Patent Application
20110242962
The present invention relates to a method of multiuser MIMO (Multiple-Input Multiple-Output) wireless communication among a transmitter and a plurality of receivers, and a device of multiuser MIMO wireless communication which communicates in the method of multiuser MIMO wireless communication.
To this day, an achievable upper limit value of communication speed in the multiuser MIMO has been theoretically derived (for example, see S. Vishwanath, N. Jindal, and A. Goldsmith, “Duality, achievable rates, and sum-rate capacity of gaussian MIMO broadcast channels”, IEEE Trans. Inform. Theory, vol. 49, no. 10, pp. 2658-2668, Oct. 2003 (Non-Patent Document 1)).
Also, to this day, as technology of obtaining performance close to a theoretical limitation in the multiuser MIMO method, there has been technology called DPC (Dirty Paper Coding) (for example, see M. H. M. Costa, “Writing on dirty paper”, IEEE Trans. Inform. Theory, vol. 29, no. 3, pp. 439-441, May 1983 (Non-Patent Document 2)).
Also, to this day, there has been a ZF-DPC (Zero-Forcing DPC) method achieving the multiuser MIMO method by subjecting a channel matrix to a QR decomposition and coding the same by DPC based on a triangular matrix L (for example, see Z. Tu and R. S. Blum, “Multiuser diversity for a dirty paper approach”, IEEE Commun. Lett., vol. 7, no. 8, pp. 370-372, Aug. 2003 (Non-Patent Document 3)).
Moreover, to this day, there has been a CZF-DPC (Cooperative ZF-DPC) method achieving the multiuser MIMO method by subjecting a channel matrix to a singular value decomposition, selecting a stream and updating a transmit weight matrix and a receive weight matrix, and repeating update of the channel matrix so that the channel matrix is orthogonal to the transmit weight matrix (for example, see Japanese Patent Application Laid-Open Publication (translation of PCT application) No. 2008-519510 (Patent Document 1)).
Patent Document 1: Japanese Patent Application Laid-Open Publication (translation of PCT application) No. 2008-519510
Non-Patent Document 1: S. Vishwanath, N. Jindal, and A. Goldsmith, “Duality, achievable rates, and sum-rate capacity of gaussian MIMO broadcast channels”, IEEE Trans. Inform. Theory, vol. 49, no. 10, pp. 2658-2668, Oct. 2003
Non-Patent Document 2: M. H. M. Costa, “Writing on dirty paper”, IEEE Trans. Inform. Theory, vol. 29, no. 3, pp. 439-441, May 1983
Non-Patent Document 3: Z. Tu and R. S. Blum, “Multiuser diversity for a dirty paper approach”, IEEE Commun. Lett., vol. 7, no. 8, pp. 370-372, Aug. 2003
The MIMO wireless communication method is used in wireless LAN, WiMAX, etc. MIMO is a communication method which improves communication speed more than the situation in which a transmitter and a receiver are using one antenna, and MIMO improves stability of communication. When MIMO is used, communications can be performed in a plurality of different signals using the same amount of time and the same frequency band.
The transmitter not only transmits a plurality of different signals to a single receiver, and also is able to transmit the signals to a plurality of receivers as illustrated in the schematic diagram of
Non-Patent Document 3 describes a ZF-DPC (Zero-Forcing DPC) method achieving the multiuser MIMO method by subjecting a channel matrix to an LQ decomposition and performing a DPC coding based on a triangular matrix L. Further, Non-Patent Document 3 describes that performance can be improved by changing the order of rows in the channel matrix upon the LQ decomposition.
Patent Document 1 describes a CZF-DPC (Cooperative ZF-DPC) method achieving the multiuser MIMO method by subjecting a channel matrix to a singular value decomposition, selecting a stream and updating a transmit weight matrix and a receive weight matrix, and repeating updating so that channel matrix is perpendicular to the transmit weight matrix. Further, Patent Document 1 describes that THP (Tomlinson-Harachima Precoding) can be used as a method of mounting the DPC.
It has been known that an achievable speed of the ZF-DPC method is low and it does not reach the communication speed of the theoretical limitation.
It has been known that a necessary computation amount is large in the CZF-DPC method. Therefore, a necessary circuit size is large for computation. In addition, as the amount of time taken for the computation is increased, a time from a generation of a request of data transmission to actual data transmission, that is, delay is increased. Moreover, the receive weight matrix can be determined only by the transmitter, it is necessary to notify the receiver about the determined receive weight matrix, and thus protocol overhead occurs.
