JP2006222872A - Space multiplex signal detection circuit - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a space multiplex signal detection circuit capable of obtaining an excellent error rate characteristic by suppressing an increase in the circuit scale if when a modulation multivalued number is particularly increased in the case of receiving a signal subjected to space multiplexing to detect the signal. <P>SOLUTION: A transmission signal candidate narrow-down circuit 108 refers to a proximity signal point data table 109 and stepwise narrows down the number of a plurality of transmission signal candidates to a prescribed number of the candidates. The proximity signal point data table 109 stores the cross-reference between signal points closer on a transmission constellation to signal points of transmission signals of each of transmission systems estimated by a transmission signal estimate circuit 107. A maximum likelihood estimate circuit 113 receives candidates of the transmission signal sequences and metrics corresponding to the candidates and outputs a transmission signal sequence whose metric value is least as a final transmission signal sequence. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a demodulator that detects a signal from a spatially multiplexed signal in a digital wireless communication system, and more particularly to spatial multiplexing that can be applied to an OFDM (Orthogonal frequency division multiplex) modulation scheme among the multicarrier modulation schemes. The present invention relates to a signal detection circuit.

  The multicarrier modulation scheme is a wireless transmission scheme that transmits information using a plurality of subcarriers. The input data signal is modulated by QPSK (Quadrature phase shift keying) or the like for each subcarrier. In this multicarrier modulation method, the orthogonal multicarrier modulation method in which the frequencies of the subcarriers are orthogonal is also called orthogonal frequency division multiplexing (OFDM), and wireless communication in which multipath propagation is a problem. Widely applied in the system.

  OFDM is a modulation scheme suitable for high-speed transmission with features that are not easily affected by multipath. However, in order to further improve the transmission rate, transmission is performed on the same frequency using multiple transmission antennas. It is conceivable to perform spatial multiplexing transmission using a spatially multiplexed signal. This spatial multiplexing transmission is called SDM (Space division multiplexing) transmission or MIMO (Multi input multi output) transmission because of the characteristics of the transmission path of the spatial multiplexing transmission system using a plurality of transmission antennas in the transceiver.

  FIG. 17 is a block diagram for explaining the spatial multiplexing transmission. In FIG. 17, four antennas are provided on each of the transmission side and the reception side. When performing spatial multiplexing transmission, there is no need to increase the frequency bandwidth. For example, when spatial multiplexing transmission is applied to OFDM, it is necessary to transmit the OFDM symbols in synchronization with the transmission timing. When performing the spatial multiplexing transmission, if it is desired to double the transmission rate, transmission is performed simultaneously from two antennas. When it is desired to triple the transmission, the transmission is performed simultaneously using three antennas. Of course, it is also possible to perform transmission from one transmission antenna without multiplexing.

Furthermore, the number of transmission systems for transmission and the number of antennas on the data reception side need not be the same as the number of antennas used on the transmission side, and these values are determined according to system requirements. Here, the transmission antenna is generalized as N and the reception antenna as M. The received signal Y can be expressed as Equation (1) using an M × N transfer function H of the propagation path using a matrix representation in the frequency domain. Here, a case where M ≧ N is shown as an example. Matrix elements such as h _MN in Equation (1) indicate transfer functions between the transmitting and receiving antennas. Here, W represents an M-dimensional noise M × 1 column vector.

  In order to detect the transmission signal of each transmission system from the spatially multiplexed reception signal represented by Equation (1) in the demodulator, a ZF (Zero-forcing) method, an MMSE (Minimum mean square error) method, an OSD (Ordered successive) detection) method, maximum likelihood detection (MLD) method, and the like. Among them, the error rate characteristic of the MLD method is the best. From a different perspective, the fact that the error rate characteristics are excellent means that communication with a low CNR is possible, which contributes to the expansion of the communication area. Further, when MLD is used, demodulation is possible even if reception is performed with one receiving antenna. Further, when a plurality of antennas are used on the reception side, there is a feature that a reception diversity effect can be obtained.

  Here, a signal detection method for performing signal detection from a spatially multiplexed signal is sometimes called signal separation or interference canceller, but the essence is that a signal transmitted from a multiplexed signal for each transmission system is used. Is to detect. Signal detection when the MLD method is used is realized as follows. Here, FIG. 18 is a conceptual diagram for explaining a conventional MLD system. The MLD scheme is not a scheme that can be applied only to the multicarrier modulation scheme, but is a technique that can be generally applied to single carrier transmission. Here, the case where it applies to the subcarrier signal after multicarrier demodulation is shown as an example. In the MLD method, signal detection is performed on the subcarrier signal, which is the FFT circuit output signal, using the channel estimation result. The principle of MLD signal detection is shown in Equation (2).

  However,

Indicates the kth transmission signal sequence candidate,

Indicates the i-row vector of the M × N-MIMO matrix of the estimated channel matrix obtained after channel estimation. In Equation (2), maximum likelihood estimation is performed using a metric μ which is one scale. In the MLD method, the signal point distance between the replica signal and the reception signal is calculated for each reception system. Thereafter, the signal point distances obtained in the respective receiving systems are added, and maximum likelihood estimation is performed using the result. Of course, signal detection by the MLD method is possible even with one receiving system, and in this case, only the result of one receiving system is used. Among the added signal point distances, the one with the shortest signal point distance is selected as the most probable combination of transmission signals. And the transmission signal of the signal of this combination is output as a signal detection estimation result. As described above, since the MLD method performs a full search for all combinations of replicas of transmission signals, the error rate characteristics are improved and excellent communication can be realized.

  On the other hand, there are the following examination examples. FIG. 19 is a block diagram showing a configuration of a conventional spatial multiplexing signal detection circuit (see, for example, Non-Patent Document 1). The operation will be described below. FIG. 19 shows a case where the demodulation circuit demodulates transmission signals transmitted from N transmission systems using the M reception system.

  The received signal S1 is multicarrier demodulated by the M FFT circuits 1 and outputs a subcarrier signal S2. The subcarrier signal S2 is input to the channel estimation circuit 2, the channel distortion of the propagation path is estimated from the subcarrier signal S2, and the channel estimation signal S3 is output. On the other hand, the transmission signal generation circuit 3 outputs an m-value transmission signal having a modulation multi-level number of m. The replica signal generation circuit 4 performs complex multiplication of the m-value transmission signal S4 and the channel estimation signal S3 to generate a replica signal S5 of the reception signal. The complex multiplier specifically required here is obtained as follows.

Considering transmission of m values-transmission signals for N transmission systems, there are m ^N possible transmission signal point candidates. Since it is necessary to perform complex multiplication for each transmission system to generate each replica, m ^N × N are required. Further, when processing is performed for a plurality of reception systems on the reception side, finally, M × m ^N × N complex multipliers are required. Next, the signal point distance calculation circuit 5 performs signal point distance calculation between the subcarrier signal S2 in each reception system and the replica signal S5 in each reception system. This signal point distance calculation circuit calculates the signal point distance in each reception system as shown in Equation (2) and adds the values of each reception system to calculate the final metric signal S6. Further, the number of times of the distance calculation can be specifically calculated as follows.

The total number of replica signals outputted from the replica signal generation circuit S4 is a m ^N pieces. Since the signal point distance in each receiving system is calculated for this replica signal, the total number of distance calculations is M × m ^N. This metric signal S6 is input to the full-search maximum likelihood estimation circuit 6 to perform maximum likelihood estimation. Signal detection is performed on the assumption that the replica signal having the smallest metric, that is, the distance between signal points among the input signals, is the most likely combination of transmission signals, and N transmission signals corresponding thereto are output.

As described above, the conventional spatial multiplexing signal detection circuit shown in FIG. 19 performs signal detection by maximum likelihood estimation of a spatially multiplexed signal. First, the signal point distance between the received signal and the replica signal is calculated for each reception system, and the replica signal indicating the distance between the signal points, which is the smallest metric for this result, is selected, and the signal constituting this replica signal Is detected.
A. van Zelst, R. van Nee and GA Awater, “Space division multiplexing (SDM) for OFDM systems,” Proc. Of VTC 2000-Spring, pp. 1070-1074.

  Currently, backward compatibility with the IEEE 802.11g standard in which the operating frequency band is changed to the 2.4 GHz band using signals in the same wireless section as the IEEE 802.11a standard or IEEE 802.11a standard of the wireless LAN transmission system. In order to realize high-quality and a high transmission rate, studies on a wireless LAN using spatial multiplexing transmission are in progress. For example, in the task group (TG) n in the IEEE 802 committee, studies aiming at application of spatial multiplexing transmission of 2 × 2 or more have been started.

  In a wireless LAN that realizes this wireless backward compatibility and uses spatial multiplexing transmission, the signal conforming to IEEE 802.11a is an OFDM signal before being multiplexed and corresponds to an existing system. An example of a packet format that takes this backward compatibility into consideration is shown in FIG. FIG. 21 shows a packet format in a two-dimensional representation using a frequency axis-time axis having a common part with the IEEE 802.11a signal.

  In these packets, in order to realize backward compatibility, a preamble signal that can be demodulated by the existing system is transmitted at the beginning of the packet, and then an existing system signal indicating the packet information is transmitted as necessary. Is done. However, since the data portion is spatially multiplexed and transmitted, application of a high-performance spatial multiplexing signal detection method is very important for improving communication quality. Furthermore, even when all transmitted packets are transmitted as spatially multiplexed signals, it is naturally important to apply a high-performance signal detection method. As a technique for realizing high-performance signal detection in this spatial multiplexing transmission, it is conceivable to apply the MLD method having excellent error rate characteristics.

Further, in order to realize further increase in transmission speed, 64QAM (m = 64) having a large modulation multi-value is often applied to subcarrier modulation of a signal before being multiplexed. When such a modulation multi-level number is large, the conventional spatial multiplexing signal detection circuit increases the circuit scale because a full search of the replica signal and the received signal is performed as shown in Equation (2). If the number of transmission systems is N, the number of reception systems at the demodulator is M, and the number of signals to be transmitted is m, the number of complex multipliers required at the time of replica generation is only M × m ^N × N. Thus, the circuit scale increases. In particular, when a modulation scheme with a large number of modulation multi-values is applied to increase the speed, the circuit scale increases exponentially, so the effect is more serious.

  Further, specifically, the complex multiplier circuit includes four real number multipliers having a large circuit scale and two adders. Therefore, using many complex multiplication circuits means that the circuit scale increases. In particular, in a wireless LAN that frequently uses a PCMCIA card or the like, there is a limit to the circuit scale allowed for the PCMCIA card.

