmission and reception of signals.

The receiving unit 24 pairs the signals Y received at the M antennas 22-1 to 22-M_kAppropriate signal processing is performed. Such signal processing includes receive front-end processing (e.g., frequency conversion and bandwidth limiting) and weighting of individual antennas. Received signal Y_kRepresented as a vector consisting of M components corresponding to M antennas. The receiving unit 24 also analyses the received signal Y_kTo determine a destination node to which to send a signal. If the signal does not reach the destination node over one hop, the relay node 14-k sends the signal to another relay node.

The channel estimator 25 estimates the channel matrix H between the source node 12 and the relay node 14-k_k. By receiving each pilot channel transmitted from the source node 12, a channel matrix H may be obtained_kOf the matrix element(s). Similarly, the channel estimator 25 estimates the channel matrix G between the relay node 14-k and the destination node 16_k. The channel estimator 25 also estimates the channel state, if necessary. The state of the wireless channel may be estimated, for example, by measuring SNR or SIR from the received signal. The level of channel state may be used in the following embodiments.

The relay signal generator 26 generates a relay signal based on the received signal Y_kAnd generating a relay signal X by the channel estimation result_k. Relay signal X_kIs a vector consisting of M components corresponding to M antennas. The relay signal will be described in detail belowA generator 26.

The transmission unit 28 performs signal processing to relay the signal X through a plurality of antennas_kTo destination node 16. The signal processing includes frequency conversion, bandwidth limiting, power amplification and weighting of the individual antennas.

Fig. 3 is a functional block diagram of the relay signal generator 26. The relay signal generator 26 has a QR decomposition unit 32, a weighting factor calculation unit 34, and a weighting unit 36.

When receiving the channel matrix H from the channel estimator 25_kAnd G_kWhen the information is related, the QR decomposition unit 32 converts the channel matrix H_kDecomposition into unitary matrices Q_kAnd a triangular matrix R_kIn the form of the product of (c). As a result, the unitary matrix Q satisfying equation (7) is determined_kAnd a triangular matrix R_k。

H_k＝Q_kR_k (7)

It should be noted that the triangular matrix R_kThe first to (i-1) th column elements of the ith row in (b) are 0 (2. ltoreq. i.ltoreq.M), which is expressed by equation (8) as follows:

QR decomposition unit 32 also converts channel matrix G_kDecomposition into a triangular matrix P represented by equation (9)_k ^HAnd unitary matrix O_k ^HIn which superscript H denotes the conjugate transpose.

G_k＝P_k ^HO_k ^H (9)

It should be noted that the triangular matrix P_kThe first to (i-1) th column elements of the ith row in (b) are 0 (2. ltoreq. i.ltoreq.M), which is expressed by equation (10) as follows:

due to the matrix P_kIs an upper triangular matrix, so P_k ^HIs a lower triangular matrix.

According to the channel matrix H_kAnd G_kAnd QR decomposition type, weighting factor calculating section 34 calculates received signal Y_kThe weighting factor of (2). Details of the calculation of the weighting factors will be described below in connection with the operation of the communication system.

The weighting unit 36 performs a predetermined matrix operation to convert the received signal Y into a digital signal_kConversion to a Relay Signal X_k。

Fig. 4 is a flow chart illustrating operation of a communication system in accordance with an embodiment of the present invention. In this communication system, a source node 12 transmits a transmission signal vector S composed of a set of M signal components from M antennas to surrounding relay nodes. Relay nodes located within a predetermined range receive the signal S from the source node 12. This range may be referred to as a 1 hop range. For convenience of explanation, it is assumed that K relay nodes receive the transmission signal S and perform similar signal processing to relay the signal to the destination node. Although only the kth relay node (1 ≦ K ≦ K) is shown in FIG. 4, other relay nodes perform similar operations.

First, the source node 12 and the destination node 16 transmit a pilot signal L, respectively_kAnd Z_kThese pilot signals are received at the relay node 14-k. In step 401, the relay node 14-k bases on the pilot signal L_kAnd Z_kChannel estimation is performed to estimate a channel matrix H between the source node 12 and the relay node 14-k, and a channel matrix G between the relay node 14-k and the destination node 16.

In step 402, the source node 12 transmits transmission signals represented as a signal vector S composed of a set of M components from M antennas to surrounding relay nodes.

At step 404, the relay node 14-k receives a signal from the source node 12. The received signal is represented as:

Y_k＝H_kS+n_k (11)

wherein H_kIs the channel matrix between the source node 12 and the kth relay node, n is as described above_kRepresenting the noise component.

At step 406, the relay node 14-k pairs the channel matrix H at the QR decomposition unit 32_kAnd G_kQR decomposition is performed (see fig. 3). In this step, the channel matrix H_kIs decomposed into a unitary matrix Q_KAnd a triangular matrix R_KProduct of (H)_k＝Q_kR_k) Of the form of a channel matrix G_kIs decomposed into a triangular matrix P_k ^HAnd unitary matrix O_k ^HProduct of (G)_k＝P_k ^HO_k ^H) In the form of (1).

At step 408, the weighting factor calculation unit 34 calculates the weighting factor according to the triangular matrix P_kAnd R_kCalculating a transformation matrix A_k(FIG. 3). If i + j is not equal to M +1(i + j ≠ M +1), then at transformation matrix A_kIs 0 in the ith row and jth column of (a). In this case, the matrix A is transformed_kRepresented by equation (12).

In other words, when these rows and columns are arranged in reverse order (inverse diagonal matrix), the transformation matrix a_kIs a matrix that becomes a diagonal matrix. If i + j equals M +1, the matrix element ${\left(A_k\right)}_{i,M-\mathrm{<figure-callout\; id="1"\; label="i\; +"\; filenames="A20051009384000371.png,A20051009384000421.png"\; state="\{\{state\}\}">i</figure-callout>}+1}=a_i^k$ Expressed as:

<math> <mrow> <msubsup> <mi>a</mi> <mi>i</mi> <mi>k</mi> </msubsup> <mo>=</mo> <mfrac> <msubsup> <mrow> <mo>(</mo> <msup> <msub> <mi>P</mi> <mi>k</mi> </msub> <mi>H</mi> </msup> <mi>&Pi;</mi> <msub> <mi>R</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mrow> <mi>i</mi> <mo>,</mo> <mi>M</mi> <mo>-</mo> <mi>i</mi> <mo>+</mo> <mn>1</mn> </mrow> <mi>H</mi> </msubsup> <mrow> <mo>|</mo> <mo>|</mo> <msubsup> <mrow> <mo>(</mo> <msup> <msub> <mi>P</mi> <mi>k</mi> </msub> <mi>H</mi> </msup> <mi>&Pi;</mi> <msub> <mi>R</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mrow> <mi>i</mi> <mo>,</mo> <mi>M</mi> <mo>-</mo> <mi>i</mi> <mo>+</mo> <mn>1</mn> </mrow> <mi>H</mi> </msubsup> <mo>|</mo> <mo>|</mo> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>13</mn> <mo>)</mo> </mrow> </mrow> </math>

where matrix n represents a commutative matrix, which is represented by equation (14):

in step 410, a relay signal X is generated_kThe relay signal is represented by equation (15):

X_k＝E_kO_kA_kQ_k ^HY_k (15)

coefficient E_kIs a scalar defined by equation (16):

<math> <mrow> <msub> <mi>E</mi> <mi>k</mi> </msub> <mo>=</mo> <msqrt> <mfrac> <mi>PM</mi> <mrow> <mi>P</mi> <mo>[</mo> <mi>tr</mi> <mo>{</mo> <mrow> <mo>(</mo> <msubsup> <mi>P</mi> <mi>k</mi> <mi>H</mi> </msubsup> <msub> <mi>A</mi> <mi>k</mi> </msub> <msub> <mi>R</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <msup> <mrow> <mo>(</mo> <msubsup> <mi>P</mi> <mi>k</mi> <mi>H</mi> </msubsup> <msub> <mi>A</mi> <mi>k</mi> </msub> <msub> <mi>R</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mi>H</mi> </msup> <mo>}</mo> <mo>]</mo> <mo>+</mo> <mi>MN</mi> <msup> <mi>&sigma;</mi> <mn>2</mn> </msup> </mrow> </mfrac> </msqrt> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>16</mn> <mo>)</mo> </mrow> </mrow> </math>

where P represents the total transmit power at the source node 12 and σ²Representing the noise level.