The present invention achieves a communication speed close to the theoretical limitation with suppressing the computation amount small and without increasing protocol overhead.
The typical ones of the inventions will be briefly described as follows. More specifically, a method of multiuser MIMO wireless communication according to the present invention transmits from a first wireless communication device to N devices of, i.e., 2-1 to 2-N wireless communication devices using a same frequency band and a same duration, the method including: a first step of estimating a channel matrix among the first wireless communication device to the 2-1 to 2-N wireless communication devices; a second step of determining a receive weight matrix to be multiplied with a received signal vector in the 2-1 to 2-N wireless communication devices; a third step of updating the channel matrix by a multiply of the receive weight matrix and the channel matrix; a fourth step of decomposing the channel matrix into a lower triangular matrix L and a unitary matrix Q by an LQ decomposition of the updated channel matrix; a fifth step of subjecting a transmit signal vector to a DPC coding based on the lower triangular matrix L; a sixth step of transmitting a transmitted signal vector obtained by multiplying the coded transmitted signal vector and a conjugate transpose matrix of the unitary matrix Q from the first wireless communication device; a seventh step of multiplying a received signal vector received by the 2-1 to 2-N wireless communication devices with the determined receive weight matrix; and an eighth step of subjecting the received signal vector obtained by multiplying the receive weight matrix to a DPC decoding.
In addition, a device of multiuser MIMO wireless communication according to the present invention transmits from a first wireless communication device to N devices of, i.e., 2-1 to 2-N wireless communication devices using a same frequency band and a same duration, in which the first wireless communication device includes a functional block performing: a first step of determining a receive weight matrix to be multiplied with a received signal vector in the 2-1 to 2-N wireless communication devices; a second step of updating the channel matrix by multiplying the receive weight matrix and the channel matrix; a third step of decomposing the channel matrix into a lower triangular matrix L and a unitary matrix Q by an LQ decomposition of the updated channel matrix; a fourth step of subjecting a transmitted signal vector to a DPC coding based on the lower triangular matrix L; and a fifth step of transmitting a transmitted signal vector obtained by multiplying the coded transmitted signal vector with a conjugate transpose matrix of the unitary matrix Q to the 2-1 to 2-N wireless communication devices.
According to the present invention, a high communication speed can be achieved in the multiuser MIMO method with a small computation amount.
In a method of multiuser MIMO wireless communication according to the present invention, a receive weight matrix is determined based on a channel matrix, the channel matrix is updated by multiplying the receive weight matrix and the channel matrix, a transmitted signal vector is generated by a ZF-DPC method based on the updated channel matrix and transmitted, and a receive weight matrix is multiplied with a received signal vector and then subjected to a DPC decoding.
One method of determining the receive weight matrix is a method of taking a conjugate transpose matrix of a left singular matrix obtained when subjecting the channel matrix to a singular value decomposition.
According to the present invention, as in the conceptual diagram illustrated in
Computation of matrix is expressed as follows.
A matrix (Expression 1) formed by aligning a plurality of matrices A1, A2, . . . , Ax in the column direction is expressed as [A1, A2, . . . , Ax].
(A1 A2 . . . Ax)=[A1, A2, . . . , Ax] Expression 1
A matrix (Expression 2) formed by aligning a plurality of matrices A1, A2, . . . , Ax in the row direction is expressed as [A1; A2; . . . ; Ax].
A conjugate transpose matrix of the matrix A is expressed as Expression 3.
AH Expression 3
Hereinafter, embodiments of the present invention will be described. Note that “t” at the end of referential marks means processing by a transmitter, and “r” means processing by a receiver.
In the present procedure, a channel matrix is estimated in a block 101, a receive weight matrix is determined in a block 102 based on the channel matrix, the channel matrix is updated by multiplying the receive weight matrix and the channel matrix in a block 103, a transmitted signal vector is generated by a ZF-DPC method based on the updated channel matrix and transmitted in a block 104, and a received signal vector is converted by being multiplied with the receive weight matrix in a block 105 and then subjected to a DPC decoding in a block 106.
The channel matrix means a complex gain upon propagation of wireless signals from each antenna of the transmitters to each antenna of the receivers. A relationship between a transmitted signal vector x=[x1; x2; . . . ; xm] composed of signals x1, x2, . . . , xm transmitted from antennas 1, 2, . . . , m of the transmitters and a received signal vector y=[y1; y2; . . . ; yn] composing signals y1, y2, yn received by antennas 1, 2, . . . , n of the receivers can be expressed as Expression 4 with a channel matrix H (noise is omitted in the expression).