Further, the number of distance calculations required for signal point distance calculation in a circuit according to the prior art is M × m ^N under the same parameters as described above, which directly leads to an increase in circuit scale. At this time, two real number multipliers and three real number adders are required for one distance calculation, and since the total number of distance calculations is large, an increase in circuit scale is inevitable.

  Furthermore, power consumption increases in proportion to the increase in circuit scale. In particular, when considering use in a wireless LAN or the like where it is difficult to use an external power source, there is a problem that an increase in power consumption accelerates battery consumption.

  The present invention has been made in consideration of such circumstances, and the object thereof is to solve these problems, and in particular, when receiving a spatially multiplexed signal and performing signal detection, the modulation multi-value number It is an object of the present invention to provide a spatial multiplexing signal detection circuit that can suppress an increase in circuit scale and realize an excellent error rate characteristic even when the signal is increased.

In order to solve the above-described problems, the present invention provides a channel estimation unit that receives a plurality of received signals and estimates channel distortion of a channel between transmission and reception antennas, and a channel estimated by the channel estimation unit. Estimated channel matrix generating means for generating an estimated channel matrix having distortion as a component, and QR decomposition means for performing unitary and triangulation (QR) decomposition operations on the estimated channel matrix generated by the estimated channel matrix generating means; , a complex conjugate transpose operation means for performing a Q ^H is the complex conjugate transpose operation with respect to the unitary matrix Q obtained by the QR decomposition unit, by using the calculation result by the complex conjugate transpose operation unit, constituting the plurality of received signals a matrix multiplication means for performing a matrix multiplication of the Q ^H × r to the received column vector signal r whose elements, the matrix multiplication means A multiplication result storage means for storing a calculation result of the transmission system; an output signal of each transmission system obtained from the multiplication result storage means; a matrix element signal of an upper triangular matrix R which is another output signal of the QR decomposition means; And a transmission signal estimation means for estimating a transmission signal of each desired transmission system using an interference signal which is a signal other than the desired transmission system, and a transmission signal and transmission of each transmission system estimated by the transmission signal estimation means A proximity signal point data table that stores correspondences between signal points with a short distance between signal points on the constellation as a database table, and a predetermined number of transmission signal estimation means estimates by referring to the proximity signal point data table A signal that selects a plurality of estimated transmission signals whose signal point distances are close to each other on the transmission constellation as a transmission signal of each transmission system Corresponding candidate narrowing means, a plurality of estimated transmission signals of each transmission system selected by the signal candidate narrowing means, and information of other transmission systems used to estimate the estimated transmission signal of the transmission system Transmission signal sequence storage means for storing each estimated transmission signal sequence, an estimated transmission signal of each estimated transmission signal sequence stored in the signal sequence storage means, and an upper triangle which is another output signal of the QR decomposition means Based on matrix element signals of the matrix, interference signal generation means for generating an interference signal that is a signal other than the desired transmission system, and each candidate of each estimated transmission signal sequence stored in the transmission signal sequence storage means On the other hand, transmission signal candidate metric calculation means for calculating a metric that is a signal point distance between the estimated reception signal and the actual reception signal when the candidate transmission signal is transmitted; The smallest metric is selected from all the metrics corresponding to each estimated transmission signal sequence calculated by the transmission signal candidate metric calculation means and stored in the transmission signal sequence storage means, and the estimated transmission signal sequence corresponding to the metric And a maximum likelihood estimator that outputs the signal as a finally estimated transmission signal sequence.

  According to the present invention, in the above invention, the QR decomposition means includes a unitary matrix Q and an upper triangular matrix R such that, among the elements of the generated upper triangular matrix R, elements having the same row number and column number are real numbers. It is characterized by selecting.

  According to the present invention, in the above invention, the transmission signal estimating means is p-th (1 ≦ p ≦ N: p is an integer) when the number of multiplexed transmission systems is N (N is an integer of 2 or more). ) In the transmission system, interference cancellation signal calculation for obtaining an interference cancellation signal obtained by subtracting the interference signal output from the interference signal generation means from the signal that is the element in the p-th row of the vector output from the matrix multiplication means And a classifying process for performing realization and imaginary part of the interference cancellation signal, which is an output from the interference cancellation signal calculating means, depending on the value of the p-th row and the p-th column of the upper triangular matrix R And a transmission signal approximate estimation unit that obtains an approximate value of the estimated transmission signal based on a table that manages the correspondence between the class value and the approximate value of the transmission signal in the classifying unit, and outputs the estimated transmission signal as the estimated transmission signal; To have And butterflies.

  According to the present invention, in the above invention, the interference signal generation unit is configured to multiply each non-zero element of the upper triangular matrix R by a signal that can be transmitted by a transmission station, and a multiplication result by the multiplication unit. And replica storage means for storing, and addition means for acquiring an interference signal, which is a signal other than a desired transmission system, by adding multiplication results stored in the replica storage means.

  In the present invention according to the above invention, the transmission signal candidate metric calculation means is stored in the transmission signal sequence storage means when the multiplexed transmission system is N (N is an integer of 2 or more). A set of transmission signal candidates from the p-th (1 ≦ p ≦ N: p is an integer) to the N-th transmission system is read, and for each transmission signal candidate group, the p-th matrix element of the upper triangular matrix R is read. Partial metric calculation means for calculating a partial metric representing the difference between the signal obtained by multiplying the signal of the transmission system corresponding to each element of the row and the value of the p-th row of the vector signal stored in the multiplication result storage means And, for each combination of transmission signal candidates from the first to the Nth transmission system, add the partial metric values obtained by the partial metric calculation means in the first to Nth transmission sequences, Metrics for transmission system Characterized by comprising a metric calculating means for calculating a click.

  According to the present invention, in the above invention, ranking means for determining superiority or inferiority of a reception state of each transmission system based on the estimated channel matrix output from the estimated channel matrix generation means, and the reception determined by the ranking means Based on the superiority or inferiority of the state, the estimated channel matrix rearranging means for changing the order of each row of the estimated channel matrix, and the estimated channel matrix rearranging means for the transmission signal sequence output from the maximum likelihood estimating means And a reverse switching means for performing a switching process opposite to the switching process of each row to be performed.

  According to the present invention, in the above invention, the estimated channel matrix rearranging unit performs rearrangement such that a transmission system having a better reception state is superior to the reception state determined by the ranking unit. It is characterized by performing.

  According to the present invention, in the above invention, the signal candidate narrowing-down means, when the multiplexed transmission system is N (N is an integer of 2 or more), the second to Nth transmission systems of the upper triangular matrix R The number of transmission signal candidates is set such that the larger the absolute value of the diagonal component of the is, the smaller or the smaller the number of candidates for narrowing down.

  The present invention, in the above invention, comprises multicarrier demodulation means for performing multicarrier demodulation on the plurality of received signals and outputting a subcarrier signal, wherein the subcarrier signal is the channel estimation means and the matrix multiplication means. And an input signal.

According to this invention, a plurality of received signals are input, channel distortion of the channel between the transmission and reception antennas is estimated by the channel estimation means, and the estimated channel distortion is calculated by the estimation channel matrix generation means. An estimated channel matrix is generated, unitary triangulation (QR) decomposition operation is performed on the estimated channel matrix by the QR decomposition unit, and complex conjugate transpose operation is performed on the unitary matrix Q by the complex conjugate transpose operation unit. perform certain Q ^H, the matrix multiplication means, by using a complex conjugate transpose operation result, performs a matrix multiplication operation Q ^H × r to the received column vector signal r to components a plurality of received signals, the matrix The calculation result of each transmission system of the multiplication means is stored in the multiplication result storage means, and the output signal of each transmission system, the matrix element signal of the upper triangular matrix R by the transmission signal estimation means The transmission signal of each desired transmission system is estimated using an interference signal that is a signal other than the desired transmission system, and a signal candidate narrowing means estimates the transmission signal of each transmission system and the signal on the transmission constellation. On the transmission constellation and the transmission signal of each transmission system estimated by the transmission signal estimation means by a predetermined number by referring to the proximity signal point data table storing the correspondence with the signal points having a short point distance A plurality of estimated transmission signals with short signal point distances are selected as candidates, a plurality of estimated transmission signals of each selected transmission system, and other transmission systems used to estimate the estimated transmission signal of the transmission system Is stored in the transmission signal sequence storage means as each estimated transmission signal sequence, and the estimated transmission signal and the upper triangular matrix of each estimated transmission signal sequence are stored by the interference signal generation means. An interference signal that is a signal other than the desired transmission system is generated based on the matrix element signal, and the transmission signal candidate metric calculation means generates the candidate transmission signal for each candidate of the estimated transmission signal sequence. The metric that is the signal point distance between the estimated received signal and the actual received signal when the signal is transmitted is calculated, and the estimated likelihood signal corresponding to each estimated transmitted signal sequence stored in the transmitted signal sequence storage unit is calculated by the maximum likelihood estimating unit. The smallest metric is selected from all the metrics, and the estimated transmission signal sequence corresponding to the metric is output as the finally estimated transmission signal sequence. Therefore, in the conventional MLD, metrics are calculated for all selectable patterns. However, in the present invention, the number of metrics to be compared is finally reduced, and candidates are narrowed down step by step. In addition, by using the proximity signal point data table, it is possible to omit an actual signal point distance comparison process and to realize an effective narrowing down.

  Further, according to the present invention, the unitary matrix Q and the upper triangular matrix R are set so that the elements having the same row number and column number among the elements of the upper triangular matrix R generated by the QR decomposition means are real numbers. select. Therefore, since the diagonal term of the upper triangular matrix R is a real number, the division processing in the QR decomposition functions to replace the division by complex numbers with the division by real numbers. As a result, the division can be replaced with a process that does not rotate the phase of a complex value. This has the advantage that the circuit scale can be reduced by simply replacing the complex divider with a real divider.

  Further, according to the present invention, in the transmission signal estimation means, when the number of transmission systems multiplexed by the interference cancellation signal calculation means is N (N is an integer of 2 or more), the pth (1 ≦ p ≦ N: p is an integer) In the transmission system, an interference cancellation signal obtained by subtracting the interference signal from the signal that is the element in the p-th row of the vector output from the matrix multiplication means is obtained, and the interference cancellation signal is obtained by the classifying means. The real number part and the imaginary number part of the above are subjected to a classifying process depending on the value of the p-th row and the p-th column of the upper triangular matrix R, and correspondence between the class value and the approximate value of the transmission signal is performed by the transmission signal approximation estimating means Based on a table for managing the estimated transmission signal, an approximate value of the estimated transmission signal is obtained and output as the estimated transmission signal. Therefore, the division of a complex number by a real number means that the phase component of the complex number is preserved and the scale of the real number component and the imaginary number component is simply converted isotropically. Therefore, if classifying is performed with a step size depending on the diagonal component of the upper triangular matrix R, it is possible to perform a hard decision or a rough soft decision equivalently without performing an actual division process. As an advantage, the circuit scale can be greatly reduced.