In step 412, relay signal X is transmitted_kTo destination node 16.

At step 414, signals from all relay nodes that relay signals from the source node 12 are received at the destination node 16. Signal Y to be received at destination node 16_RExpressed as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>Y</mi> <mi>R</mi> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>G</mi> <mi>k</mi> </msub> <msub> <mi>X</mi> <mi>k</mi> </msub> <msub> <mrow> <mo>+</mo> <mi>n</mi> </mrow> <mi>R</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msub> <mi>T</mi> <mi>k</mi> </msub> <mi>S</mi> <mo>+</mo> <mi>n</mi> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>17</mn> <mo>)</mo> </mrow> </mrow> </math>

wherein n is_RAnd n represents a noise component. From equations (7), (9), and (11), the following relationship holds:

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, from the above relationship and equations (9) and (16), the following relationship holds:

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)

Wherein, T_k＝P_k ^HA_kR_k。

The matrix T can be expressed according to equations (8), (10) and (16)_kExpressed as equation (18):

considering equation (13), it should be understood that the non-zero matrix element a_i ^kIs equal to p_ii(r_{M-i+1 M-i+1})^*/|p_ii(r_{M-i+1 M-i+1})^*Where the asterisk indicates the complex conjugate.Thus, Y_kS becomes a matrix having first to mth elements represented by equation (19).

In step 416, the transmission signal S is detected according to equations (17) and (18). Using successive interference cancellation methods (for successively cancelling T_kOff-diagonal components) to perform signal detection. Assuming that the successive cancellation method is performed in an ideal manner, the equivalent signal-to-noise ratio (λ m) of each transmission stream is calculated using equation (20-1) from the channel estimation result at the destination node 16.

<math> <mrow> <msub> <mi>&lambda;</mi> <mi>m</mi> </msub> <mo>=</mo> <mfrac> <mi>P</mi> <mi>M</mi> </mfrac> <mfrac> <msup> <mrow> <mo>(</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>k</mi> </munderover> <msub> <mrow> <mo>(</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <msup> <msub> <mi>P</mi> <mi>k</mi> </msub> <mi>H</mi> </msup> <msub> <mi>A</mi> <mi>k</mi> </msub> <msub> <mi>R</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mrow> <mi>m</mi> <mo>,</mo> <mi>M</mi> <mo>-</mo> <mi>m</mi> <mo>+</mo> <mn>1</mn> </mrow> </msub> <mo>)</mo> </mrow> <mn>2</mn> </msup> <mrow> <msubsup> <mi>&sigma;</mi> <mi>r</mi> <mn>2</mn> </msubsup> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>k</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msup> <mrow> <mo>|</mo> <mo>|</mo> <msub> <mrow> <mo>(</mo> <msup> <msub> <mi>P</mi> <mi>k</mi> </msub> <mi>H</mi> </msup> <msub> <mi>A</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mi>m</mi> </msub> <mo>|</mo> <mo>|</mo> </mrow> <mn>2</mn> </msup> <mo>+</mo> <msubsup> <mi>&sigma;</mi> <mi>d</mi> <mn>2</mn> </msubsup> </mrow> </mfrac> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>20</mn> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </math>

Wherein sigma_r ²And σ_d ²Respectively noise component n_kAnd n_RAnd P represents the total transmit power of the source node 12. According to equation (20-1), when the flow S is controlled independently₁，...，S_MThe communication capacity C between the source node 12 and the destination node 16 is represented by equation (20-2).

<math> <mrow> <mi>C</mi> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>m</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>M</mi> </munderover> <mfrac> <mn>1</mn> <mn>2</mn> </mfrac> <msub> <mi>log</mi> <mn>2</mn> </msub> <mrow> <mo>(</mo> <mn>1</mn> <mo>+</mo> <msub> <mi>&lambda;</mi> <mi>m</mi> </msub> <mo>)</mo> </mrow> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>20</mn> <mo>-</mo> <mn>2</mn> <mo>)</mo> </mrow> </mrow> </math>

Information about the rate of each flow may be reported to the source node 12 by feeding back information from the destination node to the source node 12. The power levels of the individual streams may also be controlled independently.

Eliminating T as shown in equation (19)_kAnd the signal component S of the signal vector to be obtained from the relay node 14₁To S_MEach multiplied by a positive real number. The matrix elements are combined at the destination node. Since the coefficients used in signal combining do not include imaginary components (phase components), there is little need to eliminate these components in the signal combining process, and therefore, in-phase signal combining can be achieved at the maximum ratio. In other words, the relay signals from the various relay nodes 14 may be combined phase coherently.

Since the scalar E is calculated mainly from the transform of the unitary matrix_kAnd other coefficients, so that noise increase can be reduced as compared with the conventional techniqueAdversely affecting. This arrangement is advantageous from the viewpoint of reducing signal loss. Accordingly, the degradation of the signal quality, which is a problem in the related art, can be solved.

(example 2)

Fig. 5 is a functional block diagram of the relay signal generator 26 used in the relay node 14 according to the second embodiment of the present invention. The relay signal generator 26 includes a QR decomposition unit 32, a weighting factor calculation unit 34, a first weighting unit 36, a signal detector 39, and a second weighting unit 62. In the second embodiment, destination node 16 may have the structure and function shown in fig. 5, or alternatively, it may have the structure and function shown in fig. 3.

When receiving the channel matrix H from the channel estimator 25_kAnd G_kWhen the information is concerned, the QR decomposition unit 32 converts the channel matrix H_kDecomposition into unitary matrices Q_kAnd a triangular matrix R_kProduct of (H)_k＝Q_kR_k) In the form of (1). QR decomposition unit 32 also converts channel matrix G_kDecomposed into triangular matrices P_k ^HAnd unitary matrix O_k ^HProduct of (G)_k＝P_k ^HO_k ^H) In the form of (1).

According to the channel matrix H_kAnd G_kAnd information relating to QR decomposition, weighting factor calculating section 34 calculates the weighting factor for received signal Y_kA weighting factor is calculated.

The first weighting unit 36 weights the received signal Y_kAnd a weighting factor Q estimated by the weighting factor calculation unit 34_k ^HTo extract the various components of the received signal.

The signal detector 39 detects the transmission signal S transmitted from the source node 12 based on the weighted reception signal output from the weighting unit 36 and the information on the triangular matrix_k＝(S_k1，...，S_kM)。

The second weighting unit 62 weights the detected transmission signal S_kAnd the weighting factor a calculated by the weighting factor calculation unit 34_kO_k ^HMultiply and output a relay signal A_kO_k ^HS_kThe respective components of (a).

Fig. 6 is a flowchart showing the operation of the communication system according to the second embodiment of the present invention.

First, the source node 12 and the destination node 16 transmit a pilot signal L, respectively_kAnd Z_kThe pilot signal is received at the relay node 14-k. In step 701, the relay node 14-k bases on the pilot signal L_kAnd Z_kChannel estimation is performed to estimate a channel matrix H between the source node 12 and the relay node 14-k, and a channel matrix G between the relay node 14-k and the destination node 16.

In step 702, the source node 12 transmits transmission signals represented as a signal vector S composed of a set of M components from M antennas to surrounding relay nodes.

At step 704, the relay node 14-k receives a signal from the source node 12. The received signal is represented as:

Y_k＝H_kS+n_k

at step 706, the channel matrix H is aligned_kAnd G_kQR decomposition is performed. Will channel matrix H_kDecomposition into unitary matrices Q_kAnd a triangular matrix R_kProduct of (H)_k＝Q_kR_k) Of the channel matrix G_kDecomposed into triangular matrices P_k ^HAnd unitary matrix O_k ^HProduct of (G)_k＝P_k ^HO_k ^H) In the form of (1).