To estimate the channel matrix, changes in amplitude and phase of a signal may be detected after sending a known signal decided between the transmitter and receiver. The known signal is called training signal.
Since channel matrices between the transmitter and the N devices of receivers are estimated, N pieces of channel matrices Hi (i=receiver number, 1 to N) are obtained in the block 101. In the block 102, based on the obtained channel matrices Hi, a receive weight matrix Ri to be multiplied with a received signal vector in a receiver i is determined. In the block 103, a product RixHi of the receive weight matrix Ri and the channel matrix Hi is computed, and the channel matrix Hi is updated by a result of that. In the block 104, a transmitted signal vector is processed in the technique of ZF-DPC method based on a channel matrix H′=[H1; H2; . . . ; HN] (Expression 5) to all of the receivers, in which the obtained channel matrices Hi are lined up in the row direction, and transmitted.
In the block 105, the receive weight matrix Ri is multiplied with the received signal vector of the receiver i. As the received signal vector is a vector yi=[yi1; yi2; . . . ; yin] having the same number of components as the number of the receiver antennas, a signal obtained in the block 105 is Rixyi. This signal is subjected to a DPC decoding in the block 106.
While it is necessary that the block 104 is carried out in the transmitter and the blocks 105 and 106 are carried out in the receiver, the blocks 101, 102, 103 may be carried out in both the transmitter and receiver.
H=UΣV
H Expression 6
Here, U and V are unitary matrices, and Σ is a matrix in which only diagonal components are not zero. In the present invention, U is called left singular matrix, and V is called right singular matrix. A conjugate transpose matrix of the left singular matrix obtained by a singular value decomposition in a block 112 is set as the receive weight matrix.
A=Q′R′ Expression 7
Here, Q′ is a unitary matrix, and R′ is an upper triangular matrix. The upper triangular matrix R′ is a different matrix from the receive weight matrix R. A conjugate transpose matrix of the unitary matrix Q′ obtained by the QR decomposition in the block 114 is set as the receive weight matrix.
H′=LQ Expression 8
Here, Lisa lower triangular matrix, and Q is a unitary matrix. the transmitted signal vector is subjected to a DPC coding based on the lower triangular matrix L in a block 108. In a block 109, a conjugate transpose matrix QH of the unitary matrix Q is multiplied with a signal vector which has been subjected to a DPC coding, and an obtained signal vector is transmitted in a block 110. As the signal vector which has been subjected to the DPC coding in the block 108 is a vector x=[x1; x2; . . . ; xm] having the same number of components as the number of the transmitter antennas, a signal vector obtained in the block 110 is QHxx.
Non-Patent Document 3 describes a method (Greedy algorithm) for improving performance by computing while exchanging rows of the channel matrix H, and it is effective to use the Greedy algorithm also in the present invention. In the ZF-DPC, the Greedy algorithm allocates signals in descending order of sum of squared absolute values of respective row components in the channel matrix to all the receivers. That is, it allocates in descending order of gain of the received signals. Therefore, when a receive weight matrix is previously determined and multiplied with a channel matrix as the present invention, it is possible to change an effective gain of signal and thus it is possible to adjust so that a high communication speed is obtained. As described above, there are examples of the method of subjecting the channel matrix to a singular value decomposition and taking a conjugate transpose matrix of the left singular matrix as the receive weight matrix, and the method of taking a conjugate transpose matrix of the unitary matrix obtained by the QR decomposition of the channel matrix and its conjugate transpose matrix as the receive weight matrix. In this case, by combining the received signal by the receive weight matrix, a signal having a large effective gain is generated, thereby obtaining a larger gain upon allocating signals by the Greedy algorithm.
Patent Document 1 describes that THP can be used as the DPC coding carried out in the block 108, and the DPC coding can be performed using THP also in the present invention.
The present invention is also effective in reduction of computation amount. For carrying out MIMO wireless communication, a singular value decomposition is a process having the biggest computation amount. While it is necessary to carry out the singular value decomposition for (number of receivers) x (number of parallel transmissions) cycles in the CZF-DPC, in the method of the present invention, it is only necessary to carry out for (number of receivers) cycles when using the method of determining receive weight matrix by the singular value decomposition of
Upon the determination of receive weight matrix by the QR decomposition of
It is not necessary to use the method of determining receive weight matrix all the same from the receiver 1 to receiver N. For example, a receiver which determines a receive weight matrix by the singular value decomposition and a receiver which determines a receive weight matrix by the QR decomposition may exist together.