  Further, according to the present invention, in the interference signal generation means, the multiplication means multiplies each non-zero element of the upper triangular matrix R by a signal that can be transmitted by the transmitting station, and the multiplication result is stored in replica storage means. The interference signal which is a signal other than the desired transmission system is obtained by addition of the multiplication results stored in the replica storage means. Therefore, instead of performing sequential multiplication processing each time, the value used for the calculation is pre-multiplied and stored in the memory, and when the multiplication is necessary, the result is read from the memory, so that the necessary multiplication circuit The advantage is that the number can be significantly reduced.

  Further, according to the present invention, in the transmission signal candidate metric calculation means, when the transmission system multiplexed by the partial metric calculation means is N (N is an integer of 2 or more), the transmission signal sequence storage means Are read out from the p th (1 ≦ p ≦ N: p is an integer) to the N th transmission system, and for each transmission signal candidate set, the upper triangular matrix R A partial metric representing the magnitude of the difference between the signal obtained by multiplying the signal of the transmission system corresponding to each element of the p-th row of the matrix element and the value of the p-th row of the vector signal stored in the multiplication result storage means is calculated. The value of the partial metric obtained by the partial metric computing means in the first to Nth transmission sequences for each combination of transmission signal candidates from the first to Nth transmission systems by the metric computing means. Add Calculating the metric for all transmission systems Te. Therefore, the calculation amount can be reduced by calculating as a partial metric for each transmission system and obtaining a desired metric as a total compared to the conventional method in which all metric calculations must be performed individually. Is obtained.

  Further, according to the present invention, the ranking means determines the superiority or inferiority of the reception state of each transmission system based on the estimated channel matrix output from the estimated channel matrix generating means, and the estimated channel matrix rearranging means Based on the superiority or inferiority of the reception state determined by the ranking unit, the order of each row of the estimated channel matrix is switched, and the estimated channel is transmitted to the transmission signal sequence output from the maximum likelihood estimating unit by the reverse switching unit. An exchange process opposite to the exchange process of each row performed by the matrix rearranging means is performed. Therefore, it is possible to narrow down the signal candidates in an efficient order, and as a result, there is an advantage that the processing can be limited to a smaller number of candidates if the characteristics are the same.

  Further, according to the present invention, the estimated channel matrix rearranging unit performs rearrangement such that a transmission system having a better reception state in the superiority or inferiority of the reception state determined by the ranking unit is arranged in the lower stage. Do. Therefore, it is possible to determine a transmission system with high reliability first, and it is possible to obtain an advantage that the SNIR value of the transmission system with low reliability can be increased equivalently.

  According to the present invention, when the number of transmission systems multiplexed by the signal candidate narrowing means is N (N is an integer of 2 or more), the second to Nth transmission systems of the upper triangular matrix R The number of transmission signal candidates is set so that the larger the absolute value of the diagonal component is, the smaller or smaller the number of candidates for narrowing down. Therefore, if the gain of the received signal is high, there is an advantage that the number of candidates for each signal system can be narrowed with a relatively small number of candidates.

  Further, according to the present invention, the multicarrier demodulation means performs multicarrier demodulation on the plurality of received signals, outputs a subcarrier signal, and the subcarrier signal is output from the channel estimation means and the matrix multiplication means. Input signal. Therefore, the present invention can be applied to an OFDM modulation scheme that requires multicarrier demodulation, and an advantage that an excellent error rate characteristic can be realized while reducing the circuit scale.

  Hereinafter, a spatial multiplexing signal detection circuit according to an embodiment of the present invention will be described with reference to the drawings.

A. First, the basic principle of the present invention will be described. The present invention realizes a spatial multiplexing signal detection circuit that realizes an excellent error rate characteristic equivalent to that of the MLD method while suppressing an increase in circuit scale, which is a problem in the prior art. The present invention can be applied to both a multicarrier modulation system and a single carrier modulation system. When applied to a multicarrier modulation scheme, it is applied to a subcarrier signal after multicarrier demodulation.

  In the present invention, first, the channel estimation of the transfer function between the transmitting and receiving antennas is performed using the MIMO preamble signal of the received packet signal (channel estimation means). And the estimation channel matrix H is produced | generated using the channel estimation result (estimation channel matrix production | generation means). Next, as shown in Equation (5), unitary and triangulation (QR) decomposition of this estimated channel matrix is performed (QR decomposition means). However, in Formula (5), the formula modification which applied QR decomposition | disassembly with respect to an estimation channel matrix is shown for simplification of description. The following is a case where M ≧ N. The unitary matrix Q shown in Equation (6) is an M × M orthonormal matrix, and the upper triangular matrix R shown in Equation (7) is M This is a × N matrix, which is composed of an upper triangular matrix of N rows and N columns and a zero matrix of (MN) rows and N columns. Naturally, when M = N, the upper triangular matrix R is configured as an upper triangular matrix R having no 0 matrix portion.

  However,

  Alternatively, when the unitary matrix Q is defined as an M × N matrix as another expression of Equation (5), the upper triangular matrix R is QR-decomposed as an N × N matrix.

  As a calculation method used for this QR decomposition, naturally, the application of the Gram-Schmidt orthogonalization method can be considered.

  Furthermore, the essence of the QR decomposition is to divide the transfer function matrix H into a unitary matrix Q and an upper triangular matrix R, and rotate each matrix element. As a result, the upper triangular matrix is sequentially generated. Of course, it is possible to apply the Givens method and the Householder method in which “0” elements of the upper triangular matrix are added in units of rows and columns.

  In general, the diagonal component (the component having the same row number and column number) of the upper triangular matrix R in the QR decomposition is not necessarily a real number, but by appropriately selecting the unitary transformation Q, the upper triangular matrix R It is possible to select the diagonal component of R to be a real number. In this way, by selecting the diagonal component to be a real number, it is possible to simplify the transmission signal estimation process described later.

  Furthermore, in the present invention, a complex conjugate operation of each matrix element is performed on the unitary matrix Q, and then a transpose operation process of the matrix is performed (complex conjugate transpose operation means). This complex conjugate transposition may be called by various names such as Hermitian transpose and Hermitian conjugate, but the essence of the operation here is to perform complex conjugate operation on each element of unitary matrix and generate transpose matrix of that matrix That is.

Next, the multiplication from the left side of Q ^H is the complex conjugate transpose matrix of the resulting unitary matrix by the QR decomposition on the received signal Y (matrix multiplication means). By performing this process, the received signal can be simply shown as a multiplication of the upper triangular matrix and the transmission signal, as shown in Equation (8).

Further, as a feature in the case of this matrix multiplication, it is a feature of the present invention that noise enhancement is not generated in QR decomposition because each column vector is multiplied by a unitary matrix Q ^H that is orthonormal.

It is also a further feature of the present invention to noise enhancement due to the multiplication of the Q ^H since the eigenvalues of the unitary matrix used in the QR decomposition are all 1 does not occur. In this way, a column vector Z = [z ₁ , z ₂ ,..., Z _M ] ^T which is an operation result is obtained, and each value is stored (multiplication result storage means). Here, the notation [] ^T for vectors and matrices means transposition.

Furthermore, in the present invention, after this unitary matrix multiplication, transmission signals are estimated for each transmission system in sequence. First, t _N that is a transmission signal of the transmission system corresponding to the non-zero bottom row of the upper triangular matrix is estimated. Next, the transmission signal of the transmission system corresponding to the N-1 row of the upper triangular matrix is estimated. In the present invention, a process of sequentially raising the layer of the transmission system to be repeatedly estimated while removing the interference component of the other transmission system is performed.

The specific procedure is as follows.
First, the interference signal of the other transmission system is subtracted from each element of the stored column vector Z, and the following division processing is performed using each matrix element of the desired transmission system of the upper triangular matrix R (transmission signal estimation means) ). However, in the case of the transmission system N which is the lowest layer, it is not necessary to consider the interference signal of the other transmission system (from the interference signal generating means). First, the case where calculation is performed for the N transmission systems will be described below.

The signal point t _N in the case of the transmission system N soft determines.

  Furthermore, naturally, not only the soft decision result but also the hard decision result can be used for transmission signal estimation.

_{Alternatively} , when r _NN is a real number, an approximate value of t _N can be obtained directly from the value of z _N (= r _NN × t _N ). For example, taking 16QAM as an example, constellations of 16 types of transmission signals that can be taken by 16QAM are as shown in FIG. As shown in the figure, the 16 types of transmission signals have a signal point distance of 2.0 intervals on both the I axis and the Q axis. Further, in FIG. 1 plots also to the example of t _N in Equation (10).

On the other hand, as shown in FIG. 2, the signal point distance is changed to be 2.0 × r _NN interval on both the I axis and the Q axis, and z _N in Expression (9) is plotted there. FIG. 1 and FIG. 2 are diagrams in which the scale is completely converted isotropically with respect to the I axis and the Q axis. Therefore, for example, as shown in FIG. 2 as shown in FIG. 3, if it is classified at a boundary indicated by dotted lines parallel to the I axis and the Q axis at intervals of 2.0 × r _NN and it is determined which block it belongs to, Equivalently, a hard decision can be made as a signal point included in the block (indicated by “◎” in the figure).

Also, as shown in FIG. 4, the representative point “◎” at the center of each block is classified as a step size smaller than 2.0 × _rNN and approximated to the soft decision result represented by Expression (10). May be used. Or as shown in FIG. 5, it is also possible to classify along each lattice point. In this case, the signal point plotted in this figure uses the representative point “「 ”at the center of each block connecting the grid points as an approximate value of the soft decision result. Of course, it is also possible to classify with a step size finer than fine 2.0 × _rNN while classifying along each lattice point.

Incidentally if the imaginary component of r _NN is not zero, division by r _NN Unlike changed merely isotropic scale, because it involves the rotation of the coordinate axes, grid points of FIGS. 1-5 is I-axis and Q The above processing could not be applied because it was not parallel to the axis or was distorted, and division of Expression (10) was indispensable.

  In the following, first, for simplification, the flow of processing is shown for the case of transmission signal estimation in the case of the transmission system N, and in the case of the N-1th, N-2th,. I'll show you.