At step 708, the signal Y is received by_kAnd unitary matrix Q^HThe multiplication is performed to perform the unitary transform. Receiving signal Z subjected to unitary transformation_kExpressed as:

Z_k＝Q_k ^HY_k

＝R_kS+Q_k ^Hn_k

due to the matrix R_kIs an upper triangular matrix, so if noise is ignored, the following relationship holds.

Z_k1＝r₁₁S₁+r₁₂S₂+…+r_1MS_M

Z_k2＝r₂₂S₁₂+…+r_2MS_M

…

Z_kM-1＝r_{M-1 M-1}S_M-1+r_{M-1 M}S_M

Z_kM＝r_MMS_M

In step 710, a transmission signal S is detected from a received signal subjected to unitary transformation. First, the Mth received signal component Z is focused_kMAccording to known Z_kMAnd r_MMDetecting a transmitted signal component S_M. Then focus on the (M-1) th received signal component Z_kM-1According to known r_{M-1 M-1}、r_MMAnd S_MDetecting a transmitted signal component S_M-1. In a similar manner, the transmission signal component is continuously detected.

In step 712, by detecting the transmission signal S_kAnd A_kO_k ^HMultiplying to perform a further transformation, wherein the matrix A_kIs a diagonal matrix represented as follows:

A_k＝diag(P_k ^H)

at step 714, the signal O is transformed_k ^HS_kAs a relayed signal to destination node 16.

At step 716, the signals relayed from all relevant relay nodes 14 are received at the destination node 16. Receiving signal Y_RExpressed as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>Y</mi> <mi>R</mi> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>G</mi> <mi>k</mi> </msub> <msub> <mi>A</mi> <mi>k</mi> </msub> <msup> <msub> <mi>O</mi> <mi>k</mi> </msub> <mi>H</mi> </msup> <mi>S</mi> <mo>+</mo> <mi>n</mi> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>P</mi> <mi>k</mi> </msub> <mi>diag</mi> <mrow> <mo>(</mo> <msup> <msub> <mi>P</mi> <mi>k</mi> </msub> <mi>H</mi> </msup> <mo>)</mo> </mrow> <mi>S</mi> <mo>+</mo> <mi>n</mi> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <mi>DS</mi> <mo>+</mo> <mi>n</mi> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>21</mn> <mo>)</mo> </mrow> </mrow> </math>

where n represents a noise component. Equation (21) makes use of the channel matrix G_kDecomposition to G_k＝P_kO_kThe fact of the form. Because of P_kIs a triangular matrix, so that K matrices P_kThe sum (or combination) of (a) is also a triangular matrix. The combined result is represented as a matrix D (with elements dij). The triangular matrix P may be determined by performing QR decomposition at the destination node 16_kAnd unitary matrix O_kRelevant information or, alternatively, such information may be collected from the various relay nodes 14. If the noise component is ignored, equation (21) is developed into the following form.

Y_R1＝d₁₁S₁+d₁₂S₂+…+d_1MS_M

Y_R2＝d₂₂S₂+…+d_2MS_M

…

Y_RM-1＝d_{M-1 M-1}S_M-1+d_{M-1 M}S_M

Y_RM＝d_MMS_M

At step 718, the transmitted signal S is detected at the relay node 14. First, attention is paid to the Mth received signal component Y_RMAccording to known Z_RMAnd d_MMDetecting a transmitted signal component S_M. Then, focus is given to the (M-1) th received signal component Y_RM-1According to known d_{M-1 M-1}、d_{M-1 M}And S_MTo detect the transmitted signal component S_M-1. In a similar manner, the transmission signal component is continuously detected.

In the second embodiment, destination node 16 does not have to perform a unitary transform in step 716 of fig. 6.

(example 3)

Fig. 7A and 7B are graphs showing simulation results of signal transmission according to an embodiment of the present invention. The horizontal axis represents power to noise ratio (PNR) and the vertical axis represents capacity. In fig. 7A, the number of transmission antennas and the number of reception antennas are four, respectively, and two relay nodes (K ═ 2) are located between the source node and the destination node and within one-hop communication range. The curve of theoretical limit represents the theoretical limit of capacity as a function of PNR, while the curve of the prior art represents the capacity when the signal is relayed using the zero forcing method. The curve of example 1 was obtained by carrying out the method of the first example. In fig. 7B, the number of transmission antennas and the number of reception antennas are four, respectively, and four relay nodes (K ═ 4) are located between the source node and the destination node and within one-hop communication range. From the graphs of fig. 7A and 7B, it can be understood that the system capacity increases when the transmission power increases, and the method of embodiment 1 is superior to the conventional method in terms of achieving sufficient capacity.

(example 4)

In a fourth embodiment, transmission signals are relayed between a plurality of source nodes and a plurality of destination nodes by one or more relay nodes.

Fig. 8 is a schematic diagram of a wireless communication system according to a fourth embodiment of the present invention. The system comprises: l source nodes (802-1 to 802-L), each source node having M antennas; k relay nodes (804-1 to 804-K), each relay node having N antennas; and L destination nodes (806-1 to 806-L), each destination node having M antennas. Integers N, M and L satisfy the relationship N ≧ L × M. In this example, for simplicity, all source and destination nodes have M antennas, while all relay nodes have N antennas. Of course, the nodes may have different numbers of antennas, as long as the number of antennas of the source node is equal to or less than the number of antennas of the destination node.

As described above in connection with FIG. 1, the channel conditions between a source node 802-l having M antennas and a relay node 804-k having N antennas are represented by an NxM channel matrix H_l，kAnd (4) showing. Similarly, the channel state between relay node 804-k and destination node 806-l with M antennas is represented by an M N channel matrix G_k，l(simplified representation as G)_kl) And (4) showing.

The transmission signals from the plurality of source nodes are received and relayed by the relay node. The destination node addressed by the signal transmitted from the source node receives signals from the plurality of relay nodes and recovers the signal from the source node. Therefore, the signal received at the destination node is affected by the influence (interference) of the signal transmitted from another source node in addition to the transmission signal from the desired source node. The destination node must detect the desired transmitted signal by removing interference.

Prior to describing the signal processing of the fourth embodiment, a description will be given of general signals of a conventional communication system (for example, described in the above-mentioned article by Rohit u.

Fig. 9 is a functional block diagram of one of the conventional relay nodes (kth relay node). The relay node has L reception filters 902-1 to 902-L, L transmission filters 904-1 to 904-L set corresponding to the number L of source nodes, and a signal combining unit 906.

Relaying the received signal Y at the node_kTo L receive filters 902-1 to 902-L. Because of the received signal Y_kSignals from L source nodes are included, so it is represented as:

<math> <mrow> <msub> <mi>y</mi> <mi>k</mi> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>H</mi> <mrow> <mi>l</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <mi>n</mi> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>30</mn> <mo>)</mo> </mrow> </mrow> </math>

(N × 1 matrix)

Wherein S is_lIs transmitted from the 1 st source node with M signal components (S)_l1，S_l2，…，S_lM) Is transmitted as a signal vector, and n_kRepresenting the noise components introduced between the kth relay node and the plurality of source nodes. The dimension of the received signal is N × 1.

