JP2012049733A - Demodulator and demodulation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To obtain a demodulator which requires only small number of operations to remove delay waves even in the environment having delay waves with a long delay time.SOLUTION: The demodulator comprises: a transmission line calculation part which calculates a transmission line matrix by transmission line estimation based on received signals; a column deletion part which deletes the prescribed number of columns from the beginning of the transmission line matrix; and a demodulation processing part which performs demodulation processing based on the received signals and on the transmission line matrix after the deletion by the column deletion part.

Description

  The present invention relates to a demodulator and a demodulation method.

  In a digital communication system, frequency selectivity and time fluctuation occur in a transmission path due to multipath fading caused by reflection of a transmission signal on a building or the like and Doppler fluctuation caused by movement of a terminal. In a multipath environment, the received signal is a signal in which a transmission symbol interferes with a symbol that arrives after a delay time has elapsed. In a multipath environment in which the transmission path fluctuates over time, a receiver for removing interference due to a delayed wave having a long delay time has been studied.

  Conventionally, trellis demodulation has been performed to perform maximum likelihood symbol sequence estimation. However, in the conventional trellis demodulation method, since the memory exponentially increases with the delay time, a large amount of calculation is required in an environment where a delayed wave with a long delay time exists.

  There is a branch demodulation method as a method other than the trellis demodulation. For example, Japanese Patent Application Laid-Open No. 2004-228561 is a technique of performing data rearrangement in order to optimize a link state or a bit error rate, obtaining a triangular matrix using a matrix decomposition technique, and detecting a sphere using branch demodulation. Patent Document 2 below discloses a technique for dividing the transmission line matrix and obtaining an upper triangle in order to perform optimal block demodulation with a low amount of computation. Patent Document 3 below discloses a technique for detecting a sphere using a limited memory. Sphere detection is performed by branch demodulation.

JP 2009-523373 A JP 2009-100408 A JP 2008-172340 A

  However, as described above, when trellis demodulation is used, there is a problem that a large amount of calculation is required in an environment where a delayed wave having a long delay time exists.

  Further, in Patent Documents 1 to 3 described above, a demodulation method that does not use trellis-type demodulation is disclosed, but there is no description about calculation in an environment where a delayed wave having a long delay time exists. Therefore, there has been a problem that a method for reducing the amount of calculation in an environment where a delayed wave having a long delay time exists is not disclosed.

  The present invention has been made in view of the above, and an object of the present invention is to provide a demodulator and a demodulation method capable of removing a delayed wave with a low calculation amount even in an environment where a delayed wave with a long delay time exists. .

  In order to solve the above-described problems and achieve the object, the present invention provides a transmission path calculation unit that calculates a transmission path matrix by performing transmission path estimation based on a received signal, and a predetermined number from the top of the transmission path matrix. A column deletion unit that deletes the number of columns, a demodulation processing unit that performs a demodulation process based on the received signal and the transmission path matrix that has been deleted by the column deletion unit, To do.

  According to the present invention, it is possible to remove a delayed wave with a small amount of computation even in an environment where a delayed wave having a long delay time exists.

FIG. 1 is a diagram illustrating a configuration example of a communication system according to the first embodiment. FIG. 2 is a diagram illustrating a functional configuration example of the transmitter and the receiver according to the first embodiment. FIG. 3 is a diagram illustrating a configuration example of data and pilot symbols in the first embodiment. FIG. 4 is a diagram illustrating an example of a multipath transmission line when L = 4. FIG. 5 is a diagram illustrating the concept of BB demodulation. FIG. 6 is a diagram illustrating a state of rearrangement of example # 1 in the second embodiment. FIG. 7 is a diagram illustrating a state of rearrangement of example # 2 in the second embodiment. FIG. 8 is a diagram illustrating an example of a rearrangement matrix in the case of Example # 1. FIG. 9 is a diagram illustrating an example of a rearrangement matrix in the case of Example # 2. FIG. 10 is a diagram illustrating an example of a matrix before rearrangement for explaining decomposition example # 1 of the Cholesky decomposition method of the third embodiment. FIG. 11 is a diagram illustrating an example of a matrix after rearrangement for explaining the decomposition example # 1. FIG. 12 is a diagram for explaining a decomposition example # 1 of the Cholesky decomposition method of the third embodiment. FIG. 13 is a diagram illustrating an example of a matrix A (bold) before rearrangement for explaining decomposition example # 2 of the Cholesky decomposition method. FIG. 14 is a diagram illustrating an example of a matrix after rearrangement for explaining decomposition example # 2. FIG. 15 is a diagram for explaining a decomposition example # 2 of the Cholesky decomposition method of the third embodiment. FIG. 16 is a diagram illustrating an example of a sparse transmission line matrix. FIG. 17 is a flowchart illustrating an example of the Cholesky decomposition procedure of the fourth embodiment.