Then, after the estimation of the transmission signal is performed, to narrow down the candidates of the transmitted signal points the estimated transmission signal t _N based. As an example, FIG. 6 shows an example when the estimation of the transmission signal is obtained by hard decision. For example, FIG. 6 shows a case where the number of candidates after narrowing down is nine. When a certain grid point is given as a hard decision point “x” of a transmission signal with respect to signal points arranged in a grid, the point is used as an argument, and the first proximity “「 ”between the grid point and the grid point and If managed as a database (proximity signal point data table) that gives a total of nine points of the second proximity “◯”, the narrowing process can be easily executed.

The number of candidates to be narrowed down is, for example, 5 if the first proximity is reached, and 13 if the third proximity is included. Similarly, other points can be set. Or, when classifying is performed as shown in FIG. 5, if only the first proximity from the representative point is selected, four points as shown in FIG. (Indicated by “表 記” in the figure). Also, if it is included up to the second proximity, it will be 12 points, and if it is included up to the third proximity, it will be 16 points. However, as shown in FIG. 8, there is a case where there is no adjacent point at an isotropic position in the end region on the constellation. In such a case, it is asymmetric with respect to t _N depending on the situation. The data table may be constructed so as to select a nearby point.

  By performing the above-described processing, various numbers of narrowing down can be selected. In the narrowing down, even when the approximate value of the soft decision shown in FIG. 4 is obtained, the narrowing result corresponding to the predetermined number of narrowing down may be provided for each representative point as a database.

  Here, the narrowing-down process described above obtains an approximate solution to the calculation represented by Equation (11) using the signal point distance between the signal point s of the m-valued transmitted multilevel signal and the estimated transmission signal. There is nothing else.

Here, hat [t _N (l _N )] represents a signal point s in which the value of ‖t _N −s‖ ² is l _Nth smallest among signal points s that can be selected as a transmission signal. The above hat corresponds to the symbol (^) in the mathematical formula, and hereinafter the same notation. In addition, here, the case where the square of the Euclidean distance is used as an example of the distance calculation is shown.

  Using the above processing, a predetermined number of transmission signal candidates in the Nth transmission system are selected from the one with the smaller signal point distance (transmission candidate narrowing means). For example, when the upper 32 signal candidates are selected as an example, a plurality of transmission signal candidates of the Nth transmission system are selected as shown in Equation (12).

  Then, the transmission signal of this transmission system is stored (transmission sequence addition storage means) and applied to the following metric calculation (interference signal generation means).

On the other hand, using the stored element signal z _N of the column vector Z and r _NN which is an element signal of the upper triangular matrix subjected to QR decomposition, the following partial metrics M are obtained for each of the 32 transmission signal candidates. _N (l _N ) is calculated as in Expression (13) (transmission signal candidate metric calculation means and partial metric calculation means).

  Here, as the signal point distance calculation and the partial metric used in the present invention, the above description shows the case where the square of the Euclidean distance is used. However, for the specific distance calculation, the Euclidean distance (of each element of the vector signal) is used. The square root of the sum of the square value of the real part and the square value of the imaginary part), the Manhattan distance (the sum of the absolute value of the real part and the absolute value of the imaginary part of each element of the vector signal), and the square of those distances Naturally, it can be realized by using various distance measures such as distance.

  Next, this partial metric is stored together with the combination information of the transmission signals of each transmission system. Further, it may be stored as a cumulative addition metric AM after being added to the previous metrics (metric calculation means). The above is the signal processing flow in the Nth transmission system, which is the lowest layer in the present invention.

Next, processing in the (N-1) th transmission system will be described. In order to estimate the transmission signal of the (N-1) th transmission system, an operation for generating an interference signal from the estimated signal of the Nth transmission system is required. When calculating for the N-1 transmission system, the interference signal is as follows for all combinations of the element signal r _{N−1, N} of the upper triangular matrix and the estimated transmission signal hat [t _N (l _N )]. Is calculated (interference signal generating means).

  The basic operation when detecting a signal of the N-1 transmission system can be considered as follows. The signal detection procedure of the N-1 transmission system is basically the same operation as the signal detection of the Nth transmission system. The difference is that the estimated signal of the other transmission system (output from the interference signal generating means) (Nth transmission system larger than N-1 in the case of the N-1th transmission system) is subtracted (transmission signal estimation). Means and transmission signal candidate metric calculation means).

Basically, as shown in Expression (12), when a plurality of estimated signals of the N-th transmission system are expressed as hat [t _N (l _N )], they are soft according to Expression (15) below. Determination calculation is performed (transmission signal estimation means).

Note that this estimate of the N-1 transmission signal of the transmission lines, will be performed for each l _N of the N transmission lines, estimates of the transmitted signals of the N-1 transmission system for l _N is t _{N −1} (l _N ). However, in practice, as in the case of the N transmission system, the soft decision approximate value obtained by the above-described grading process from the equation (16) is used without executing the equation (15), or the hard decision is performed. It is also possible to apply the results.

Next, after the transmission signal of the (N-1) th transmission system is estimated, the transmission signal point candidates are narrowed down based on the estimated transmission signal t _N-1 (l _N ). The transmission signal is narrowed down using the proximity signal point data table as in the case of the Nth transmission system. The narrowing-down process here is nothing but finding an approximate solution for the calculation shown in Equation (17).

Here, hat [t _N-1 (l _N−1 , l _N )] means that the transmission signal from the N-th transmission system is hat [t _N (l _N )] represents a signal point s in which the value of ‖t _N-1 (l _N ) -s ‖ ² is l _N-1 smallest.

Using the above processing, for each hat [t _N (l _N )], a predetermined number is selected as a plurality of transmission signal candidates in the (N−1) -th transmission system from the smaller signal distance (signal candidate narrowing means). . For example, when estimating the upper 16 transmission signal candidates, the estimation is shown as in Expression (18).

  Next, a partial metric is calculated using a plurality of estimated transmission signals of the N-1 transmission system obtained in this way (transmission signal candidate metric calculation means and partial metric calculation means).

Here, the partial metric for determining transmission signal candidates for the (N-1) th transmission system is the transmission signal hat [t _{N of} the Nth transmission system, which is assumed when selecting transmission signal candidates for the (N-1) th transmission system. (L _N )]. In other words, it is necessary to obtain a partial metric for each combination of l _N−1 and l _N. Then, this part metric and combination information (e.g., _{l N-1} and _{l N} of the transmission signal of each transmission _{system, hat [t N-1 (} l N-1, l N)] and hat _[t N (l _N )] and other information).

  Similarly, the above processing is continued with the N-2th transmission system, the N-3th transmission system,..., And finally the processing is executed up to the first transmission system.

Here, a major feature of the present invention is that there is little deterioration in characteristics even when the number of signal points considered as candidates is reduced as the estimated transmission system increases. Since the metric is calculated using the transmission signals estimated sequentially, the diversity effect is improved. In the estimation of the upper transmission signal, the transmission signal is estimated after canceling a signal from another transmission system. For higher-order signal estimation, estimation is not performed using only information on the transmission system. For example, for the estimation of the N-1 transmission system, a replica signal of the signal of the N transmission system is subtracted from the matrix multiplication result zN _{-1 of the N-1} transmission system. In this way, the diversity effect is generated because the transmission signal is estimated using the information of the two transmission systems N-1 and N. Therefore, the diversity effect can be obtained even if the number of transmission signal candidates narrowed down as candidates at the upper level is reduced. Of course, the present invention can be applied without decreasing the number of candidates as it goes up.

  The description of the decrease in the number of transmission signal candidates will be specifically shown. In the present invention, for example, in the case of 64QAM having four transmission systems, the number of narrowing down can be reduced by limiting to “32, 16, 8, 1”, and the number depends on the transmission signal point. It is not necessary to generate all the number of replica signals, and realization with a simple circuit scale is possible. The number of candidates for the first transmission system is set to 1.

  Finally, after processing up to the first transmission system is performed, a metric AM that is an accumulated value of the partial metrics so far is obtained (transmission signal candidate metric calculation means and metric calculation means).

  In the addition at this time, for example, the cumulative value of the partial metric up to the pth (1 ≦ p ≦ N) transmission system is a value obtained by adding the partial metric of the pth transmission system to the cumulative value of the partial metric up to the (p + 1) th transmission system. Therefore, it may be obtained sequentially as in Expression (21).

The intent of this recurrence formula is that when the transmission signal candidates of the p-th transmission system are selected, if the processing of Formula (19) and Formula (21) is sequentially performed, the formula is finally put together. It is not necessary to calculate the metric of (20), and the metric amount corresponding to the equation (21) can be sequentially stored in the memory.

  FIG. 9 shows a state in which all signals up to one transmission system are estimated and signal sequences of all transmission systems are calculated repeatedly in sequence. In this case, the number of transmission signal candidates in the N-th transmission system is 32, the number of transmission signal candidates in the (N-1) -th transmission system is 16, and finally the number of candidates in the highest transmission system is 1. Yes.

  After the series of processes described above, in the present invention, a final transmission signal sequence is estimated (maximum likelihood estimation means). A transmission signal sequence indicating the smallest value among the transmission signal sequence candidates (transmission signal sequence storage means) and the corresponding cumulative addition metric (transmission signal candidate metric calculation means) is output as a transmission signal sequence.

As described above, according to the present invention, it is possible to greatly reduce the generation of a huge amount of replica signals that is necessary in the prior art. In the present invention, the number of complex multiplications required in the prior art by M × m ^N × N is the maximum number of operations (Max [l ₁ ] × Max [l ₂ ] ×... × Max [l _N ]]. ) And can be significantly reduced. Here, Max [l _N ] is the number of transmission signal candidates to be narrowed down in the Nth transmission system. In the present invention, for example, in the previous example, when 64QAM with m = 64 is used in an N = 4 transmission system, operation is possible with a number of candidates of about 32 × 16 × 8 × 1. In this case, the number of complex multiplications can be significantly reduced from 4 × 4 × 64 ⁴ = 2684435456 (conventional method) to 4 × 4 × 32 × 16 × 8 × 1 = 65536.

  Further, in narrowing down transmission signal candidates in each transmission system, a single transmission signal that has been primarily estimated is referred to a database table to easily narrow down to a plurality of transmission signals (candidates). Selection).