The receive weight determination method can be other methods than those illustrated in
The transmitter transmits a training signal for estimating a channel matrix in a block 201t. A receiver i (i=1 to N) receives the training signal in a block 202r, estimates a channel matrix in a block 203r based on the training signal, and transmits channel matrix information to the transmitter in a block 204r. The transmitter receives the channel matrix information transmitted from the receiver i in a block 205t. As the channel matrix information is transmitted from the N pieces of receivers, in accordance with that, a plurality of channel matrix information items are received in the block 205t. The transmitter determines a receive weight matrix Ri based on the obtained channel matrix Hi in a block 206t. The transmitter calculates a product of the receive weight matrix Ri and the channel matrix Hi in a block 207t and updates the channel matrix Hi according to a result of the calculation. The transmitter transmits data in the technique of ZF-DPC based on the updated channel matrix Hi in a block 208t. In the block 208t, data may be transmitted with adding a training signal thereto. The receiver i receives in a block 209r. The receiver determines a receive weight matrix Ri in a block 206r and multiplies the receive weight matrix Ri with a received signal in a block 210r, and performs a DPC decoding in a block 211r.
The procedure of the present embodiment is such that the estimation of channel matrix of the block 101 in the first embodiment is carried out by the receiver, and the determination of receive weight matrix in the block 102 and the update of channel matrix in the block 103 are carried out by the transmitter.
Based on the channel matrix obtained in the block 203r or the newly estimated channel matrix Hi upon receiving in the block 209r, the receive weight matrix Ri can be determined in the block 206r. This is possible because the receive weight matrix Ri of the receiver i can be determined only by the channel matrix Hi of the receiver i. For example, while the receive weight matrix is multiplied with the received signal also in CZF-DPC, since it is necessary to know all the channel matrices H1, H2, . . . , HN to determine the receive weight matrix Ri of the receiver i, it is not possible to determine the receive weight matrix by the receiver same as the present invention. Therefore, in CZF-DPC, the transmitter has to transmit information of the receive weight matrix to all the receivers and protocol overhead is increased; this problem is solved according to the present invention. In the present invention, while one extra step is added when using a channel matrix in other receivers than the receiver in the determination of receive weight matrix, there is no problem if information of the receive weight matrix determined in the block 206t is previously notified by the transmitter to the receiver.
To determine the receive weight matrix Ri based on the channel matrix Hi obtained in the block 203r, the receive weight matrix Ri can be determined prior to data reception by arranging the block 206r before the block 209r.
A transmitter transmits a training signal for estimating channel matrix in a block 201t. A receiver i (i=1 to N) receives a training signal in a block 202r and estimates a channel matrix Hi in a block 203r based on that, and determines a receive weight matrix Ri based on the obtained channel matrix Hi in a block 206r. A product of the receive weight matrix Ri and the channel matrix Hi is calculated in a block 207r, and the channel matrix Hi is updated by a result of the calculation. Channel matrix information updated in a block 204r is transmitted to the transmitter. Channel matrix information transmitted from a receiver i is received by the transmitter in a block 205t. As the channel matrix information is transmitted from N pieces of receivers, in accordance with that, a plurality of channel matrix information items are received in the block 205t. The transmitter transmits data in the technique of ZF-DPC based on the channel matrix Hi updated in a block 208t. In the block 208t, data may be transmitted with adding a training signal thereto. The receiver i receives in a block 209r. The receive weight matrix Ri is multiplied with the received signal in a block 210r and subjected to a DPC decomposition in a block 211r.
The procedure of the present embodiment is such that the estimation of channel matrix in the block 101, the determination of receive weight matrix in the block 102, and the update of channel matrix in the block 103 are carried out by the receiver.
The receive weight matrix Ri multiplied in the block 210r can be a receive weight matrix determined in the block 206r, or a receive weight matrix obtained by newly estimating a channel matrix upon data reception in the block 209r and recalculating based on the channel matrix. As the receive weight matrix can be determined by the receiver, protocol overhead does not occur as explained in the description of the second embodiment.