  Embodiments of a demodulator and a demodulation method according to the present invention will be described below in detail with reference to the drawings. Note that the present invention is not limited to the embodiments.

Embodiment 1 FIG.
FIG. 1 is a diagram illustrating a configuration example of a communication system according to the present embodiment. As shown in FIG. 1, the communication system according to the present embodiment includes a transmitter 1 and a receiver 2. The transmitter 1 and the receiver 2 according to the present embodiment communicate in a multipath environment. The receiver 2 receives the direct wave W1 to which the transmission signal transmitted from the transmitter 1 directly reaches, and the transmission signal transmitted from the transmitter 1 is reflected by the reflectors 3-1 and 3-2 such as buildings. The received interference waves W2-1 and W2-2 are received. Interference waves W2-1 and W2-2 are received with a delay from direct wave W1. Further, thermal noise 4 is added to the received direct wave W1 and interference waves W2-1 and W2-2.

  FIG. 2 is a diagram illustrating a functional configuration example of the transmitter 1 and the receiver 2 according to the present embodiment. The transmitter 1 of the present embodiment includes a symbol encoder 11 and a transmission antenna 12, and is the same as a transmitter that has been conventionally used. In addition, the receiver 2 of the present embodiment includes a reception antenna 13, a reception filter 14, and a demodulation unit 15.

First, the transmission operation of the transmitter 1 will be described. When Q is an integer that is a power of 2, in the transmitter 1, the symbol encoder 11 maps log ₂ Q bits to one of Q signal points. There is no restriction on the mapping method, and any method may be used, but generally, Phase Shift Keying, Quadrature Amplitude Modulation, etc. are used. The mapped signal is sent from the transmission antenna 12. It is assumed that the information bit transmitted at time k is b _k and the symbol output by mapping the information bit by the symbol encoder 11 is s _k .

  Next, the demodulation operation of the receiver 2 will be described. In the receiver 2, the signal transmitted from the transmitter 1 is received by the reception antenna 13 as a radio wave, and the reception filter 14 removes a signal other than a desired frequency region from the received signal, and the signal after the removal is obtained. Output to the demodulator 15. The demodulator 15 demodulates the signal input from the reception filter 14 and obtains the signal transmitted from the transmitter 1. In general, an A / D (analog / digital) converter or the like is provided between the reception filter 14 and the demodulator 15, but these are not shown here for the sake of simplicity.

Hereinafter, a demodulation process performed by the demodulation unit 15 of the present embodiment will be described. The demodulator 15 of the present embodiment uses a transmission path calculation section 151 that performs transmission path estimation, a column deletion section 152 that performs column deletion of the transmission path matrix, and a BB (Branch and Branch) using the transmission path matrix after column deletion. Bound) demodulation processing unit 153 that performs demodulation processing by demodulation or the like. In the following description, symbols used are shown.
T _S : 1 symbol time (seconds)
N _T : Number of pilot symbols L: Number of multipaths in transmission line N: Number of symbols to be demodulated [A (bold)] _ij : ij element of matrix A (bold) A (bold) = diag (a) is , The diagonal component of the matrix A (bold) is formed by the vector a, and the other components are 0. I _N is an N × N diagonal matrix, and 0 (bold) _{N × M} is an N × M 0 matrix. Further, the Cholesky decomposition of the matrix A (bold) is expressed by the following equation (1). U (bold) indicates an upper triangular matrix and L (bold) indicates a lower triangular matrix.

When N received signals sampled at the symbol time T _S after passing through the transmission path having L multipaths and the reception filter 14 are represented by a received signal vector y (bold), y (bold) is expressed by the following equation: (2). The received signal vector y (bold) is y (bold) = [y _N-1 , y _N-2 ,..., Y ₀ ] ^T , and the transmission signal vector s (bold) is s (bold) = [s _{N-1, s N-2} ..., and s _0] ^T, then the white thermal noise vector w (in bold) w _{(bold) = [w N-1,} w N-2 ..., and w _0] ^T, the transmission path Let the matrix be H (bold).