  As a further configuration for improving the characteristics of the present invention, each channel estimation result is used to receive a received signal input from a receiving antenna, such as a received SNR (signal to noise ratio), SNIR (signal to noise and interference ratio), Alternatively, a ranking method or the like corresponding to the transmission signal system can be applied according to the received power. Specifically, for example, ranking is performed according to the estimated SNIR, SNR, and the like shown in Expression (22) to Expression (24), and the column vectors of the estimated channel matrix are replaced according to this ranking.

Here, _{W i} is receiving power of the i transmission system, the _{W i} in Equation (23) the estimated SNIR of the i transmission system, _{W i} is the estimated SNR of the i transmission system of equations (24) Equation (22) It is estimated that the larger these values are, the better the communication quality is. For example, it is possible to estimate a transmission signal from a transmission system having a high reception SNR by performing an operation of replacing column vectors of the following estimation channel matrix using A which is an N × N matrix. Here, when N = 4 and M = 4, the received SNR of each transmission system is in the order of “1, 2, 3, 4” in descending order.

  When this weighting is used, Equation (5) can be transformed into Equation (26) using the transformation matrix A.

The meaning of this equation (26) is that the column vectors are replaced by HA, and are arranged from the right in order of increasing SNR. In addition, rearrangement is performed from the bottom in order from the highest SNR of the transmission signal vector by A ^H × T. As described above, when ranking based on SNR is performed, QR decomposition can be performed using HA as an estimated channel matrix instead of H as shown in Expression (27).

  The subsequent calculation procedure is the same as that in the case where ranking is not used, and can be expressed by Equation (28).

  By using this ranking process, the estimation accuracy for narrowing down transmission signal candidates is improved. As a result, the entire estimated candidate transmission signal can be reduced, and the circuit scale can be reduced.

Here, an example is shown in which the SNRs are arranged in order from the bottom in the descending order of SNR, but other arrangement orders may be used. For example, to arrange from the formula (23) in the first line only a small signal system of most W _i with either equation (25), the other signal lines may be arranged such that optional.

Also, using either the equation (22) formula (24), after rearrangement according to W _i of each transmission system performs QR decomposition, the diagonal elements of the upper triangular matrix R after QR decomposition The number of candidates for narrowing down each transmission system may be selected in accordance with the absolute value of. This is because the diagonal component of the upper triangular matrix R that is a coefficient multiplied by the transmission signal in Equation (9) and Equation (16) is considered to correspond to the reception gain for each signal system. This is because a signal system having a large gain is considered to have a true transmission signal point relatively close to the estimated transmission signal represented by the equations (10) and (15).

  For example, in the previous example, the number of the fourth transmission system is 32, the third transmission system is 16, the second transmission system is 8, the first transmission system is 1, and so on. Assuming that the absolute values of the diagonal components of the second to fourth transmission systems are in the order of 3 → second → fourth, the third transmission system is 32, the second transmission system is 16, and the fourth transmission system. May be 8. At this time, the number of candidates for the first transmission system is 1. Thus, the optimal number of candidate points can be distributed by reducing the number of candidate points for signals considered to have a large reception gain and increasing the number of candidate points for signals considered to have a small gain.

  Note that the method for determining the number of candidates based on the size of the diagonal components of the upper triangular matrix R described above does not necessarily require ranking processing, and the number of candidates can be determined based on the size of the diagonal components of the upper triangular matrix R without performing ranking processing. You can decide. Further, when setting the number of candidates according to the ranking order, there is no necessity that all the numbers of candidates are different, for example, from the smallest absolute value of the diagonal components of the second to fourth transmission systems, the third → second → If it is the fourth order, there may be a plurality of the same number of candidates such as 32 for the third transmission system, 12 for the second transmission system, and 12 for the fourth transmission system. Of course, all the same candidates may be used except for the first transmission system.

When the ranking and rearrangement processing described above is performed, the transmission signal finally estimated by the maximum likelihood estimation is a signal after rearrangement in signal processing, and the actual transmission system It is not a corresponding transmission signal. Since the final estimated transmission signal is A ^H × T, in order to finally obtain the transmission signal T, the transformation matrix A is multiplied from the left of A ^H × T as shown in Equation (29). Then, rearrange the correct transmission signals and output the estimated transmission signal T (reverse switching means).

  As described above, the ranking method applied to the present invention is easily realized by a circuit that performs an operation of replacing the column vector of the transfer function matrix in accordance with the SNR of the received signal (ranking means).

B. First Embodiment Next, a first embodiment of the present invention will be described.
B-1. Configuration of First Embodiment FIG. 10 is a block diagram showing a configuration of a spatial multiplexing signal detection circuit according to the first embodiment of the present invention. The channel estimation circuit 101 estimates the distortion of the propagation path between the transmitting and receiving antennas using the MIMO preamble signal of the received packet signal S101 or the preamble signal of the existing system. The estimated channel matrix generation circuit 102 generates an estimated channel matrix H using the channel estimation result. The QR decomposition circuit 103 performs a QR decomposition operation on the estimated channel matrix, and decomposes it into a unitary matrix Q and an upper triangular matrix R. The complex conjugate transpose operation circuit 104 performs a complex conjugate operation of each matrix element on the unitary matrix Q and performs a transpose operation process of the matrix.

Matrix multiplication circuit 105 performs multiplication from the left side of Q ^H is the complex conjugate transpose matrix of the resulting unitary matrix by QR decomposition on the received packet signal S101. The multiplication result storage circuit 106 stores the calculation result of the matrix multiplication circuit 105 over one symbol time. The transmission signal estimation circuit 107 includes the element S107 of the Nth transmission system in the Nth row stored in the multiplication result storage circuit 106 and the Nth row of the upper triangular matrix R output separately from the QR decomposition circuit 103. A soft decision or a hard decision of the transmission signal of the Nth transmission system is estimated for the component S108 corresponding to the row vector, and is supplied to the transmission signal candidate narrowing circuit 108 as an estimation signal S110.

  The transmission signal candidate narrowing circuit 108 refers to the proximity signal point data table 109 and selects a predetermined number of transmission signal candidates S111 located in the vicinity of the input estimated signal S110. The proximity signal point data table 109 stores a correspondence between the transmission signal of each transmission system estimated by the transmission signal estimation circuit 107 and a signal point having a short distance between signal points on the transmission constellation. The transmission signal sequence storage circuit 110 stores the plurality of transmission signal candidates S111. The transmission signal candidate metric calculation circuit 111 obtains a metric S114 with respect to the estimated combinations of signals of all transmission systems.

  The interference signal generation circuit 112 determines the upper triangular matrix R according to the transmission signal sequence candidate S112 stored in the transmission signal sequence storage circuit 110 and the upper triangular matrix R information S108 separately output from the QR decomposition circuit 103. A component signal corresponding to the row vector of the (N-1) th row and a replica signal S113 of a transmission system larger than (N-1) transmission system are generated. The maximum likelihood estimation circuit 115 determines the transmission signal sequence indicating the smallest value among the transmission signal sequence candidates S112 stored in the transmission signal sequence storage circuit 111 and the metric S114 corresponding thereto as the final transmission signal. Output as series S116.

B-1. Operation of First Embodiment Next, the operation of the first embodiment described above will be described. Received signal S101 received by M receiving antennas is input to channel estimation circuit 101. The channel estimation circuit 101 estimates the propagation path distortion between the transmitting and receiving antennas using the MIMO preamble signal or the preamble signal of the existing system in the received signal. The estimated distortion signal S102 is input to the estimated channel matrix generation circuit 102 and output as the estimated channel matrix S103.

  The QR decomposition circuit 103 performs a QR decomposition operation on the estimated channel matrix S103 to perform decomposition into a unitary matrix Q and an upper triangular matrix R. The unitary matrix S104 is input to the complex conjugate transpose operation circuit 104 and output as a complex conjugate transpose matrix S105 of the unitary matrix. The matrix multiplication circuit 105 performs a process of multiplying the input received signal S101 by the complex conjugate transpose matrix S105 from the left side after converting the received signal into a column vector notation. The multiplication result S106 is stored in the multiplication result storage circuit 106 for one symbol time. Thereafter, the signal processing operation of the signal detection circuit shifts to signal detection.

  First, the transmission signal of the N transmission system which is the lowest layer is estimated. In the transmission signal estimation circuit 107, the element S107 of the N-th transmission system in the N-th row shown in Expression (8) stored in the multiplication result storage circuit 106, and the upper triangular matrix output separately from the QR decomposition circuit 103 The component S108 corresponding to the row vector of the Nth row of R is input. On the other hand, the soft decision or the hard decision of the transmission signal of the Nth transmission system is estimated using Expression (9) or Expression (10), and is input to the transmission signal candidate narrowing circuit 108 as the estimation signal S110. . The transmission signal candidate narrowing circuit 108 refers to the proximity signal point data table 109 and selects a predetermined number of transmission signal candidates S111 located in the vicinity of the input estimation signal S110. The multiple transmission signal candidates S111 are stored in the transmission signal sequence storage circuit 110.

  Next, in the signal detection process, the operation is shifted to the (N-1) th transmission system. For estimation of the (N-1) th transmission system, it is necessary to remove an interference signal of a transmission system larger than N-1, that is, the Nth transmission system. Transmission signal S112 of the Nth transmission system stored in transmission signal sequence storage circuit 110 is input to interference signal generation circuit 112. Further, information S108 of the upper triangular matrix R that is separately output from the QR decomposition circuit 103 is also input to the interference signal generation circuit 112. The interference signal generation circuit 112 generates a component signal corresponding to the row vector of the (N-1) th row of the upper triangular matrix R and a replica signal S113 of a transmission system larger than N-1 (that is, the Nth) transmission system, and estimates the transmission signal Input to the circuit 107. Thereafter, the transmission signal estimation circuit 107 performs signal estimation in the same manner as in the Nth transmission system. The main difference between the signal processing of the Nth transmission system and the signal processing of the (N-1) th transmission system in the transmission signal estimation circuit 107 is only the difference whether or not the replica signal S113 output from the interference signal generation circuit 112 is used. It is.

  In the transmission signal estimation circuit 107, the element S107 of the (N-1) th transmission system shown in the mathematical expression (8) stored in the multiplication result storage circuit 106 is changed to the Nth transmission system that is the output signal of the interference signal generation circuit 112. The corresponding replica S113 is subtracted, and the transmission signal of the (N-1) th transmission system is estimated using Equation (16) or Equation (15). As an estimation result of the soft decision or the hard decision of this transmission signal, an estimation signal S110 of the (N-1) th transmission system is output. In the transmission signal candidate narrowing-down circuit 108 to which the estimation signal S110 is input, a predetermined number in the vicinity of the estimated signal of the (N-1) th transmission system is referred to by referring to the proximity signal point data table 109 as before. Are selected as candidates, and the result is output to the transmission signal sequence storage circuit 110 as a plurality of transmission signal candidates S111.