What is different from the second embodiment is to perform the update of channel matrix in the block 207r by the receiver instead of the transmitter. Therefore, processing about the receive weight matrix is not necessary in the transmitter, and the procedure by the transmitter is the same as the existing ZF-DPC method. According to the present embodiment, the transmitter does not change the procedure of the existing ZF-DPC method, and only the receiver is changed to the procedure of the present invention, thereby obtaining the effects of the present invention.
A receiver i (i=1 to N) transmits a training signal for estimating channel matrix in a block 201r. A transmitter receives a training signal in a block 202t, and estimates a channel matrix Hi based on that in a block 203t. As the training signal is transmitted from the N pieces of receivers, in accordance with that, a plurality of training signals are received in the block 202t. The transmitter determines a receive weight matrix Ri in a block 206t. The transmitter calculates a product of the receive weight matrix Ri and the channel matrix Hi in a block 207t and updates the channel matrix Hi based on a result of the calculation. The transmitter transmits data in the technique of ZF-DPC based on the updated channel matrix Hi in a block 208t. In the block 208t, data may be transmitted with adding the training signal thereto. The receiver i receives in a block 209r. As the training signal is normally added upon data transmission, the receiver i can estimate the channel matrix Hi based on the received training signal in a block 203r. A receive weight matrix is determined based on the channel matrix Hi in a block 206r, and the receive weight matrix is multiplied with a received signal in a block 210r. An obtained signal is subjected to a DPC decoding in a block 211r.
The procedure of the present embodiment is such that the estimation of channel matrix in the block 101, the determination of receive weight matrix in the block 102, and the update of channel matrix in the first embodiment are carried out by the transmitter.
In the present embodiment, using the duality between the channel matrix in the propagation from the transmitter to receiver and the channel matrix of the propagation from the receiver to transmitter, the channel matrix is estimated by the transmitter. In the second and third embodiments, two kinds of communication of the transmission of the training signal from the transmitter to the receiver before the data transmission and the transmission of channel matrix from the receiver to transmitter have been necessary. As compared with that, in the present embodiment, the procedure is finished only by the transmission of training signal from the receiver to transmitter, and thus protocol overhead can be reduced. Also, the receive weight matrix can be determined by the receiver, and thus protocol overhead does not occur as explained in the description of the third embodiment.
After the transmitted data is converted to a transmitted signal in a transmitted signal generating portion 301t, a ZF-DPC coding is carried out in a ZF-DPC coding portion 302t. Then, the transmitted data is subjected to a multiplexing in a signal multiplexing portion 303t, and added with a training signal in a training signal adding portion 304t, and inputted to an analog front end 309t and transmitted. A received signal is inputted from the analog front end 309t to a data reconstructing portion 308t and a channel matrix obtaining portion 305t. Received data is obtained from the data reconstructing portion 308t. The channel matrix obtaining portion 305t estimates a channel matrix by receiving a training signal and so forth and a receive weight matrix is determined in a receive weight matrix generating portion 306t. From a result of that, the channel matrix is updated in a channel matrix updating portion 307t, and inputted to the ZF-DPC coding portion 302t.
The transmitted signal generating portion 301t includes, in accordance with its communication method, a scrambler, an error correction coding portion, a QAM mapper, a transmission power controlling portion, etc. Also, an outputted signal is a signal vector having components of the number of parallel transmissions in the multiuser MIMO wireless communication.
The signal multiplexing portion 303t is necessary for a multiplex communication method such as the OFDM method. In the OFDM method, an inverse discrete Fourier transform is carried out in the signal multiplexing portion 303t. When the signal multiplexing such as single carrier transmission is not necessary, the signal multiplexing portion 303t is also not necessary.
While the training signal adding portion 304t is disposed immediately before the analog front end 309t, the disposing portion of the training signal adding portion 304t is not particularly limited to this. The training signal adding portion 304t may be disposed before the signal multiplexing portion 303t or before the ZF-DPC coding portion 302t.
The analog front end 309t mutually converts signals and radio waves and transmits and receives the same, including an antenna, a filter, a signal amplifier, an up/down converter, a quadrature modulator (orthogonal modulator)/demodulator, an AD converter, a DA converter, etc.
The data reconstructing portion 308t reconstructs data transmitted from the receiver to the transmitter, including an OFDM splitting portion (discrete Fourier transfer), a MIMO receive weight matrix multiplexing portion, a QAM demapper, an error correction decoding portion, a descrambler, etc. in accordance with its communication method.