1 st receive filter 902-1 receives signal y represented by M vector components_kMultiplying by a weighting matrix w^b _kl. The weighting matrix w^b _klIs an M × N matrix and satisfies the relationship:

<math> <mrow> <msup> <mrow> <mo>[</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mn>1</mn> </mrow> <mi>bT</mi> </msubsup> <mo>.</mo> <mo>.</mo> <mo>.</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>bT</mi> </msubsup> <mo>.</mo> <mo>.</mo> <mo>.</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>L</mi> </mrow> <mi>bT</mi> </msubsup> <mo>]</mo> </mrow> <mi>T</mi> </msup> <mo>=</mo> <msup> <mrow> <mo>(</mo> <msubsup> <mi>H</mi> <mi>k</mi> <mi>H</mi> </msubsup> <mo>&CenterDot;</mo> <msub> <mi>H</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>&CenterDot;</mo> <msubsup> <mi>H</mi> <mi>k</mi> <mi>H</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>31</mn> <mo>-</mo> <mn>1</mn> <mo>)</mo> </mrow> </mrow> </math>

(ML × N matrix)

The relational expression represents an ML × N matrix. H_kIs a matrix comprising a plurality of channel matrices and is defined as:

H_k＝[H_lk，…，H_Lk] (31-2)

as can be understood from equations (31-1) and (31-2), w^b _klAnd H_lkAre orthogonal to each other. Using this orthogonality, receive filter 902-1 passes received signal y as shown in equation (32)_kMultiplying by a weighting matrix w^b _klTo convert the received signal vector to y'_kl。

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <msup> <mi>y</mi> <mo>&prime;</mo> </msup> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <mo>=</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>b</mi> </msubsup> <msub> <mi>y</mi> <mi>k</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>b</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>32</mn> <mo>)</mo> </mrow> </mrow> </math>

(M × 1 matrix)

Then, transmission filter 904-l converts converted reception signal y'_klMultiplication by another weighting matrix w^f _kl. The weighting matrix w^f _klIs an N × M matrix and satisfies the relationship:

<math> <mrow> <mo>[</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mn>1</mn> </mrow> <mi>f</mi> </msubsup> <mo>.</mo> <mo>.</mo> <mo>.</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>f</mi> </msubsup> <mo>.</mo> <mo>.</mo> <mo>.</mo> <msubsup> <mi>W</mi> <mi>kL</mi> <mi>f</mi> </msubsup> <mo>]</mo> <mo>=</mo> <msup> <mrow> <msubsup> <mi>G</mi> <mi>k</mi> <mi>H</mi> </msubsup> <mrow> <mo>(</mo> <msub> <mi>G</mi> <mi>k</mi> </msub> <msubsup> <mrow> <mo>&CenterDot;</mo> <mi>G</mi> </mrow> <mi>k</mi> <mi>H</mi> </msubsup> <mo>)</mo> </mrow> </mrow> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>33</mn> <mo>)</mo> </mrow> </mrow> </math>

(NxML matrix)

The relational expression represents an N × ML matrix. G_kIs a matrix comprising a plurality of channel matrices and is defined as:

G_k＝[G_lk，…，G_Lk] (34)

signal w to be multiplied^f _kl*y’_klIs provided to a signal combining unit 906. The signal combining unit 906 combines the output signals from the transmission filters 904-1 to 904-L to generate a relay signal x_k. The relay signal x_kExpressed as:

<math> <mrow> <msub> <mi>x</mi> <mi>k</mi> </msub> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>f</mi> </msubsup> <mo>&CenterDot;</mo> <msubsup> <mi>y</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mo>&prime;</mo> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>35</mn> <mo>)</mo> </mrow> </mrow> </math>

(N × 1 matrix)

Wherein E is_kIs a scalar for normalizing the transmission power of the relay node. The relay signal x_kSent to the destination node.

Among the plurality of destination nodes, the l-th destination node 806-l receives signals from the K relay nodes, each of which reflects a transmission signal transmitted from the l-th source node and addressed to the l-th destination node. Thus, the signal r received at the l-th destination node₁Is represented as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>r</mi> <mn>1</mn> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>x</mi> <mn>1</mn> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <mrow> <mo>(</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>b</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mo>+</mo> <msub> <mi>z</mi> <mi>l</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>36</mn> <mo>)</mo> </mrow> </mrow> </math>

(M × 1 matrix)

Wherein z is_lRepresenting the noise components introduced between the plurality of relay nodes and the l-th destination node. Using channel matrix G_klAnd a weighting matrix W^f _klOrthogonal relationship therebetween, to estimate the received signal r defined in equation (36)₁。

In equation (36), the received signal r₁Is linearly dependent on the desired transmitted signal s_lOf the corresponding signal component. Therefore, it is possible to directly detect a desired transmission signal s from a reception signal_lWithout performing the complex signal separation typically performed in MIMO schemes.

However, in this way, by the weighting factor w^b _klAmplifying noise n_kTherefore, degradation of the received signal quality at the destination node is a concern. By applying a coefficient E_kSet smaller to reduce the contribution of the weighting matrix to noise amplification. However, because of the coefficient E_kAlso for the desired signal s_lSo as to follow the coefficient E_kThe desired signal component also becomes smaller. With the conventional technique, the signal detection accuracy at the destination node may be degraded.

Fig. 10 is a functional block diagram showing a relay node according to a fourth embodiment of the present invention. The relay node is one of the relay nodes shown in fig. 8 (kth relay node 804-k). Other relay nodes also have the same structure and function. The relay node 804-k has L receive filters 1002-1 to 1002-L, L receive filter estimators 1004-1 to 1004-L, L intermediate filters 1006-1 to 1006-L, L intermediate filter estimators 1008-1 to 1008-L, L transmit filters 1010-1 to 1010-L, L transmit filter estimators 1012-1 to 1012-L and a signal combining unit 1014.

Fig. 11 shows arithmetic operations performed in the ith reception filter estimator 1004-l, the ith intermediate filter estimator 1008-l, and the ith transmission filter estimator 1012-l.

As shown in fig. 10, a signal y to be received at the kth relay node_kTo L receive filters 1002-1 to 1002-L. Because of the received signal y_kContains signals from L source nodes, so it is represented by equation (30) above.

The L (1. ltoreq. L. ltoreq.L) reception filter 1002-L receives a signal y represented by M vector components_kMultiplication by a first unitary matrix U_kl. The first unitary matrix has dimensions of N rows and N-M (L-1) columns (N ≧ LM), and is estimated by a receive filter estimator 1004-L.

Of the L channel matrices between the (kth) relay node 804-k and the L source nodes of interest, the L-th receive filter estimator 1004-L considers the matrix H^(l) _kContains L-1 channel matrices, except for one between the 1 st source node and the relay node 804-k, which are represented as:

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

it should be noted that unlike equation (31-2), matrix H^(l) _kDoes not contain a channel matrix H_lk. Thus, H^(l) _kWith dimensions of N rows and M (L-1) columns. By pairing matrix H as shown in equation (38)^(l) _kPerforming singular value decomposition to obtain the first unitary matrix U_kl。

(NxM (L-1) matrix)

In equation (38), Λ^(l) _k，l，…，Λ^(l) _k，L-1Are M x M diagonal matrices and their diagonal components are H^(l) _kThe singular value of (a). Matrix [ U ]^(l) _k，l，…，U^(l) _k，L-1]Having dimensions of N rows and M (L-1) columns, and comprising a matrix H^(l) _kA basis vector of the defined signal space. Similarly, [ V ]^(l) _k，l，…，V^(l) _k，L-1]^TComprises a matrix H^(l) _kA basis vector of the defined signal space, and is represented by a square matrix of M (L-1). times.M (L-1). U shape_k，lIs a first unitary matrix having N rows and N-M (L-1) columns. The matrix corresponds to the basis vectors of the null space of the signal space.

As shown in equation (39), the 1 st receive filter 1002-1 will receive signal y_klMultiplication by a first unitary matrix U^H _klTo convert the received signal vector to y'_k，l。

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msubsup> <mi>y</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mo>&prime;</mo> </msubsup> <mo>=</mo> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>y</mi> <mi>k</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>H</mi> <mrow> <mi>l</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>39</mn> <mo>)</mo> </mrow> </mrow> </math>

Because of the first unitary matrix U_klAnd is formed by H^(l) _kBasis vector correspondence of null space of defined signal spaceTherefore, when the received signal is multiplied by the first unitary matrix, the transmission signal from the l source node can be separated from the transmission signals from other source nodes. It should be noted that, unlike equation (32), interference between signal components transmitted from the ith source node is not removed. In contrast, at this stage, the noise component n is prevented_kAmplification of (1).