The above-described processing regarding the (N-1) th transmission system is performed individually for each candidate in the Nth transmission system. That is, for each individual hat [t _N (l _N )], the interference signal generation circuit 112 generates an individual replica (interference signal) S113. For each individual interference signal, the transmission signal estimation circuit 107 An individual estimated signal S110 is output. The transmission signal candidate narrowing circuit 108 outputs candidates for each individual estimated signal S110. Therefore, for example, when the number of candidates for the Nth transmission system is 32 and the number of candidates for the N-1th transmission system is 16, the information stored in the transmission signal sequence storage circuit 110 is 32 types of l _N (or t _{N (l N))} and, _{l N-1} of 16 types for each _{l N} (or _{t N-1 (l N-} 1, l N)) , and the respective is 32 × 16 pieces of information transmitted as the total number It will be recorded as a series.

  Thereafter, similarly, processing is performed with the (N−2) -th transmission system and the (N−3) -th transmission system, and finally transmission signal candidates for the first transmission system are selected.

On the other hand, the transmission signal candidate metric calculation circuit 111 obtains the metric represented by the formula (20) for the estimated combinations of the signal candidates of all transmission systems. Partial metrics M ₁ (l ₁ , l ₂ ,..., L _N ), M ₂ (l ₂ , l ₃ ,..., L _N ), ... M _N (l _N ) in formula (20) are expressed by formula (13). , Given by equation (19) and the like. Since the metric as the cumulative value of the partial metrics is given by a recurrence formula like Formula (21), Formula (21) is sequentially executed for each signal narrowing process of each transmission system, and the last first The metric of Expression (20) may be obtained as a value after processing of the transmission system. Alternatively, the metric calculation may be finally performed on the combinations of all the estimated signal candidates of all transmission systems.

  After obtaining all the above-mentioned metrics, the transmission signal sequence S112 (a combination of transmission system signal candidates) recorded in the transmission signal sequence storage circuit 110 and the metric S114 of each transmission signal sequence are set to the maximum likelihood estimation circuit 113 as a set. Entered. The estimated transmission sequence S116 having the smallest metric is output.

C. Second Embodiment Next, a second embodiment of the present invention will be described.
In the second embodiment, in the QR decomposition circuit 104, the unitary matrix Q and the upper triangular matrix R are selected so that the diagonal components of the upper triangular matrix R are all real numbers. For example, if the Gram-Schmidt orthogonalization method is used as the QR decomposition algorithm, the diagonal terms of the upper triangular matrix R can be easily converted to real numbers. Alternatively, in an arbitrary unitary matrix Q and upper triangular matrix R, for example, a matrix Q ′ obtained by multiplying all elements of the p-th column (1 ≦ p ≦ N) of the unitary matrix Q by a complex number exp (θj) and an upper triangular matrix The product of the matrix R ′ obtained by multiplying all the elements in the p-th row of R by the complex number exp (−θj) is equal to the original matrix Q · R. After obtaining R, the above processing may be performed so as to cancel the complex phase of each diagonal component of the upper triangular matrix R. Here, j is an imaginary unit.

D. Third Embodiment Next, a third embodiment of the present invention will be described. The third embodiment relates to the transmission signal estimation circuit 107 in the spatial multiplexing signal detection circuit of the first embodiment described above.

D-1. Configuration of Third Embodiment FIG. 11 is a block diagram showing a configuration of a transmission signal estimation circuit 107 according to the third embodiment of the present invention. In the figure, the transmission signal estimation circuit 107 includes an interference cancellation signal calculation circuit 121, a classifying circuit 122, a class threshold generation circuit 123, and a transmission signal approximate estimation circuit 124. The interference cancellation signal calculation circuit 121 subtracts the interference signal S113 from the interference signal generation circuit 112 from the elements of the Nth transmission system in the Nth row from the multiplication result storage circuit 106 (information obtained by unitary conversion of the reception signal). Then, an amount corresponding to the left side of Expression (16) is acquired. The class threshold generation circuit 123 sets a classifying threshold in proportion to the diagonal component of the upper triangular matrix for each transmission system. As shown in FIGS. 3 to 5, the classifying circuit 122 determines which class it belongs to based on the threshold information generated by the class threshold generating circuit 123. The transmission signal approximate estimation circuit 124 converts the determination result information S123 into coordinate point information on a normal constellation as shown in FIG. 1, and outputs it to the transmission signal candidate narrowing circuit 108 as an estimation signal S110.

D-2. Operation of Third Embodiment Next, the operation of the third embodiment described above will be described. After the channel estimation and the QR decomposition, when the upper triangular matrix element S108 is input from the QR decomposition circuit 103 to the class threshold generation circuit 123, the class threshold generation circuit 123 performs the diagonal of the upper triangular matrix for each transmission system. Set the stratification threshold in proportion to the components. As shown in FIGS. 3 to 5, in the case of the Nth transmission system, the stratification threshold is a 2 × _rNN interval that is twice the diagonal component _rNN , or a step size of an integer. Set class boundaries. This is performed for all of the first to Nth transmission systems. The above processing is performed only once when data is received.

Thereafter, in the data processing, when the interference removal signal calculation circuit 121 receives the information S107 obtained by unitary conversion of the reception signal from the multiplication result storage circuit 106 and the interference signal S113 from the interference signal generation circuit 112, they are processed. Subtraction is performed to obtain an amount corresponding to the left side of Equation (16). This result signal S121 is input to the classifying circuit 122, and based on the threshold information generated by the class threshold generating circuit 123 as shown in FIGS. 3 to 5, it is determined to which class. The determination result information S123 is input to the transmission signal approximate estimation circuit 124, which is converted into coordinate point information on a normal constellation as shown in FIG. 1, and the estimated transmission signal t _p (1 ≦ p ≦ N). ) Information S110 to the transmission signal candidate narrowing circuit 108. In the case of the Nth transmission system, there is no signal input from the interference signal generation circuit 112 because there is no interference signal as shown in Equation (9).

E. Fourth Embodiment Next, a fourth embodiment of the present invention will be described. The fourth embodiment relates to the interference signal generation circuit 112 in the spatial multiplexing signal detection circuit of the first embodiment described above.

E-1. Configuration of Fourth Embodiment FIG. 12 is a block diagram showing the configuration of the interference signal generation circuit 112 according to the fourth embodiment of the present invention. In the figure, the interference signal generation circuit 112 includes a replica generation circuit 131, a transmission signal generation circuit 132, a replica storage circuit 133, and an interference replica addition circuit 134. The replica generation circuit 131 multiplies each non-zero element of the upper triangular matrix by all transmission signals. The transmission signal generation circuit 132 generates all transmission signal points that can be selected as transmission signals of a certain transmission system. The replica storage circuit 133 stores all the result information S132 generated by the replica generation circuit 131. The interference replica adding circuit 134 notifies the replica storage circuit 133 of information on the signal candidates of the determined transmission system and information on the number of the transmission system to be obtained at present (S133), and uses this information as an argument for the replica storage circuit The replica information S134 stored in 133 is acquired, added, and output to the transmission signal estimation circuit 107 or the transmission signal candidate metric calculation circuit 111 as an interference signal S113.

E-2. Operation of Fourth Embodiment Next, the operation of the fourth embodiment will be described. After the channel estimation and the QR decomposition, an upper triangular matrix element signal S108 is input from the QR decomposition circuit 103 to the replica generation circuit 131. The transmission signal generation circuit 132 generates all transmission signal points that can be selected as transmission signals of a certain transmission system, and inputs the transmission signal points to the replica generation circuit 131 as transmission signal information S131. The replica generation circuit 131 multiplies each element of the upper triangular matrix by all the transmission signals, and stores each result information S132 in the replica storage circuit 133.

  Thereafter, in the data processing, at the time of generating the interference signal, the interference replica adding circuit 134 uses the information of the signal candidate of the transmission system determined and the information of the number of the transmission system currently being obtained as notification information S133. The replica storage circuit 133 is notified, and the replica information S134 stored in the replica storage circuit 133 is acquired using this information as an argument. When there are a plurality of transmission signals as interference sources, replica information is individually acquired from the replica storage circuit 133 and added to obtain an interference signal. Finally, the interference signal S113 is output to the transmission signal estimation circuit 107 or the transmission signal candidate metric calculation circuit 111.

F. Fifth Embodiment Next, a fifth embodiment of the present invention will be described. The fifth embodiment relates to a transmission signal candidate metric calculation circuit 111 in the spatial multiplexing signal detection circuit of the first embodiment described above.

F-1. Configuration of Fifth Embodiment FIG. 13 is a block diagram showing a configuration of a transmission signal candidate metric calculation circuit 111 according to the fifth embodiment of the present invention. The transmission signal candidate metric calculation circuit 111 includes a partial metric calculation circuit 141, a recurrence equation calculation circuit 142, and a cumulative metric storage circuit 143. The partial metric calculation circuit 141 includes a transmission signal candidate S112 for each transmission system from the transmission signal sequence storage circuit 110, an interference signal S113 for the corresponding transmission signal candidate from the interference signal generation circuit 112, and a multiplication result storage circuit 106. The partial metric defined by Equation (13), Equation (19), etc. is calculated in accordance with the information S107 obtained by unitarily transforming the received signal stored in (1). For example, when Equation (21) is used as the recurrence formula, the cumulative metric storage circuit 143 uses the second term (1) for the signal S141 corresponding to the first term (partial metric of the p-th transmission system) on the right side. Signal S142 corresponding to (cumulative metric in the (p + 1) th transmission system) is stored. The recursive equation circuit 142 is obtained using the signal S142, and obtains a cumulative value from the calculation result S141. The calculation result is stored in the cumulative metric storage circuit 143 as a signal S143.

F-2. Operation of Fifth Embodiment Next, the operation of the fifth embodiment will be described. Information S 112 that is a transmission signal candidate of each transmission system from the transmission signal sequence storage circuit 110, an interference signal S 113 for the corresponding transmission signal candidate from the interference signal generation circuit 112, and the reception result stored in the multiplication result storage circuit 106 When the information S107 obtained by unitarily converting the signal is input, the partial metric calculation circuit 141 calculates a partial metric defined by Equation (13), Equation (19), and the like. The calculation result S141 is supplied to the recurrence equation calculation circuit 142, and the recurrence equation calculation circuit 142 obtains an accumulated value.