Note that although an example in which demodulation is performed using a signal sampled at the symbol time T _S is described in this embodiment, demodulation may be performed using a signal sampled at an interval shorter than the symbol time.

When a vector in which M zeros are arranged is 0 (bold) _M , an impulse response h (bold) _k of a multipath transmission line of k sample time can be expressed by the following equation (3).

Here, D _l is the delay symbol time of the l-th pass (the delay time is expressed in symbol units), and K is a value that satisfies the relationship shown in the following equation (4).

The direct wave transmission path is _denoted by h _{k, 0} . Transmission path estimation is performed using known pilot symbols or the like, and transmission path values h _{i, j} can be obtained. FIG. 3 is a diagram illustrating a configuration example of data and pilot symbols according to the present embodiment. Thus, it is assumed that the transmitter 1 transmits a transmission signal including pilot symbols. The configuration of data and pilot symbols is not limited to the example of FIG. 3, and any configuration may be used.

FIG. 4 is a diagram illustrating an example of a multipath transmission line when L = 4. In the example of FIG. 4, the impulse response h _{k, 1} corresponding to the direct wave W1 (preceding wave) is delayed from D ₁ = 5 symbols, and the impulse response h _{k, 1} exists, and further, the impulse responses h _{k, 1} to D ₂ = 3 symbol delay exists impulse response h _{k, 2,} further from the impulse response h _{k, 2} and D ₃ = 4-symbol delay exists impulse response h _{k, 3.} FIG. 4 is an example, and each impulse response of the multipath transmission path is not limited to the case of FIG.

  The transmission line matrix H (bold) is an upper triangular matrix, and the (i, j) element can be expressed by the following equation (5).

As an example, the relationship between the transmission path matrix of L = 2, N = 6, and D ₁ = 2 and the received signal vector is shown in the following formula (6). Note that k = 0 indicates the start position of the data frame.

By multiplying the above equation (2) by H (bold) ^H from the left side, the processing shown in the following equation (7) is performed.

A = H and A (bold) ^H A (bold) Cholesky decomposition upper triangular matrix is U (bold) (U (bold) is U (bold) ^H U (bold) = A (bold) ^H A (Satisfying bold), U (bold) ^-H H (bold) When the white noise vector rearranged by ^H is n (bold) = U (bold) ^-H H (bold) ^H w (bold) Equation (2) can be modified as the following Equation (8).

  In the present embodiment, the Cholesky decomposition method using the upper triangular matrix has been described for the sake of simplification. However, the column deletion method of the present embodiment is applied to the Cholesky decomposition using the lower triangle. Is also possible. Here, it is assumed that the demodulation processing unit performs the above-described Cholesky decomposition.

  Then, demodulation processing is performed based on z (bold) shown in the above equation (8). In the present embodiment, BB (Branch and Bound) demodulation, which is a well-known technique, is used as a demodulation processing technique. However, the present invention is not limited to BB demodulation, and other re-likelihood sequence demodulation schemes or “E. Viterbo and Like the sphere search described in J. Boutros, “A universal lattice code decoder for fading channels”, IEEE Transactions on Information Theory, vol. 45, no. 5, July 1999, pp. 1639-1642, etc. A demodulation method close to the demodulation accuracy of the re-likelihood sequence demodulation method may be used.

FIG. 5 is a diagram illustrating the concept of BB demodulation. As shown in FIG. 5, BB demodulation performs demodulation while going down the hierarchy (time), and the current estimated error ε _k-1 shown in the following equation (9) is smaller than the minimum estimated error ε _min found so far. If it is higher, the search is interrupted and the search is performed from the upper hierarchy. Note that z _i is a matrix element of z (bold) corresponding to time i, and assuming that N transmission symbols are transmitted in time order, the (i + 1) -th element from the bottom of z (bold) is z _i. It is. z _i ^{^} (hat) is an expected value of z _i and corresponds to U (bold) s (bold) in the above equation (8). The example shown in FIG. 5 is an example using BPSK, and the demodulation results are s ₀ = −1, s ₁ = −1, s ₂ = −1, and s ₃ = −1.