In the following, checking from the procedure of the transmitter of the second embodiment will be made. By transmitting the training signal added in the training signal adding portion 304t from the analog front end 309t, the training signal transmission of the block 201t is carried out. By obtaining a channel matrix in the channel matrix obtaining portion 305t from the signal received by the analog front end 309t, the reception of channel matrix of the block 205t is carried out. In the receive weight matrix generating portion 306t and the channel matrix updating portion 307t, the determination of receive weight matrix of the block 206t and the update of channel matrix of the block 207t are carried out. By transmitting the signal subjected to processing in the ZF-DPC coding portion 302t from the analog front end 309t, the data transmission of the block 208t is carried out. In this manner, the transmission procedure described in the second embodiment can be carried out.
The transmitted data is converted to a transmitted signal in a transmitted signal generating portion 301r, and then transmitted by an analog front end 309r. The received signal is inputted from the analog front end 309r to a signal splitting portion 313r and a training signal extracting portion 316r. In the signal splitting portion, multiplexed signals are split and inputted to a receive weight matrix multiplying portion 314r. In the receive weight matrix multiplexing portion 314r, a receive weight matrix obtained from a receive weight matrix generating portion 306r is multiplied with the received signal, and a result of it is decoded by a DPC decoding portion 315r, thereby obtaining received data from a data reconstructing portion 308r. A training signal is extracted from the received signal in a training signal extracting portion 316r, and a channel matrix is obtained in a channel matrix estimating portion 317r based on a result of the extraction and is inputted to the receive weight matrix generating portion 306r. In the receive weight matrix generating portion 306r, a receive weight matrix is calculated from the channel matrix and inputted to the receive weight matrix multiplying portion 314r. The channel matrix obtained in the channel matrix estimating portion 317r is inputted also to the transmitted signal generating portion 301r, and used in the transmission to the transmitter.
The transmitted signal generating portion 301r includes, in accordance with its communication method, a scrambler, an error correction coding portion, a QAM mapper, a transmission power control portion, a MIMO transmission weight multiplying portion, an OFDM multiplexing portion (inverse discrete Fourier transform), etc.
The signal splitting portion 313r is necessary in a multiplex communication method such as the OFDM method. In the OFDM method, a discrete Fourier transfer is carried out in the signal splitting portion 313r. When a signal multiplexing such as single carrier transmission is not necessary, the signal splitting portion 313r is also not necessary.
A decoding processing corresponding to a DPC coding is performed in a DPC decoding portion 315r. Specifically, when THP is used in the DPC coding, a residue calculation is performed in the DPC decoding portion 315r.
The data decoding portion 308r includes a QAM demapper, an error correction decoding portion, a descrambler, etc.
The analog front end 309r mutually converts signals and radio waves and transmits and receives the same, including an antenna, a filter, a signal amplifier, an up/down converter, an orthogonal modulator/demodulator, an AD converter, a DA converter, etc.
In the following, checking from the procedure of the transmitter of the second embodiment will be made. A training signal is extracted in the training signal extracting portion 316r from the signal received in the analog front end 309r, and a channel matrix is estimated in the channel matrix estimating portion 317r, thereby carrying out the reception of training signal of the block 202r and the estimation of channel matrix of the block 203r. The channel matrix information is converted to a signal in the transmitted signal generating portion 301r and transmitted from the analog front end 309r, thereby carrying out the transmission of channel matrix of the block 204r. As the analog front end 309r receives signals, the data reception of the block 209r is carried out. The determination of receive weight matrix of the block 206r is carried out in the receive weight matrix generating portion 306r. In the receive weight matrix multiplying portion 314r and the DPC decoding portion 315r, the multiply of receive weight matrix of the block 210r and the DPC decoding of the block 211r are carried out, respectively. In this manner, the reception procedure described in the second embodiment can be carried out.
Transmitted data is converted to a transmitted signal in a transmitted signal generating portion 301r and then transmitted by an analog front end 309r. A received signal is inputted from the analog front end 309r to a signal splitting portion 313r and a training signal extracting portion 316r. Multiplexed signals are split in the signal splitting portion 313r and inputted to a receive weight matrix multiplying portion 314r. In the receive weight matrix multiplying portion 314r, a receive weight matrix obtained from a receive weight matrix generating portion 306r is multiplied with the received signal, and a result of the multiplying is decoded in a DPC decoding portion 315r, thereby obtaining received data from a data reconstructing portion 308r. A training signal is extracted from the received signal in a training signal extracting portion 316r, and a channel matrix is obtained in a channel matrix estimating portion 317r based on a result of the extraction and inputted to the receive weight matrix generating portion 306r. A receive weight matrix is calculated from a channel matrix in the receive weight matrix generating portion 306r and inputted to the receive weight matrix multiplying portion 314r and a channel matrix updating portion 307r. In the channel matrix updating portion 307r, the channel matrix is updated using the inputted receive weight matrix and inputted to the transmitted signal generating portion 301r to transmit to a transmitter.