L intermediate filter 1006-l converts the converted reception signal y'_klMultiplication by a transformation matrix phi_klThe transformation matrix phi_klGenerated by the intermediate filter estimator 1008-l. However, in this embodiment, the transformation matrix Φ_klIs an identity matrix and therefore, the intermediate filter 1006-l and the intermediate filter estimator 1008-l do not perform specific processing. Of course, as described below in another embodiment, the intermediate filter estimator 1008-l may generate a different matrix than the identity matrix.

L-th transmission filter 1010-1 converts reception signal y'_klMultiplying by a second unitary matrix A_kl. The second unitary matrix has dimensions of N rows and N-M (L-1) columns (N ≧ LM), and is generated by the transmit filter estimator 1012-L.

Of the L channel matrices between the (kth) relay node 804-k of interest and the L destination nodes, the L-th transmit filter estimator 1012-L considers matrix G^(l) _kIncludes L-1 channel matrices other than one between the ith destination node and the relay node 804-k, which are represented as:

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

it should be noted that unlike equation (34), matrix G is^(l) _kDoes not contain a channel matrix G_lk. Thus, G^(l) _kWith dimensions of N rows and M (L-1) columns. By pairing matrix G as shown in equation (40)^(l) _kPerforming singular value decomposition to obtain the second unitary matrix A_kl。

(NxM (L-1) matrix)

In equation (40), Ω^(l) _k，l，…，Ω^(l) _k，L-1Is an M × M diagonal matrix, and its diagonal elements are G^(l) _kThe singular value of (a). Matrix [ A ]^(l) _k，l，…，A^(l) _k，L-1]Having dimensions of N rows and M (L-1) columns, and comprising a matrix G^(l) _kA basis vector of the defined signal space. Similarly, [ B ]^(l) _k，l，…，B^(l) _k，L-1]^TComprises a matrix G^(l) _kA basis vector of the defined signal space, and is represented by a square matrix of M (L-1). times.M (L-1). A. the_k，lIs a second unitary matrix having dimensions of N rows and N-M (L-1) columns. The matrix corresponds to the basis vectors of the null space of the signal space.

L-th transmission filter 1010-1 converts signal y'_klMultiplying by a second unitary matrix A_kl. Multiplying the signal A_kly’_k，lIs provided to the signal combination unit 1014. The signal combining unit 1014 combines the signals output from the transmission filters 1010-1 to 1010-L to generate a relay signal x_k. The relay signal x_kQuilt watchShown as follows:

<math> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>x</mi> <mi>k</mi> </msub> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <mo>&CenterDot;</mo> <msub> <mi>y</mi> <mi>k</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>U</mi> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msubsup> <mi>U</mi> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> </mtd> </mtr> </mtable> </mfenced> </math>

(N × 1 matrix)

Wherein E_kIs a scalar used to normalize the transmit power of the relay node 804-k. The relay signal x_kSent to the destination node.

Among the plurality of destination nodes, an l-th destination node 806-l to which the transmission signal from the l-th source node 802-l is addressed receives K relay signals from K relay nodes. A signal r to be received at the l-th destination node 806-l_lExpressed as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>r</mi> <mn>1</mn> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>x</mi> <mi>k</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>A</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>H</mi> <mrow> <mi>l</mi> <mo>,</mo> <mi>k</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>A</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <msub> <mrow> <mo>+</mo> <mi>z</mi> </mrow> <mi>l</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>41</mn> <mo>)</mo> </mrow> </mrow> </math>

(M × 1 matrix)

Wherein z is_lIs a noise component introduced between the plurality of relay nodes and the l-th destination node. Equation (41) is estimated using the following facts: if l ≠ l', the channel matrix G_klAnd a second unitary matrix A_kl’Are orthogonal to each other. If l is l', then the result is G_klA_klThe matrix represented is a normal matrix other than the identity matrix.

As is clear from equation (41), in the received signal r_lWill be derived fromOf the source node s_lWith transmitted signals s from other source nodes_l(l.noteq.l') isolation. In other words, interference between source nodes is substantially reduced; however, there is also interference remaining between multiple signal components within the transmitted signal from the desired source node. This is because, in general, G is represented by_klA_klU^H _klH_lkThe matrix represented is not a diagonal matrix. Therefore, the destination node must perform normal signal separation by performing in the MIMO scheme to detect the desired signal s from the received signal_l. The signal detection itself may become somewhat complex compared to conventional techniques.

However, this method has the advantage that noise n can be prevented at the relay node_kThe advantage of amplification of (a). In equation (41), G is associated with the noise n_kAmong those matrices multiplied, matrices that must be introduced. Because of the matrix A_klAnd U_klAre unitary matrices, so these matrices do not amplify noise. Therefore, it is not necessary to use the coefficient E as small as equation (36)_kIn equation (36), the noise is represented by the weighting matrix W^b _klAnd (4) amplifying. This means that the reduction in the accuracy of signal detection, which is a problem of concern in the conventional art, can be eliminated or reduced by the present embodiment.

(example 5)

Fig. 12 shows another example of arithmetic operations performed in the ith reception filter estimator 1004-l, the ith intermediate filter estimator 1008-l, and the ith transmission filter estimator 1012-l. The operations performed by the reception filter 1002-1, the reception filter estimator 1004-l, the transmission filter 1010-l, and the transmission filter estimator 1012-l are the same as in the fourth embodiment.

In the fifth embodiment, the l intermediate filter 1006-1 converts the signal y 'output from the reception filter 1004-1'_klMultiplication by a transformation matrix phi_klTo generate a signal phi_kly’_kl. The transformation matrix phi_klIs calculated by the ith intermediate filter estimator 1008-1.

The intermediate filter estimator 1008-1 pairs the matrix U as shown in equation (50)^H _klH_lkQR decomposition is performed.

U^H _klH_lk＝Q_1klR_1kl (50)

Wherein Q is_1klIs a unitary matrix with dimensions of N-M (L-1) rows and M columns, and R_1klIs an M x M upper right triangular matrix. Similarly, the intermediate filter estimator 1008-1 pairs the matrix (G) as shown in equation (51)^H _klA_lk)^HQR decomposition is performed.

(G^H _klA_lk)^H＝Q_2klR_2kl (51)

Wherein Q is_2klIs a unitary matrix with dimensions of N-M (L-1) rows and M columns, and R_2klIs an M x M upper right triangular matrix. The intermediate filter estimator 1008-1 also estimates the matrix Θ using a triangular matrix satisfying equations (50) and (51)_kl. The matrix theta_klIs represented as:

wherein matrix pi is defined as:

the intermediate filter estimator 1008-1 uses these estimation matrices to estimate the transformation matrix Φ_klIt is defined as:

Φ_kl＝Q_2klΘ_klQ^H _1kl (53)

the matrixΦ_klIs a matrix of (N-M (L-1)) × (N-M (L-1)).

The l-th intermediate filter 1006-1 converts the signal phi_kly’_klOutput to transmit filter 1010-1. The transmit filter 1010-l multiplies the input signal by the matrix A described in the fourth embodiment_klAnd outputs the multiplied signals to the signal combining unit 1014. The signal combining unit 1014 adds the signals from the L transmission filters 1010-1 to 1010-L and outputs a relay signal x_k. The relay signal x_kIs represented as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>X</mi> <mi>k</mi> </msub> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msub> <mi>&Phi;</mi> <mi>kl</mi> </msub> <msubsup> <mi>y</mi> <mi>kl</mi> <mo>&prime;</mo> </msubsup> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>Q</mi> <mrow> <mn>2</mn> <mi>kl</mi> </mrow> </msub> <msub> <mi>&Theta;</mi> <mi>kl</mi> </msub> <msubsup> <mi>Q</mi> <mrow> <mn>1</mn> <mi>kl</mi> </mrow> <mi>H</mi> </msubsup> <mo>)</mo> </mrow> <mrow> <mo>(</mo> <msubsup> <mi>U</mi> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>H</mi> <mi>lk</mi> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <msubsup> <mrow> <mo>+</mo> <mi>U</mi> </mrow> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msub> <mi>Q</mi> <mrow> <mn>2</mn> <mi>kl</mi> </mrow> </msub> <msub> <mi>&Theta;</mi> <mi>kl</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>R</mi> <mrow> <mn>1</mn> <mi>kl</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <msubsup> <mrow> <mo>+</mo> <mi>Q</mi> </mrow> <mrow> <mn>1</mn> <mi>kl</mi> </mrow> <mi>H</mi> </msubsup> <msubsup> <mi>U</mi> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msub> <mi>Q</mi> <mrow> <mn>2</mn> <mi>kl</mi> </mrow> </msub> <msub> <mi>&Theta;</mi> <mi>kl</mi> </msub> <msub> <mi>R</mi> <mrow> <mn>1</mn> <mi>kl</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msub> <mi>&Phi;</mi> <mi>kl</mi> </msub> <msubsup> <mi>U</mi> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>54</mn> <mo>)</mo> </mrow> </mrow> </math>

wherein E is_kIs a scalar used to normalize the transmit power of the relay node 804-k. In estimating the relay signal x_kIn the process of (1), equations (39) and (50) are used. The estimated relay signal x_kSent to the destination node.