Next, a column deletion method for the transmission line matrix according to the present embodiment will be described. In order to always perform the above-described Cholesky decomposition, H (bold) ^H H (bold) needs to be a positive definite matrix. H (bold) ^{H If} H (bold) is not a positive definite matrix, there is no Cholesky decomposition matrix. Therefore, in this embodiment, the above equation (5) is rewritten as the following equation (11). That is, the first _NTR column is deleted from the transmission path matrix H (bold). N _TR is the number of columns to be deleted from H (bold). Here, for simplification of the expression display, the transmission path does not change with time, that is, the following expression (10) is used, and the time index is omitted. However, the present embodiment is the same in an environment where the transmission path changes over time. Can be applied to.

Here, for simplification, _NTR is set to a value as shown in the following formula (12), but may be set as shown in formula (13). The upper limit of N _TR is N.

For example, by deleting the first N _TR = 2 columns, the matrix illustrated in the above equation (6) becomes the following equation (14).

By deleting the first _NTR column in this way, the diagonal element of A (bold) ′ = H (bold) ′ ^H H (bold) ′ is expressed by the following equation (15), and A (bold) ) ′ Is a positive definite matrix.

  Therefore, U (bold) is obtained using H (bold) instead of the transmission line matrix H, z (bold) is calculated by the above equation (8) using this U (bold), and BB demodulation is performed. be able to.

As described above, in this embodiment, H (bold) ′ is generated by deleting the first _NTR column of the transmission line matrix H, and Cholesky decomposition is performed using H (bold) instead of the transmission line matrix H. To obtain U (bold). Therefore, when performing demodulation using Cholesky decomposition without using trellis decoding, Cholesky decomposition can be reliably and efficiently performed. Therefore, even in an environment where a delayed wave having a long delay time exists, the delayed wave can be removed with a small amount of computation.

Embodiment 2. FIG.
Next, a receiver demodulation method according to Embodiment 2 of the present invention will be described. The configuration of the communication system, the configuration of the transmitter, and the receiver of this embodiment are the same as those of the first embodiment. However, the demodulation unit 15 of the present embodiment further includes a rearrangement unit in addition to the transmission path calculation unit 151, the column deletion unit 152, and the demodulation processing unit 153 of the first embodiment. In the BB demodulation described in the first embodiment, the calculation amount at both ends of the data frame increases by deleting the column of the transmission path matrix. This is because in the received signal model using the transmission path matrix shown in Expression (11), the delay wave or the preceding wave is missing at both ends of the frame, and the model error increases. If the model error increases at both ends of the data frame, the number of searches in branch demodulation increases.

  In order to suppress such an increase in the amount of calculation, in this embodiment, the rearrangement unit performs a rearrangement process that collects data locations with many model errors in the vicinity of k = 0 in the data frame. Here, the rearrangement method of Example # 1 and Example # 2 will be described. However, the present invention is not limited to this, and any rearrangement method is possible as long as the data portions with many model errors are collected in the vicinity of k = 0 in the data frame. May be used.

  FIG. 6 is a diagram illustrating how the example # 1 is rearranged. In Example # 1, the position of the data 21 is left as it is in both ends of the data frame of the received signal (data 21 near k = 0, data 22 near k = N−1), and data near k = N−1. 22 is rearranged so as to be arranged immediately after the data 21. The data amount M to be rearranged may be arbitrarily set. In this way, portions with a lot of model errors are gathered in one place at the beginning of the data, so that the movement up and down the hierarchy during BB demodulation can be reduced. It is assumed that rearrangement corresponding to rearrangement of received signals is performed on the transmission path matrix used for BB demodulation.

  FIG. 7 is a diagram illustrating how the example # 2 is rearranged. In example # 2, of the two ends of the data frame of the received signal (data 21 near k = 0, data 22 near k = N−1), the data 22 is set near k = 0, that is, the head of the data frame, and the data 22 Data 21 is arranged after the.

Next, received signals rearranged using mathematical expressions are shown. It is assumed that the rearrangement matrix for performing the rearrangement as described above is P (bold), and the rearranged transmission line matrix A (bold) ⁺ is expressed by the following equation (16).

Then, the rearranged received signals can be expressed as P (bold) y (bold) = A (bold) ⁺ P (bold) s (bold) + P (bold) w (bold). A (bold) ^{+ H} A (bold) ⁺ The upper triangular matrix obtained by Cholesky decomposition of ⁺ is U ⁺ , and the rearranged white noise vector is n (bold) n (bold) = (U (bold) ⁺ ) ^{If -H} A (bold) ^{+ HP} (bold) w (bold), the received signal shown in equation (2) can be expressed as in equation (17) below.