The transmitted signal generator 301r includes, in accordance with its communication method, a scrambler, an error correction coding portion, a QM mapper, a transmission power control portion, a MIMO transmission weight multiplying portion, an OFDM multiplexing portion (inverse discrete Fourier transfer), etc.
The signal splitting portion 313r is necessary in a multiplex communication method such as the OFDM method. In the OFDM method, a discrete Fourier transfer is carried out in the signal splitting portion 313r. When the signal multiplexing such as single carrier transmission is not necessary, the signal splitting portion 313r is also not necessary.
In the DPC decoding portion 315r, a decoding processing corresponding to the DPC coding is performed. Specifically, when THP is used in the DPC coding, a residue calculation is performed in the DPC decoding portion 315r.
The data reconstructing portion 308r includes, in accordance with its communication method, a QAM demapper, an error correction decoding portion, a descrambler, etc.
The analog front end 309r mutually converts signals and radio waves and transmits and receives the same, including an antenna, a filter, a signal amplifier, an up/down converter, an orthogonal modulator/demodulator, an AD converter, a DA converter, etc.
In the following, checking from the procedure of the receiver of the third embodiment will be made. The training signal is extracted in the training signal extracting portion 316r from the signal received in the analog front end 309r, and the channel matrix is estimated in the channel matrix estimating portion 317r, thereby carrying out the reception of training signal of the block 202r and the estimation of channel matrix of the block 203r. In the receive weight matrix generating portion 306r and the channel matrix updating portion 307r, the determination of receive weight matrix of the block 206r and the update of channel matrix of the block 207r are carried out, respectively. Channel matrix information is converted to a signal and transmitted from the analog front end 309r, thereby carrying out the transmission of channel matrix of the block 204r. As the signal is received in the analog front end 309r, the data reception of the block 209r is carried out. In the receive weight matrix multiplying portion 314r and the DPC decoding portion 315r, the multiplying of receive weight matrix of the block 210r and the DPC decoding of the block 211r are carried out, respectively. In this manner, the reception procedure described in the third embodiment can be carried out.
Transmitted data is converted to a transmitted signal in a transmitted signal generating portion 301t and then subjected to the ZF-DPC coding in a ZF-DPC coding portion 302t. Thereafter, the transmitted data is subjected to a multiplexing in a signal multiplexing portion 303t, added with the training signal in a training signal adding portion 304t, and transmitted after being inputted to an analog front end 309t. The received signal is inputted from the analog front end 309t to a data reconstructing portion 308t and a training signal extracting portion 316t. Received data is obtained from the data reconstructing portion 308t. In the training signal extracting portion 316t, the training signal is extracted from the received signal and a channel matrix is obtained in a channel matrix estimating portion 317t based on a result of the extraction and inputted to a receive weight matrix generating portion 306t. In the receive weight matrix generating portion 306t, a receive weight matrix is calculated from the channel matrix and inputted to a receive weight matrix updating portion 307t; and the channel matrix is updated in the channel matrix updating portion 307t and inputted to the ZF-DPC coding portion 302t. The transmitted signal generating portion 301t includes, in accordance with its communication method, a scrambler, an error correction coding portion, a QAM mapper, a transmission power control portion, etc. Also, the signal to be outputted is a signal vector having a component of a parallel transmission number in the multiuser MIMO wireless communication.
The signal multiplexing portion 303t is necessary in a multiplex communication method such as the OFDM method. In the OFDM method, an inverse discrete Fourier transfer is carried out in the signal multiplexing portion 303t. When a signal multiplexing such as single carrier transmission is not necessary, the signal multiplexing portion 303t is also not necessary.
While the training signal adding portion 304t is disposed immediately before the analog front end 309t, the disposing position of the training signal adding portion 304t is not particularly limited to this. The training signal adding portion 304t may be disposed before the signal multiplexing portion 303t or before the ZF-DPC coding portion 302t.
The analog front end 309t mutually converts signals and radio waves and transmits and receives the same, including an antenna, a filter, a signal amplifier, an up/down converter, an orthogonal modulator/demodulator, an AD converter, a DA converter, etc.