Among the plurality of destination nodes, the l-th destination node 806-l to which the transmission signal from the l-th source node 802-l is addressed receives the relay signals from the K relay nodes. Thus, the signal r received at the l-th destination node 806-l_lIs represented as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>r</mi> <mn>1</mn> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>x</mi> <mi>k</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msubsup> <mi>R</mi> <mrow> <mn>2</mn> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mrow> <msub> <mi>&Theta;</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <mi>R</mi> </mrow> <mrow> <mn>1</mn> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>A</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>&Phi;</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <msub> <mrow> <mo>+</mo> <mi>z</mi> </mrow> <mi>l</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>55</mn> <mo>)</mo> </mrow> </mrow> </math>

(M × 1 matrix)

Wherein z is_lIs a noise component introduced between the plurality of relay nodes and the l-th destination node. Equation (55) is estimated using the following facts: if l ≠ l', the channel matrix G_klAnd a second unitary matrix A_kl’Are orthogonal to each other. The fact that equation (51) holds when l ═ l' is also utilized.

As is clear from equation (55)At the received signal r_lIn (1), a transmission signal s from a desired source node is transmitted_lWith transmitted signals s from other source nodes_l’(l.noteq.l') isolation. In other words, interference between multiple source nodes is substantially reduced; however, there is also interference remaining between multiple signal components within the transmitted signal from the desired source node. Therefore, the destination node must perform normal signal separation, which is usually performed in the MIMO scheme, to detect the desired signal s from the received signal_l。

Incidentally, as understood from the definition of the respective matrices, for transmitting the signal s_lOf (2) matrix Q^H _2kΘ_klR_1klIs the lower right triangular matrix. Thus, if one of the signal components (e.g., s) that depends only on the upper-right triangular matrix element is determined_lM) The transmitted signal s can be determined successively one by one_lOf the signal component (c). Therefore, the calculation workload of the signal calculation can be reduced as compared with the fourth embodiment.

Further, matrix elements arranged from the upper right corner to the lower left corner of the lower right triangular matrix (i.e., those elements whose sum of row number and column number (i + j) is equal to a prescribed number (column number plus 1)) are paired with the transmission signal s_lIs larger than the other matrix elements. Such matrix elements are positive real numbers and do not include imaginary components. Thus, the contributions Σ E from the L relay nodes are combined in phase_kR^H _2kΘ_klR_1klAnd the signal to noise power ratio can be increased at the destination node. In addition, because of the matrix A_klAnd U_klIs a unitary matrix, so the noise n_kNot amplified by these matrices. Therefore, the accuracy of detecting the signal at the destination node can be further improved.

Fig. 15 is a graph showing the simulation result of the fifth embodiment compared with the related art. The graph shows the traversal capacity as a function of the power-to-noise ratio (PNR). The method of the fifth embodiment and the prior art method are simulated with the number of relay nodes K2 and K8. The number of source nodes and the number of destination nodes are also two. The number of antennas of the source node and the destination node is four (4), and the number of antennas of the relay node is eight (8). In general, as PNR increases (i.e., as the signal power level increases), capacity increases. When the number of relay nodes increases, the capacity increases. As clearly shown in the graph, the technique of the fifth embodiment is superior to the prior art in that the capacity is improved by about 5bps/Hz with the same number of relay nodes.

(example 6)

Fig. 13 shows another example of arithmetic operations performed in the ith reception filter estimator 1004-l, the ith intermediate filter estimator 1008-l, and the ith transmission filter estimator 1012-l. The operations performed by the reception filter 1002-1 and the reception filter estimator 1004-l are the same as those described in the fourth embodiment. The operations performed by the transmit filter 1010-1 and the transmit filter estimator 1012-l are the same as those of the known art.

In the sixth embodiment, the l intermediate filter 1006-1 converts the signal y 'output from the reception filter 1004-1'_klMultiplication by a transformation matrix phi_klTo generate a signal phi_kly’_kl. The transformation matrix phi_klIs calculated by the ith intermediate filter estimator 1008-1.

The intermediate filter estimator 1008-1 pairs the matrix U as shown in equation (60)^H _klH_lkQR decomposition is performed.

U^H _klH_lk＝Q_1klR_1kl (60)

Wherein Q is_1klIs a unitary matrix with dimensions of N-M (L-1) rows and M columns, R_1klIs an M x M upper right triangular matrix. The intermediate filter estimator 1008-1 estimates the matrix Θ using a triangular matrix satisfying equation (60)_klIt is represented as:

intermediate filter estimator 1008-1 finally estimates transformation matrix Φ based on the above matrix_klIt is represented as:

Φ_kl＝Θ_klQ^H _1kl (61)

wherein phi_klIs a matrix of M (N-M (L-1)).

The l-th intermediate filter 1006-1 converts the signal phi_kly’_klOutput to transmit filter 1010-l. The operations performed by the transmit filter 1010-1 and the transmit filter estimator 1012-L are the same as those of the conventional art, and thus, the L matrices A are determined_kl(l＝1，...，L)(w^f _kl＝A_kl) So as to satisfy:

[A_kl，…，A_lL]＝G^H _k(G_kG_k ^H)^-1，

wherein G is_kIs defined as:

G_k＝[G^H _kl，…，G^H _kL]。

transmit filter 1010-1 multiplies an input signal by a matrix A_klAnd outputs the multiplied signal.

The output of the transmit filter 1010-l is connected to the input of a signal combining unit 1014. The signal combining unit 1014 generates a relay signal x_k. The relay signal x_kIs represented as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>X</mi> <mi>k</mi> </msub> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msub> <mi>&Phi;</mi> <mi>kl</mi> </msub> <msubsup> <mi>y</mi> <mi>kl</mi> <mo>&prime;</mo> </msubsup> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <mrow> <mo>(</mo> <msub> <mi>&Theta;</mi> <mi>kl</mi> </msub> <msubsup> <mi>Q</mi> <mrow> <mn>1</mn> <mi>kl</mi> </mrow> <mi>H</mi> </msubsup> <mo>)</mo> </mrow> <mrow> <mo>(</mo> <msubsup> <mi>U</mi> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>H</mi> <mi>lk</mi> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <msubsup> <mrow> <mo>+</mo> <mi>U</mi> </mrow> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msub> <mi>&Theta;</mi> <mi>kl</mi> </msub> <msub> <mi>R</mi> <mrow> <mn>1</mn> <mi>kl</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mi>kl</mi> </msub> <msub> <mi>&Phi;</mi> <mi>kl</mi> </msub> <msubsup> <mi>U</mi> <mi>kl</mi> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>62</mn> <mo>)</mo> </mrow> </mrow> </math>

wherein E is_kIs a scalar used to normalize the transmit power of the relay node 804-k. In estimating the relay signal x_kIn the process of (1), equations (39) and (60) are used. The estimated relay signal x_kSent to the destination node.