  Then, BB demodulation is performed in the same manner as in the first embodiment using z (bold) shown in the above equation (17). FIG. 8 is a diagram illustrating an example of the rearrangement matrix P (bold) in the case of Example # 1. FIG. 9 is a diagram illustrating an example of a rearrangement matrix P (bold) in the case of Example # 2.

  In the present embodiment, an example in which rearrangement of received signals is performed when performing column deletion of the transmission line matrix described in the first embodiment has been described. However, when column deletion of the transmission line matrix is not performed ( When the column deletion unit 152 is not provided, the received signals of this embodiment may be rearranged. The operations of the present embodiment other than those described above are the same as those of the first embodiment.

  As described above, in the present embodiment, the received signal and the transmission path matrix are rearranged so that data portions with many model errors are grouped around k = 0 in the data frame. For this reason, since a portion with many model errors is gathered in one place at the beginning of the data, the vertical movement of the layer during BB demodulation can be reduced, and the amount of calculation can be reduced. Further, this embodiment can be similarly applied not only to model errors but also to an environment where there are many transmission path estimation errors in one place in a frame.

Embodiment 3 FIG.
Next, a receiver demodulation method according to Embodiment 3 of the present invention will be described. The configuration of the communication system, the configuration of the transmitter, and the receiver of this embodiment are the same as those of the first embodiment. However, the demodulation unit 15 of the present embodiment further includes a rearrangement unit in addition to the transmission path calculation unit 151, the column deletion unit 152, and the demodulation processing unit 153 of the first embodiment. In the present embodiment, the rearrangement unit may perform the rearrangement described in the second embodiment, or may perform rearrangement different from that in the second embodiment.

FIG. 10 is a diagram illustrating an example of a matrix A (bold) before rearrangement for explaining decomposition example # 1 of the Cholesky decomposition method of the present embodiment. Assume that the rearrangement unit of the present embodiment performs rearrangement on this matrix A (bold). FIG. 11 is a diagram illustrating an example of a matrix after rearrangement for explaining the decomposition example # 1. When A (bold) = H (bold) ^H H (bold), both ends of A (bold) correspond to both ends of the data frame of the received signal. In the decomposition example # 1 shown in FIG. 11, as in the second embodiment, rearrangement is performed so that portions corresponding to both ends of the data frame of the received signal where model errors are expected to increase are combined in one place. Yes.

That is, the upper left M × M size matrix of the matrix A shown in FIG. 10 is B (bold) ₀ , and the upper right M × M size matrix is B (bold) ₂ . A matrix between B (bold) ₀ and B (bold) ₂ is B (bold) ₁ . Similarly, the lower left M × M size matrix is C (bold) ₀ , and the lower right M × M size matrix is C (bold) ₂ . A matrix between C (bold) ₀ and C (bold) ₂ is C (bold) ₁ . A square matrix excluding the regions (B (bold) ₀ , B (bold) ₂ , C (bold) ₀ , C (bold) ₂ )) at both ends is defined as A (bold) ₁ . A (bold) ₁ left, and respectively the right of the remaining matrix A (bold) _0, A (bold) _2.

In this embodiment, to perform initially a Cholesky decomposition with A (bold) _1, as shown in FIG. 11, rearranged to A (bold) ₁ is disposed at the upper left. Therefore, when performing Cholesky decomposition of A (bold) ⁺ after the rearrangement, rather than performing the entire A (bold) ⁺ matrix to the subject, Cholesky decomposition on the divided matrix A (bold) ₁ And the remaining elements are calculated based on the result of the Cholesky decomposition.

Hereinafter, an example of a decomposition-type Cholesky decomposition algorithm that consists of two steps of Cholesky decomposition of a divided matrix and calculation of the remaining elements based on the result of this Cholesky decomposition will be described. Here, as an example, Cholesky decomposition when H (bold) ^H H (bold) = U (bold) ^H U (bold) (ie, A = H (bold) ^H H (bold)) An example of In the following, H (bold) ^H H (bold) is also referred to as a transmission line matrix.

FIG. 12 is a diagram for explaining a decomposition example # 1 of the Cholesky decomposition method of the present embodiment. As shown in (1) of FIG. 12, first, the lower triangular matrix U (bold) ₁₁ is obtained by performing Cholesky decomposition on the upper left divided matrix A ₁ of A (bold) ⁺ shown in FIG.