  At this time, Equation (21) is used as the recurrence formula, but in Equation (21), the signal S141 corresponds to the first term (partial metric of the p-th transmission system) on the right side, and the cumulative metric. The signal S142 from the storage circuit 143 corresponds to the second term (cumulative metric in the (p + 1) th transmission system). The calculation result is output to and stored in the cumulative metric storage circuit 143 as a signal S143. When the metric value obtained by adding the partial metrics up to the first transmission system is finally obtained, the result (in some cases, information on combinations of transmission signal candidates corresponding to the result) S114 is input to the maximum likelihood estimation circuit 113. Is done. Here, since the recursive equation calculation circuit 142 finally adds all the partial metrics via the cumulative metric storage circuit 143, it operates as a metric calculation means.

  The above-described processing can be realized by performing a recurrence formula in order when the Nth transmission system, the (N-1) th transmission system, the (N-2) th transmission system,. In this case, the processing delay can be shortened rather than performing all the operations collectively at the end. Here, the example is based on the assumption that the recursion formula (21) is used, but it goes without saying that the metric may be directly obtained using the formula (20). In this case, the cumulative metric storage circuit 143 is omitted by performing the calculation of the formula (20) by the recurrence formula calculation circuit 142.

G. Sixth Embodiment Next, a sixth embodiment of the present invention will be described. The sixth embodiment is characterized in that the received signal S101 is a subcarrier signal after multicarrier demodulation, compared to the second embodiment described above. Therefore, in the sixth embodiment, the received signal is a multicarrier modulation signal.

G-1. Configuration of Sixth Embodiment FIG. 14 is a block diagram showing a configuration of a spatial multiplexing signal detection circuit according to a sixth embodiment of the present invention. Note that portions corresponding to those in FIG. 10 are denoted by the same reference numerals and description thereof is omitted. In the figure, the sixth embodiment is characterized in that a ranking circuit 151 and a reverse replacement processing circuit 152 are further provided to the first embodiment shown in FIG. The ranking circuit 151 is added between the estimated channel matrix generation circuit 102 and the QR decomposition circuit 103. In the ranking circuit 151, ranking is performed using evaluation indices such as Equation (22) to Equation (24) based on the signal S103 from the estimated channel matrix generation circuit 102, and the ranking order as in Equation (25). A permutation matrix corresponding to is generated. The reverse permutation processing circuit 152 is arranged at the subsequent stage of the maximum likelihood estimation circuit 113. The reverse replacement processing circuit 152 performs the order of the transmission system in the ranking circuit 151 for the output signal S116 from the maximum likelihood estimation circuit 113 in accordance with the information S151 of the matrix column replacement content notified from the ranking circuit 151. Replacement is performed as the reverse of the process.

G-2. Operation of Sixth Embodiment Next, the operation of the sixth embodiment will be described. In the ranking circuit 151, ranking is performed using evaluation indices such as Equation (22) to Equation (24) based on the signal S103 from the estimated channel matrix generation circuit 102, and the ranking order as shown in Equation (25). A permutation matrix corresponding to is generated. Using this generated matrix, a process corresponding to Equation (27), that is, a matrix column replacement process is performed. The matrix information S152 after the replacement process is input to the QR decomposition circuit 103. Further, the information S151 of the matrix column permutation contents is notified to the reverse permutation processing circuit 152, and the order of the transmission system is reversed with respect to the output signal S116 from the maximum likelihood estimation circuit 113 as compared with that performed by the ranking circuit 151. Replacement is performed as the process. The replacement here can also be expressed as Equation (29). This replacement process result is output as the final estimation result S153. The other processing contents are the same as those described with reference to FIG.

H. Seventh Embodiment Next, a seventh embodiment of the present invention will be described. In the seventh embodiment, in the above-described sixth embodiment, the ranking circuit 151 shown in FIG. 14 is a transmission system having a large transmission index value such as Expression (22) to Expression (24). It is characterized by rearranging in the larger one. Otherwise, the configuration is the same as that shown in FIG.

I. Eighth Embodiment Next, an eighth embodiment of the present invention will be described. In the eighth embodiment, in the transmission signal candidate narrowing circuit 108 shown in FIG. 10 according to the first embodiment described above, the size comparison of diagonal components excluding the first row and first column components of the R matrix after QR decomposition is performed. The number of candidates for narrowing is set to a large value for a transmission system having a small absolute value, and the number of candidates for narrowing is set to a small value for a transmission system having a large absolute value. Other than that, there is no difference from FIG. 10 and FIG. At this time, the value of the diagonal term can be set flexibly. However, as a simple method, for example, the number of candidates is determined in advance as 16, 16, or 32, and the diagonal component is determined. This number should correspond to the result of the size comparison.

J. et al. Ninth Embodiment Next, a ninth embodiment of the present invention will be described. The ninth embodiment is characterized in that the received signal is a subcarrier signal after multicarrier demodulation, compared to the first or sixth embodiment described above. Therefore, in the ninth embodiment, the received signal is a multicarrier modulation signal.

J-1. Configuration of the ninth embodiment (part 1)
FIG. 15 is a block diagram showing a configuration of a spatial multiplexing signal detection circuit in which the ninth embodiment is applied to the first embodiment. Note that portions corresponding to those in FIG. 10 are denoted by the same reference numerals and description thereof is omitted. In the figure, FFT circuits 161-a to 161-b perform multicarrier demodulation on a multicarrier modulation signal S161, which is a received signal, and supply it to a channel estimation circuit 101 and a matrix multiplier 105 as a subcarrier signal S162.

J-2. Operation of the ninth embodiment (part 1)
Next, the operation of the above-described ninth embodiment will be described. Multicarrier modulation signal S161, which is a received signal, is input to FFT circuits 161-a to 161-b and subjected to multicarrier demodulation to obtain subcarrier signal S162. The subcarrier signal S162 is input to the channel estimation circuit 101 and the matrix multiplication circuit 105. The other operations are the same as those in the first embodiment shown in FIG.

J-3. Configuration of the ninth embodiment (2)
FIG. 16 is a block diagram showing a configuration of a spatial multiplexing signal detection circuit in which the ninth embodiment is applied to the sixth embodiment. Note that portions corresponding to those in FIG. 14 are denoted by the same reference numerals and description thereof is omitted. In the figure, FFT circuits 171-a to 171-b demodulate a multicarrier modulation signal S 171 that is a received signal, and supply it to the channel estimation circuit 101 and the matrix multiplier 105 as a subcarrier signal S 172.

J-4. Operation of the ninth embodiment (2)
Next, the operation of the above-described ninth embodiment will be described. Multicarrier modulation signal S171 as a reception signal is input to FFT circuits 171-a to 171-b and subjected to multicarrier demodulation to obtain subcarrier signal S172. This subcarrier signal S 172 is input to channel estimation circuit 101 and matrix multiplication circuit 105. The other operations are the same as those in the sixth embodiment shown in FIG.

  According to the first to ninth embodiments described above, when receiving a spatially multiplexed signal and performing signal detection, particularly when the number of modulation multilevels is increased, an increase in circuit scale is suppressed, and an excellent error is achieved. Rate characteristics can be realized. In particular, a signal detection method with excellent error rate characteristics can be realized even in a wireless LAN or the like in which the circuit scale or power consumption is severely limited.

  Further, according to the first to ninth embodiments described above, the calculation of QR decomposition is performed based on the channel estimation result, and the calculation is performed only at the head portion of the received packet. Therefore, it is not necessary to perform QR decomposition over the data portion, and power consumption can be reduced.

  Furthermore, even when a signal detection method that has a small circuit scale in terms of circuit scale and power consumption but has a poor error rate characteristic has to be used in the past, the present invention provides an excellent error rate characteristic. The signal detection method can be applied, and an effect that a high-quality wireless system can be provided is also obtained. Furthermore, since a feature with excellent error rate characteristics can be used, it means that communication with a low CNR is possible while reducing the circuit scale, which can also contribute to the expansion of the communication area.

  In the first to ninth embodiments described above, it is specified that a signal is transmitted in a general expression with a multi-level signal (m). However, naturally, the m-value signal includes m. Wireless communication system such as value QAM signal, m value PSK signal, m value ASK, m value FSK signal, m value CCK (Complementary Code Keying) signal, SS (Spread Spectrum) signal multiplexed with m codes Of course, general modulation schemes applied to the above are applicable. In this case, at the time of data reception, a classifying process condition corresponding to the modulation multi-level number of the received data may be selected.

  In the ninth embodiment, the FFT circuits 161-a to 161-b and 171-a to 171-b that perform multicarrier demodulation remove the repetitive signal section called a guard interval, and then perform multicarrier demodulation. Of course it is possible. Furthermore, it is naturally possible to apply after multicarrier demodulation in the multicarrier modulation system. Even in the multicarrier modulation scheme, signal processing is performed on each subcarrier after multicarrier demodulation. Therefore, the method of the present invention can naturally be applied not only to the multicarrier modulation method but also to the case of using a single carrier modulation method.

  Further, in the channel estimation circuit 101. A plurality of configurations are conceivable depending on in what signal format the channel estimation preamble signal for the spatially multiplexed signal on the transmission side is transmitted. Various configurations can be considered, such as when synchronous detection is performed using the preamble signal provided in the receiver to obtain a channel estimation result, or when inverse matrix calculation is performed, but naturally these configurations are applied. Is possible.

  In the first to ninth embodiments described above, naturally, as shown in Equation (5), the transmission system can be expanded even when the number of transmission systems is three or more, and the number of reception systems is also increased. Naturally, expansion is possible in the case of three or more. Further, in a configuration of three or more multiplexes, it is not necessary to provide as many reception sequences as the number of transmission systems to be transmitted. Depending on system requirements, the number of transmission systems is the same as or less than the number of transmission systems. Various combinations are naturally possible depending on the case. Further, even if the maximum number of multiplexing that can be adapted is applied and operated while changing the number of multiplexing during operation, it is naturally applicable.

  In addition, in the configuration according to the first to ninth embodiments described above, it is possible to reduce the power consumption by operating each unit and each circuit only when necessary, instead of always operating each unit and each circuit.

  In the first to ninth embodiments described above, each unit and each circuit are basically implemented on hardware, but may be executed in a computer system. In this case, a series of processing steps by the above-described units and circuits is stored in a computer-readable recording medium in the form of a program, and the above-described processing is performed by the computer reading and executing the program. . That is, each unit and each circuit is realized by a central processing unit such as a CPU reading the above program into a main storage device such as a ROM or RAM and executing information processing / calculation processing. Also good.