At a plurality of destination nodesOf these, the lth destination node 806-l to which the transmission signal from the lth source node 802-l is addressed receives the relayed signals from the K relay nodes. Thus, the signal r received at the l-th destination node 806-l_lIs represented as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>r</mi> <mn>1</mn> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>x</mi> <mi>k</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msub> <mi>&Theta;</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>R</mi> <mrow> <mn>1</mn> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msub> <mi>&Phi;</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>U</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <msub> <mrow> <mo>+</mo> <mi>z</mi> </mrow> <mi>l</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>63</mn> <mo>)</mo> </mrow> </mrow> </math>

(M × 1 matrix)

Wherein z is_lIs a noise component introduced between the plurality of relay nodes and the l-th destination node. Equation (63) is estimated using the following facts: channel matrix G_klAnd unitary matrix a_kl’Are orthogonal to each other.

As is clear from equation (63)At the received signal r_lIn (1), a transmission signal s from a desired source node is transmitted_lWith transmitted signals s from other source nodes_l’(l ≠ l') is separated, thereby substantially reducing interference between multiple source nodes. However, there is also interference remaining between multiple signal components within the transmitted signal from the desired source node. Therefore, the destination node must perform normal signal separation, which is usually performed in a MIMO scheme, to detect the desired signal s from the received signal_l。

Incidentally, as understood from the definition of the respective matrices, for transmitting the signal s_lMatrix theta of_klR_1klIs the upper right triangular matrix. Thus, if one of the signal components (e.g., s) that depends only on the bottom-right matrix element is determined_lM) The transmitted signal s can be determined successively one by one_lOf the signal component (c). Therefore, the calculation workload of the signal calculation can be reduced as compared with the fourth embodiment.

In addition, the matrix Θ_klR_1klIs sent as a signal s_lIs larger than the other matrix elements. Such matrix elements are positive real numbers and do not include imaginary components. Thus, the contributions Σ E from L relay nodes are combined in phase_kΘ_klR_1klAnd the signal to noise power ratio can be increased at the destination node. In addition, because of the matrix A_klAnd U_klIs a unitary matrix, so the noise n_kNot augmented by these matrices. Therefore, the signal detection accuracy at the destination node can be further improved.

(example 7)

In embodiments 4, 5 and 6, the reception by the reception signal y is used_kMultiplication by a unitary matrix U^H _klAnd the signal y 'obtained'_klThe unitary matrix U^H _klIs estimated by singular value decomposition. However, the signal processing described in embodiments 4, 5 and 6 can be applied by converting the reception signal y_kMultiplying by a weighting matrix W^b _klAnd the signal y 'obtained'_klAs in the conventional technique (W)^b _klY_k＝s_l+W^b _kln_k)。

In this case, transmit filter 1010-l may output a transmit signal y'_kl(＝W^b _kly_k) Multiplied by the unitary matrix a described in the fourth embodiment_klAnd the resulting signal.

Alternatively, transmit filter 1010-l may output a transmit signal y'_kl(＝W^b _kly_k) Multiplied by the matrix phi described in the fifth embodiment_kl(＝Q_2klΘ_klQ^H _kl) And unitary matrix a_klAnd the resulting signal.

Still alternatively, transmit filter 1010-l may output a transmit signal y'_kl(＝W^b _kly_k) Multiplied by the matrix phi described in the sixth embodiment_kl(equal to Θ_klQ^H _1kl) And unitary matrix a described in the fourth embodiment_klAnd the resulting signal.

(example 8)

In the eighth embodiment, as in the seventh embodiment, W is added^b _klApplied to the receive filter. As shown in fig. 14, the reception filter, the intermediate filter, and the transmission filter of the relay node perform arithmetic operations to generate signals. The relay node generates a second unitary matrix a as in the fourth embodiment using the method described in the first embodiment_kl. The second unitary matrix a is obtained by performing singular value decomposition on a plurality of channel matrices as shown in equation (40)_kl。

Then, to the matrix (G)_klA_kl)^HQR decomposition is performed.

(G_klA_kl)^H＝Q_2klR_2kl

Wherein Q is_2klIs (N-M (L-1))X M matrix whose column vectors are mutually orthogonal (referred to as unitary matrix in this application), and R_2klIs an M × M matrix and is an upper right triangular matrix.

Estimating a diagonal matrix Θ using the triangular matrix_kl. The diagonal matrix Θ_klIs defined as:

based on diagonal matrix theta_klAnd unitary matrix Q_2klEstimating MxM transformation matrix phi_klAs shown in equation (65).

Φ_kl＝Q_2klΘ_kl (65)

The relay node also estimates a weighting matrix W^b _klWhich is defined by equation (66).

<math> <mrow> <msup> <mrow> <mo>[</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mn>1</mn> </mrow> <mi>bT</mi> </msubsup> <mo>.</mo> <mo>.</mo> <mo>.</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>bT</mi> </msubsup> <mo>.</mo> <mo>.</mo> <mo>.</mo> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>L</mi> </mrow> <mi>bT</mi> </msubsup> <mo>]</mo> </mrow> <mi>T</mi> </msup> <mo>=</mo> <msup> <mrow> <mo>(</mo> <msubsup> <mi>H</mi> <mi>k</mi> <mi>H</mi> </msubsup> <mo>&CenterDot;</mo> <msub> <mi>H</mi> <mi>k</mi> </msub> <mo>)</mo> </mrow> <mrow> <mo>-</mo> <mn>1</mn> </mrow> </msup> <mo>&CenterDot;</mo> <msubsup> <mi>H</mi> <mi>k</mi> <mi>H</mi> </msubsup> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>66</mn> <mo>)</mo> </mrow> </mrow> </math>

(ML × N matrix)

Equation (66) is the same as equation (31-1) already described in the fourth embodiment.

Relay signal x is generated by equation (67) using unitary, transform and weighting matrices_kAnd sends it to the destination node.

<math> <mrow> <msub> <mi>x</mi> <mi>k</mi> </msub> <mo>=</mo> <msub> <mi>E</mi> <mi>k</mi> </msub> <munderover> <mi>&Sigma;</mi> <mrow> <mi>l</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>L</mi> </munderover> <msub> <mi>A</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>&Phi;</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>b</mi> </msubsup> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>b</mi> </msubsup> <mo>&CenterDot;</mo> <msub> <mi>y</mi> <mi>k</mi> </msub> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>67</mn> <mo>)</mo> </mrow> </mrow> </math>

(N × 1 matrix)

A signal r received at a target destination node (for convenience, it will be referred to as the lth destination node)₁Is represented as:

<math> <mrow> <mfenced open='' close=''> <mtable> <mtr> <mtd> <msub> <mi>r</mi> <mn>1</mn> </msub> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>G</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>x</mi> <mi>k</mi> </msub> </mtd> </mtr> <mtr> <mtd> <mo>=</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msubsup> <mi>R</mi> <mrow> <mn>2</mn> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>&Theta;</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msub> <mi>s</mi> <mi>l</mi> </msub> <mo>+</mo> <munderover> <mi>&Sigma;</mi> <mrow> <mi>k</mi> <mo>=</mo> <mn>1</mn> </mrow> <mi>K</mi> </munderover> <msub> <mi>E</mi> <mi>k</mi> </msub> <msubsup> <mi>R</mi> <mrow> <mn>2</mn> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>H</mi> </msubsup> <msub> <mi>&Phi;</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> </msub> <msubsup> <mi>W</mi> <mrow> <mi>k</mi> <mo>,</mo> <mi>l</mi> </mrow> <mi>b</mi> </msubsup> <msub> <mi>n</mi> <mi>k</mi> </msub> <msub> <mrow> <mo>+</mo> <mi>z</mi> </mrow> <mi>l</mi> </msub> </mtd> </mtr> </mtable> </mfenced> <mo>-</mo> <mo>-</mo> <mo>-</mo> <mrow> <mo>(</mo> <mn>68</mn> <mo>)</mo> </mrow> </mrow> </math>

(M × 1 matrix)

Wherein, the first term (R) on the right side^H _2klΘ_kl) Is the lower left triangular matrix and its diagonal elements are positive real numbers. Thus, when the K relay signals from the K relay nodes are combined at the destination node, the diagonal elements are combined in phase. As a result, the power-to-noise ratio at the destination node can be improved, and the transmission signal s can be accurately detected using the successive interference cancellation method_l。

(example 9)

In the ninth embodiment, the relay node 14 detects a signal based on a method using a unitary matrix or a Zero Forcing (ZF) method according to a channel state. When a unitary matrix is used, the unitary matrix is estimated by the singular value decomposition described above. When zero forcing is used, the ZF weighting factor is calculated by the Moore-Penrose inverse matrix.