Then, as shown in (2) of FIG. 12, based on the value of each element of U (bold) ₁₁ , a predetermined rule for determining each element of the lower triangular matrix U after Cholesky decomposition is used. (Bold) Calculate ₁₂ elements. Next, each element of U (bold) ₁₃ is calculated based on the value of each element of U (bold) ₁₂ as shown in (3) of FIG. Thereafter, similarly, as (4) to (6) in FIG. 12, U (bold) ₂₂ , U (bold) ₂₃ , U (bold) ₃₃ (U () in order based on each element of the matrix obtained in the previous processing. Bold) Calculate the triangular area below ₂₃ ). As the above-mentioned predetermined rule, any rule for calculating each element of Cholesky decomposition that is generally used may be used.

FIG. 13 is a diagram illustrating an example of a matrix A (bold) before rearrangement for explaining decomposition example # 2 of the Cholesky decomposition method. FIG. 14 is a diagram illustrating an example of a matrix after rearrangement for explaining decomposition example # 2. The examples of FIGS. 13 and 14 correspond to, for example, the case where it is known that the model error of the received signal becomes particularly large at the start position of the data frame. In this case, as shown in FIGS. 13 and 14, the square matrix A ₁ to be subjected to Cholesky decomposition can be a matrix excluding the region corresponding to the start position portion of the data frame. For the matrix shown in FIG. 14 as well, Cholesky decomposition can be performed using a partition matrix as follows.

FIG. 15 is a diagram for explaining a decomposition example # 2 of the Cholesky decomposition method of the present embodiment. As shown in (1) of FIG. 15, first, a lower triangular matrix U (bold) ₁₁ is obtained by performing Cholesky decomposition on the upper left divided matrix A ₁ of A (bold) ⁺ shown in FIG.

Next, as shown in (2) of FIG. 15, each element of U (bold) ₁₂ is calculated based on the value of each element of U (bold) ₁₁ . Then, each element of U (bold) ₂₂ is calculated based on the value of each element of U (bold) ₁₂ as shown in (3) of FIG.

  A specific calculation example using the Cholesky decomposition method of the present embodiment is shown below. The rearranged matrix of the above-described decomposition example # 2 is subjected to decomposition. Here, the matrix A (bold) to be decomposed is a matrix represented by the following equation (18).

The matrix A (bold) ⁺ after rearrangement of the matrix A (bold) shown in the equation (18) can be expressed by the following equation (19).

Next, Cholesky decomposition is performed by the Cholesky decomposition method of the present embodiment. U (bold) ₁₁ after the Cholesky decomposition of the matrix A ₁ shown in FIG. 15 is expressed by the following equation (20).

Based on the above U (bold) ₁₁ , each element of U (bold) ₁₂ is obtained by the following equation (21).

Finally, based on U (bold) ₁₂ , each element of U (bold) ₂₂ is obtained by the following equation (22). Note that the rearrangement unit performs rearrangement on the received signal so as to correspond to rearrangement of the transmission path matrix.

  In the present embodiment, an example of performing rearrangement when performing column deletion of the transmission line matrix described in the first embodiment has been described. The forms may be rearranged. The operations of the present embodiment other than those described above are the same as those of the first embodiment.

Embodiments 1 to 3 can also be used for transmission paths that vary with time. When the data block shown in FIG. 3 is demodulated in order to estimate the time-varying transmission path, the transmission path between the data blocks can be estimated using both pilot symbol blocks before and after the data block. . Techniques for performing this transmission path estimation are described in, for example, “T.K. Moon and W. C. Stirling,“ Mathematical methods and algorithms for signal processing ”, Prentice Hall, Upper Saddle River, New Jersey, 2000”. Interpolation method, “D.K. Borah and B.D.Hart,“ Robust receiver structure for time-varying frequency-flat Rayleigh fading channels ”, IEEE Trans.on Commun.Vol.47, no.3, March 1999 (pp. 360-364) may be used to estimate the transmission path using the basis expansion model or the like to obtain h (bold) _k .