  Here, the computer-readable recording medium means a magnetic disk, a magneto-optical disk, a CD-ROM, a DVD-ROM, a semiconductor memory, or the like. Alternatively, the computer program may be distributed to the computer via a communication line, and the computer that has received the distribution may execute the program.

It is a conceptual diagram for demonstrating the constellation of the transmission signal in 16QAM in this invention. It is a conceptual diagram for explaining the constellation of a transmission signal whose scale has been changed so that the signal point distance is 2.0 × _{rNN for} both the I axis and the Q axis. It is a conceptual diagram which shows the 1st example regarding the setting of the threshold value in a classification process. It is a conceptual diagram which shows the 2nd example regarding the setting of the threshold value in a classification process. It is a conceptual diagram which shows the 3rd example regarding the setting of the threshold value in a classification process. It is a conceptual diagram which shows the 1st example regarding the setting of the threshold value in the narrowing-down process of a transmission signal. It is a conceptual diagram which shows the 2nd example regarding the setting of the threshold value in the narrowing-down process of a transmission signal. It is a conceptual diagram which shows the 3rd example regarding the setting of the threshold value in the narrowing-down process of a transmission signal. It is a conceptual diagram for demonstrating a mode until all the signals from the Nth transmission system to one transmission system are estimated. It is a block diagram which shows the structure of the spatial multiplexing signal detection circuit by 1st Embodiment of this invention. It is a block diagram which shows the structure of the spatial multiplexing signal detection circuit by 3rd Embodiment of this invention. It is a block diagram which shows the structure of the spatial multiplexing signal detection circuit by 4th Embodiment of this invention. It is a block diagram which shows the structure of the spatial multiplexing signal detection circuit by 5th Embodiment of this invention. It is a block diagram which shows the structure of the spatial multiplexing signal detection circuit by 6th Embodiment of this invention. It is a block diagram which shows the structure of the spatial multiplexing signal detection circuit by 9th Embodiment (the 1) of this invention. It is a block diagram which shows the structure of the spatial multiplexing signal detection circuit by 9th Embodiment (the 2) of this invention. It is a block diagram for demonstrating spatial multiplexing transmission. It is a conceptual diagram for demonstrating the conventional MLD system. It is a block diagram which shows the structure of the conventional spatial multiplexing signal detection circuit. It is a conceptual diagram which shows an example of the packet format used with a spatial multiplexing transmission system. It is a conceptual diagram at the time of showing in two dimensions in the frequency-time of an example packet format used with a spatial multiplexing transmission system.

Explanation of symbols

101 channel estimation circuit (channel estimation means)
102 Estimated channel matrix generation circuit (estimated channel matrix generation means)
103 QR decomposition circuit (QR decomposition means)
104 Complex conjugate transpose arithmetic circuit (complex conjugate transpose arithmetic means)
105 Matrix multiplication circuit (matrix multiplication means)
106 Multiplication result storage circuit (multiplication result storage means)
107 Transmission signal estimation circuit (transmission signal estimation means)
108 Transmission signal candidate narrowing circuit (signal candidate narrowing means)
109 Proximity signal point data table (Proximity signal point data table)
110 Transmission signal sequence storage circuit (transmission signal sequence storage means)
111 Transmission signal candidate metric calculation circuit (transmission signal candidate metric calculation means)
112 Interference signal generation circuit (interference signal generation means)
113 Maximum likelihood estimation circuit (maximum likelihood estimation means)
121 interference cancellation signal calculation circuit (interference cancellation signal calculation means)
122 Classifying circuit (classifying means)
123 class threshold value generation circuit 124 transmission signal approximate estimation circuit (transmission signal approximate estimation means)
131 Replica generation circuit (multiplication means)
132 Transmission signal generation circuit 133 Replica storage circuit (replica storage means)
134 Interference replica adding circuit (adding means)
141 Partial metric calculation circuit (partial metric calculation means)
142 Recursive equation circuit (metric calculation means)
143 Cumulative metric storage circuit 151 Ranking circuit (ranking means, estimated channel matrix rearrangement means)
152 Reverse replacement processing circuit (reverse replacement means)
161-a to 161-b FFT circuit (multi-carrier demodulating means)
171-a to 171-b FFT circuit (multicarrier demodulation means)

Claims (9)

A channel estimation means for receiving a plurality of received signals and estimating channel distortion of the propagation path between the transmitting and receiving antennas;
An estimated channel matrix generating means for generating an estimated channel matrix having the distortion of the channel estimated by the channel estimating means as a component;
QR decomposition means for performing unitary and triangulation (QR) decomposition operations on the estimated channel matrix generated by the estimated channel matrix generation means;
A complex conjugate transpose operation means for performing a Q ^H is the complex conjugate transpose operation with respect to the unitary matrix Q obtained by the QR decomposition unit,
Matrix multiplication means for performing a Q ^H × r matrix multiplication operation on a received column vector signal r having a plurality of received signals as constituent elements, using a calculation result obtained by the complex conjugate transposition calculating means;
Multiplication result storage means for storing calculation results of each transmission system of the matrix multiplication means;
Using the output signal of each transmission system obtained from the multiplication result storage means, the matrix element signal of the upper triangular matrix R which is another output signal of the QR decomposition means, and the interference signal which is a signal other than the desired transmission system A transmission signal estimation means for estimating a transmission signal of each desired transmission system;
A proximity signal point data table that stores a correspondence between a transmission signal of each transmission system estimated by the transmission signal estimation means and a signal point having a short distance between signal points on a transmission constellation as a database table;
By referring to the proximity signal point data table, a predetermined number of transmission signals of each transmission system estimated by the transmission signal estimation means and a plurality of estimated transmission signals whose signal point distances are close on the transmission constellation are candidates. A signal candidate narrowing means to select;
Each estimated transmission signal selected by the signal candidate narrowing means is associated with a plurality of estimated transmission signals of each transmission system and information of other transmission systems used to estimate the estimated transmission signal of the transmission system. Transmission signal sequence storage means for storing as a sequence;
Other than the desired transmission system based on the estimated transmission signal of each estimated transmission signal sequence stored in the signal sequence storage means and the matrix element signal of the upper triangular matrix that is another output signal of the QR decomposition means Interference signal generating means for generating an interference signal that is a signal of
For each candidate of each estimated transmission signal sequence stored in the transmission signal sequence storage means, a metric that is a signal point distance between the estimated reception signal and the actual reception signal when the candidate transmission signal is transmitted is calculated. Transmitting signal candidate metric calculating means to perform,
The smallest metric is selected from all the metrics corresponding to each estimated transmission signal sequence calculated by the transmission signal candidate metric calculating means and stored in the transmission signal sequence storage means, and the estimated transmission signal corresponding to the metric And a maximum likelihood estimating means for outputting the sequence as a finally estimated transmission signal sequence.   The QR decomposition unit selects the unitary matrix Q and the upper triangular matrix R so that elements having the same row number and column number are real numbers among the elements of the generated upper triangular matrix R. Item 2. The spatial multiplexing signal detection circuit according to Item 1. The transmission signal estimation means includes
When the number of multiplexed transmission systems is N (N is an integer of 2 or more)
In a p-th (1 ≦ p ≦ N: p is an integer) transmission system, an interference signal that is an output from the interference signal generation unit is obtained from a signal that is an element in the p-th row of a vector output from the matrix multiplication unit. Interference cancellation signal computing means for obtaining a subtracted interference cancellation signal;
Classifying means for performing a classifying process depending on the value of the p-th row and the p-th column of the upper triangular matrix R with respect to the real part and the imaginary part of the interference cancellation signal that is an output from the interference cancellation signal calculation means;
Transmission signal approximate estimation means for obtaining an approximate value of the estimated transmission signal based on a table for managing the correspondence between the class value and the approximate value of the transmission signal in the classifying means, and outputting the estimated transmission signal as the estimated transmission signal. The spatially multiplexed signal detection circuit according to claim 2. The interference signal generating means includes
Multiplying means for multiplying each non-zero element of the upper triangular matrix R by a signal that can be transmitted by a transmitting station;
Replica storage means for storing a multiplication result by the multiplication means;
The addition means which acquires the interference signal which is signals other than a desired transmission system | strain by the addition of the multiplication result memorize | stored in the said replica memory | storage means is provided, The any one of Claim 1 thru | or 3 characterized by the above-mentioned. Spatial multiplexed signal detection circuit. The transmission signal candidate metric calculation means includes:
When the multiplexed transmission system is N (N is an integer of 2 or more),
A candidate set of transmission signals from the p-th (1 ≦ p ≦ N: p is an integer) stored in the transmission signal sequence storage means to the N-th transmission system is read, and for each candidate set of transmission signals, The magnitude of the difference between the signal obtained by multiplying the signal of the transmission system corresponding to each element of the p-th row of the matrix elements of the upper triangular matrix R and the value of the p-th row of the vector signal stored in the multiplication result storage means A partial metric calculation means for calculating a partial metric representing
For each combination of transmission signal candidates from the first to the Nth transmission system, the value of the partial metric obtained by the partial metric calculating means in the first to Nth transmission sequences is added to the entire transmission system. 5. The spatial multiplexing signal detection circuit according to claim 1, further comprising: metric calculation means for calculating a metric for. Ranking means for determining the superiority or inferiority of the reception state of each transmission system based on the estimated channel matrix output from the estimated channel matrix generation means;
Estimated channel matrix rearranging means for switching the order of the columns of the estimated channel matrix based on the superiority or inferiority of the reception state determined by the ranking means;
A reverse permutation means for performing a reverse permutation process and a permutation process for each column performed by the estimated channel matrix rearrangement means for the transmission signal sequence output from the maximum likelihood estimation means. The spatially multiplexed signal detection circuit according to claim 1.   The estimated channel matrix rearranging unit performs rearrangement such that a transmission system having a better reception state in the superiority or inferiority of the reception state determined by the ranking unit is arranged in the lower stage. 7. The spatial multiplexing signal detection circuit according to 6.   The signal candidate narrowing means has a large absolute value of the diagonal components of the second to Nth transmission systems of the upper triangular matrix R when the multiplexed transmission system is N (N is an integer of 2 or more). 8. The spatial multiplexed signal detection circuit according to claim 1, wherein the number of candidates for transmission signals is set so that the number of candidates for narrowing is the same or smaller. Multicarrier demodulation is performed on the plurality of received signals, and multicarrier demodulation is performed to output a subcarrier signal.
9. The spatially multiplexed signal detection circuit according to claim 1, wherein the subcarrier signal is used as an input signal to the channel estimation means and the matrix multiplication means.

Images