Generating a relay signal x according to a channel status_kAnd transmits the relay signal to the destination node. The destination node detects the transmission signal transmitted from the source node in the manner described above. The quality of the channel state may be determined at the relay node by a channel estimator (fig. 3 or fig. 5). Alternatively, it is possible toThe quality of the channel state is determined based on the ratio of the power level of the desired wave to the power level of the undesired wave (e.g., SIR or SNR).

For example, a relay node estimates the channel state SNR between a source node and the relay node_HAnd channel state SNR between the relay node and the destination node_G。

If SNR_H＞＞SNR_GThe channel state between the source node and the relay node is very good. Thus, even if zero forcing is applied between the source node and the relay node, the noise amplification is sufficiently small and can be neglected. On the other hand, since the influence of noise amplification increases between the relay node and the destination node, a method using a unitary matrix is applied between the relay node and the destination node (similar to fig. 14).

On the other hand, if the SNR_H＜＜SNR_GThe reverse process is performed (shown in fig. 13).

The intermediate filter may be appropriately selected from those shown in fig. 12 to 14. By adaptively changing the relay scheme at the relay node 14 according to the quality of the channel state, the reception quality characteristics of the destination node can be improved.

The present patent application is based on and claims priority from Japanese patent application No. 2004-. The entire contents of which are hereby incorporated by reference.

Claims (13)

1. A wireless communication system for transmitting a transmission signal from a desired source node among a plurality of source nodes to a target destination node through a relay node,

wherein the relay node comprises:

a first unitary matrix estimation unit configured to estimate a first unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of source nodes except for the desired source node;

a second unitary matrix estimation unit configured to estimate a second unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of destination nodes other than a target destination node; and

a transmitting unit configured to transmit a relay signal generated by multiplying a received signal by the first and second unitary matrices to the target destination node;

wherein the destination node detects a transmission signal transmitted from a desired source node from the received relay signal.

2. A communication node for relaying a transmission signal transmitted from a desired source node to a target destination node between a plurality of source nodes and a plurality of destination nodes, comprising:

a first unitary matrix estimation unit configured to estimate a first unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of source nodes except for the desired source node;

a second unitary matrix estimation unit configured to estimate a second unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of destination nodes other than a target destination node; and

a transmitting unit configured to transmit a relay signal generated by multiplying a received signal by the first and second unitary matrices to the target destination node.

3. The communications node of claim 2, further comprising:

a transform matrix estimation unit configured to estimate a transform matrix composed of a product of a matrix in which matrix elements of an ith row and a jth column are zero if i + j does not satisfy a prescribed value and one or more unitary matrices;

wherein the transmission unit transmits a relay signal generated by multiplying a reception signal by the first unitary matrix, the transformation matrix, and the second unitary matrix to the destination node.

4. The communications node of claim 2, further comprising:

a transform matrix estimation unit configured to estimate a transform matrix composed of a product of an diagonal matrix and a unitary matrix derived from a channel matrix between the source node and the relay node or between the relay node and the destination node;

wherein the transmission unit transmits a relay signal generated by multiplying a reception signal by the first unitary matrix, the transformation matrix, and the second unitary matrix to the destination node.

5. A communication method for relaying a transmission signal transmitted from a desired source node among a plurality of source nodes to a destination node through a relay node, comprising the steps of:

estimating, at the relay node, a first unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of source nodes other than the desired source node, and estimating a second unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and a plurality of destination nodes other than the destination node;

transmitting a relay signal generated at the relay node by multiplying a received signal by the first and second unitary matrices to the destination node; and

at the destination node, a transmission signal transmitted from a desired source node is detected from the received relay signal.

6. A wireless communication system for transmitting a transmission signal from a desired source node among a plurality of source nodes to a target destination node through a relay node, comprising:

a matrix estimation unit configured to estimate a Moore-Penrose inverse matrix derived from a plurality of channel matrices between the relay node and a plurality of nodes;

a relay signal generation unit configured to generate a relay signal by multiplying a received signal by a weighting matrix defining the Moore-Penrose inverse matrix and a unitary matrix obtained by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of nodes other than a prescribed node; and

a transmitting unit configured to transmit the relay signal to the destination node;

wherein the destination node detects the transmission signal from the received relay signal.

7. A communication node for relaying a transmission signal transmitted from a desired source node among a plurality of source nodes to a destination node, comprising:

a matrix estimation unit configured to estimate a Moore-Penrose inverse matrix derived from a plurality of channel matrices between the relay node and a plurality of nodes;

a relay signal generation unit configured to generate a relay signal by multiplying a received signal by a weighting matrix defining the Moore-Penrose inverse matrix and a first unitary matrix obtained by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of nodes other than a prescribed node; and

a transmitting unit configured to transmit the relay signal to the destination node.

8. The communications node of claim 7, further comprising:

a transform matrix estimation unit configured to estimate a transform matrix composed of a product of a matrix in which matrix elements of an ith row and a jth column are zero if i + j does not satisfy a prescribed value and one or more unitary matrices;

wherein the transmission unit transmits a relay signal generated by multiplying a reception signal by the transform matrix and the first unitary matrix to the destination node.

9. The communications node of claim 7, further comprising:

a transform matrix estimation unit configured to estimate a transform matrix composed of a product of an diagonal matrix and a unitary matrix derived from a channel matrix between the source node and the relay node or between the relay node and the destination node;

wherein the transmission unit transmits a relay signal generated by multiplying a reception signal by the transform matrix and the first unitary matrix to the destination node.

10. A communication method for relaying a transmission signal transmitted from a desired source node among a plurality of source nodes to a destination node through a relay node, comprising the steps of:

estimating, at the relay node, a Moore-Penrose inverse matrix derived from a plurality of channel matrices between the relay node and a plurality of nodes;

generating a relay signal by multiplying a received signal by a weighting matrix defining the Moore-Penrose inverse matrix and a unitary matrix obtained by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of nodes except a prescribed node;

sending the relay signal to the destination node; and

detecting, at the destination node, the transmission signal from the received relay signal.

11. A communication node for relaying a transmission signal transmitted from a desired source node among a plurality of source nodes to a destination node, comprising:

a matrix estimation unit configured to estimate a Moore-Penrose inverse matrix derived from a plurality of channel matrices between the relay node and a plurality of nodes;

a first unitary matrix estimation unit configured to estimate a first unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of source nodes other than the desired source node;

a second unitary matrix estimation unit configured to estimate a second unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of destination nodes other than the destination node; and

a relay signal generating unit configured to generate a relay signal by multiplying a received signal by two of a weighting matrix defining the Moore-Penrose inverse matrix, the first unitary matrix, and the second unitary matrix; and

a transmitting unit configured to transmit the relay signal to the destination node.

12. The communication node of claim 11, wherein the two of the matrices are selected based on a quality of a channel state.

13. A communication method for relaying a transmission signal transmitted from a desired source node among a plurality of source nodes to a destination node through a relay node, comprising the steps of:

estimating, at the relay node, a Moore-Penrose inverse matrix derived from a plurality of channel matrices between the relay node and a plurality of nodes;

estimating, at the relay node, a first unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of source nodes other than the desired source node, and estimating a second unitary matrix by performing singular value decomposition on one or more channel matrices between the relay node and the plurality of destination nodes other than the destination node;

generating a relay signal by multiplying a received signal by two of a weighting matrix defining the Moore-Penrose inverse matrix, the first unitary matrix, and the second unitary matrix; and

and sending the relay signal to the destination node.

Images