As described above, in the present embodiment, when performing the Cholesky decomposition of the matrix A (bold) ⁺ which is the rearranged transmission path, among the divided matrices obtained by dividing the matrix A (bold) ⁺ to be decomposed Cholesky decomposition is performed on the upper left one to obtain U (bold) ₁₁ , U ₁₁ is arranged at the upper left of U (matrix) after Cholesky decomposition of matrix A (bold) ⁺ , and thereafter based on U ₁₁ Then, the elements of U (matrix) are determined, and the elements of U (matrix) are determined based on the elements of U (matrix) determined thereafter. Therefore, the amount of calculation can be reduced as compared with the case where the entire matrix is subjected to Cholesky decomposition.

Embodiment 4 FIG.
Next, a receiver demodulation method according to the fourth embodiment of the present invention will be described. The configuration of the communication system, the configuration of the transmitter, and the receiver of this embodiment are the same as those of the first embodiment. The demodulation unit 15 of the present embodiment includes a transmission path calculation unit 151, a column deletion unit 152, and a demodulation processing unit 153 as in the first embodiment, but the Cholesky decomposition method performed by the demodulation processing unit is not implemented. Different from Form 1.

In an environment where a delayed wave having a long delay time exists, Cholesky decomposition of a sparse transmission line matrix H (bold) ^H H (bold) is performed. In the present embodiment, a method for performing Cholesky decomposition with a small amount of computation by making use of the characteristics of a sparse transmission line matrix will be described.

In the present embodiment, a delay environment (D _i = D) with an equal delay time is assumed. FIG. 16 is a diagram illustrating an example of a sparse transmission line matrix (H (bold) ^H H (bold)). Note that black dots in FIG. 16 indicate elements having values (not 0). Note that the number of delay symbols (D) is calculated based on the transmission path estimation result and the like.

As an example, the following equation (23) shows an H (bold) ^H H (bold) matrix when the transmission line matrix H (bold) shown in the above equation (11) is used.

In this embodiment, first, a Cholesky decomposition of H (in bold) ^H H small matrix A contracted (bold) (bold) _s = H (in bold) _s ^H H _s (bold) (A (bold) _s = H (bold) _s ^H H _s (bold) = U (bold) _s ^H U _s (bold)). Then, U (bold) obtained by Cholesky decomposition of H (bold) ^H H (bold) is obtained using U _s (bold). In this way, since the Cholesky decomposition is performed on a matrix having a small size, it is possible to reduce the amount of calculation and the number of memories for storing values.

FIG. 17 is a flowchart showing an example of the Cholesky decomposition procedure of the present embodiment. First, a reduced matrix A (bold) _s = H (bold) _s ^H H _s (bold) is obtained based on the following equation (24) (step S1). The size of the matrix A (bold) _s is N _s = R (N / D−1) (R (x) is a rounding function (x is rounded to the nearest integer)).

Cholesky decomposition is performed on the reduced matrix A (bold) _s (A (bold) _s = U (bold) _s ^H U _s (bold): step S2). As variable initialization processing, l = 1, j = 0, d = 1, U (bold) = 0 (bold) _{N × N} (step S3).

  It is determined whether (l-1) D + d + jD ≦ N (step S4). If (l-1) D + d + jD ≦ N (step S4 Yes), the process proceeds to step S5, and (l-1) D + d + jD> N. In the case of (No in step S4), the process proceeds to step S6.

In step S5, [U (bold)] _{(l-1) D + d, (l-1) D + jD + d} = [U (bold) _s ] _{l, l +} j is set as the copy operation (step S5). ). In step S6, d = d + 1 is set (step S6). Next, it is determined whether or not d ≦ D (step S7). If d ≦ D (step S7 Yes), the process returns to step S4. If d> D (No in step S7), the process proceeds to step S8.

  In step S8, d = 1 and j = j + 1 are set (step S8). Then, it is determined whether or not j ≦ D (step S9). In step S9, if j ≦ D (step S9 Yes), the process returns to step S4. If j> D (No at step S9), the process proceeds to step S10.

In step S10, j = 0 and l = 1 + 1 are set (step S10). Then, it is determined whether or not l ≦ N _s (step S11). If l ≦ N _s (Yes in step S11), the process returns to step S4. If l> N _s (No in step S11), the calculation is terminated (step S12).

  Next, a specific calculation example is shown. The transmission path matrix H (bold) is shown in the following formula (25). The transmission path parameters are D = 2 and L = 2. In the case of the example shown in Expression (25), the size N of the matrix subject to Cholesky decomposition is N = 4.