The data reconstructing portion 308t reconstructs data transmitted from the receiver to the transmitter, and includes, in accordance with its communication method, an OFDM splitting portion (discrete Fourier transfer), a MIMO receive weight matrix multiplying portion, a QAM demapper, an error correction decoding portion, a descrambler, etc.
In the following, checking from the procedure of the transmitter of the fourth embodiment will be made. A training signal is extracted by a training signal extracting portion 316t from a signal received in an analog front end 309t and a channel matrix is estimated by a channel matrix estimating portion 317t, thereby carrying out the signal reception of the block 202t and the estimation of channel matrix of the block 203t. In a receive weight matrix generating portion 306t and a channel matrix updating portion 307t, the determination of receive weight matrix 206t and the update of channel matrix of the block 207t are carried out, respectively. As a signal processed in a ZF-DPC coding portion 302t is transmitted from the analog front end 309t, the data transmission of the block 208t is carried out. A training signal adding portion 304t adds a training signal upon data transmission so that the channel matrix can be estimated by the receiver. In this manner, the transmission procedure described in the fourth embodiment can be carried out.
Transmitted data is converted to a transmitted signal in a transmitted signal generating portion 301r and then added with a training signal in a training signal adding portion 304r and transmitted by an analog front end 309r. A received signal is inputted from the analog front end 309r to a signal splitting portion 313r and a training signal extracting portion 316r. Multiplexed signals are split in the signal splitting portion and inputted to a receive weight matrix multiplying portion 314r. The receive weight matrix multiplying portion 314r multiplies a receive weight matrix obtained from a receive weight matrix generating portion 306r with the received signal, and a result of that is subjected to decoding in a DPC decoding portion 315r, thereby obtaining received data from a data reconstructing portion 308r. A training signal is extracted from the training signal extracting portion 316r, and a channel matrix is obtained by a channel matrix estimating portion 317r based on a result of the extraction and inputted to the receive weight matrix generating portion 306r. The receive weight matrix generating portion 306r calculates a receive weight matrix from the channel matrix and inputs the same to the receive weight matrix multiplying portion 314r.
The transmitted signal generating portion 301r includes, in accordance with its communication method, a scrambler, an error correction coding portion, a QAM mapper, a transmission power control portion, a MIMO transmission weight multiplying portion, an OFDM multiplexing portion (inverse discrete Fourier transfer), etc.
The signal splitting portion 313r is necessary in a multiplex communication method such as the OFDM method. In the OFDM method, a discrete Fourier transfer is performed in the signal splitting portion 313r. When a signal multiplexing such as single carrier transmission is not necessary, the signal splitting portion 313r is also not necessary.
In the DPC decoding portion 315r, a decoding processing corresponding to DPC coding is performed. Specifically, when THP is used in the DPC coding, a residue calculation is performed in the DPC decoding portion 315r.
The data decoding portion 308r includes, in accordance with its communication method, a QAM demapper, an error correction decoding portion, a descrambler, etc.
The analog front end 309r mutually converts signals and radio waves and transmits and receives the same, including an antenna, a filter, a signal amplifier, an up/down converter, an orthogonal modulator/demodulator, an AD converter, a DA converter, etc.
In the following, checking from the procedure of the receiver of the fourth embodiment will be made. By transmitting the training signal added by the training signal adding portion 304r, the transmission of training signal of the block 201r is carried out.
As the signal is received by the analog front end 309r, the data transmission of the block 209r is carried out. A training signal is extracted in the training signal extracting portion 316r and a channel matrix is estimated by the channel matrix estimating portion 317r, thereby carrying out the estimation of channel matrix of the block 203r. The determination of receive weight matrix of the block 206r is carried out in the receive weight matrix generating portion 306r. In a receive weight matrix multiplying portion 314r and a DPC decoding portion 315r, the multiplying of receive weight matrix of the block 210r and the DPC decoding of the block 211r are carried out, respectively. In this manner, the receiving procedure described in the fourth embodiment can be carried out.
According to the present invention, a computing amount regarding the multiuser MIMO method can be reduced. Accordingly, a reduction of circuit size and shortening of delay to data transmission can be achieved.
Also, achievement of a communication speed close to a theoretical limitation.
| Filing Document | Filing Date | Country | Kind | 371c Date |
|---|---|---|---|---|
| PCT/JP2008/069540 | 10/28/2008 | WO | 00 | 6/22/2011 |