Contracted channel matrix A (bold) _s and A (bold) _s of Cholesky decomposition by resulting triangular matrix U (in bold) _s is given by the following equation (26).

  The copying operation in step S5 is performed as in the following equation (27).

  As a result, the result U (bold) of the Cholesky decomposition of the matrix A (bold) is obtained as the following equation (28).

  In the present embodiment, an example of performing rearrangement when performing column deletion of the transmission line matrix described in the first embodiment has been described. You may perform the Cholesky decomposition method of the form. The operations of the present embodiment other than those described above are the same as those of the first embodiment.

As described above, in the present embodiment, the contracted transmission line matrix A (bold) _s is obtained, U _s (bold) is obtained by A (bold) _s Cholesky decomposition, and H is used using U _s (bold). (Bold) U (bold) obtained by Cholesky decomposition of ^H H (bold). Therefore, Cholesky decomposition can be performed with a small amount of computation in an environment where a delayed wave having a long delay time exists.

  As described above, the demodulator and the demodulation method according to the present invention are useful for a digital communication system under a multipath environment, and are particularly suitable for a digital communication system under an environment where a delayed wave having a long delay time exists. .

DESCRIPTION OF SYMBOLS 1 Transmitter 2 Receiver 3-1 and 3-2 Reflector W1 Direct wave 11 Symbol encoder 12 Transmission antenna 13 Reception antenna 14 Reception filter 15 Demodulator W2-1, W2-2 Interference wave 21, 22 Data 151 Transmission path calculation Unit 152 column deletion unit 153 demodulation processing unit

Claims (10)

A transmission path calculation unit that calculates a transmission path matrix by performing transmission path estimation based on the received signal;
A column deletion unit that deletes a predetermined number of columns from the top of the transmission line matrix;
A demodulation processing unit that performs demodulation processing based on the received signal and the transmission path matrix that has been deleted by the column deletion unit;
A demodulator. The received data is rearranged by moving the data of a predetermined size at the end of the data frame to the area immediately before or after the data of the predetermined size at the head of the same data frame, and the column deletion unit A rearrangement unit that performs rearrangement corresponding to rearrangement of the received signals with respect to the transmission path matrix that has been deleted;
Further comprising
The demodulation processing unit performs demodulation processing based on the received signal rearranged by the rearrangement unit and the transmission path matrix.
The demodulator according to claim 1. A transmission path calculation unit that calculates a transmission path matrix by performing transmission path estimation based on the received signal;
The rearrangement of the transmission path matrix is performed on the received signal by rearranging the divided matrix, which is a square partial matrix excluding the end area of the transmission path matrix, in the upper left area of the rearranged matrix. A reordering unit that performs reordering corresponding to
A division decomposition matrix which is a matrix obtained by Cholesky decomposition of the division matrix is obtained, a Cholesky decomposition result of the transmission line matrix is obtained based on the division decomposition matrix, and the received signal rearranged by the rearrangement unit; A demodulation processing unit that performs a demodulation process based on the Cholesky decomposition result,
A demodulator. A transmission path calculation unit that calculates a transmission path matrix by performing transmission path estimation based on the received signal;
Obtaining a reduced matrix obtained by reducing the transmission line matrix; obtaining a reduced decomposition matrix that is a matrix obtained by performing Cholesky decomposition on the reduced matrix; obtaining a Cholesky decomposition result of the transmission line matrix based on the division decomposition matrix; A demodulation processing unit that performs demodulation processing based on the signal and the Cholesky decomposition result;
A demodulator. Determining a matrix size of the reduced matrix based on a delay time;
The demodulator according to claim 4. The demodulation process is a Branch and Bound demodulation process.
The demodulator according to claim 1, wherein the demodulator is a demodulator. A transmission path calculation step of calculating a transmission path matrix by performing transmission path estimation based on the received signal;
A column deletion unit that deletes a predetermined number of columns from the top of the transmission line matrix;
A demodulation step for performing demodulation processing based on the received signal and the transmission path matrix that has been deleted by the column deletion unit;
The demodulation method characterized by including. A transmission path calculation step of calculating a transmission path matrix by performing transmission path estimation based on the received signal;
Reorder the received signal to move the data of a predetermined size at the end of the data frame to the area immediately before or after the data of the predetermined size at the top of the same data frame, A rearrangement step for performing rearrangement corresponding to the rearrangement of the received signals;
A demodulation step for performing demodulation processing based on the received signal rearranged by the rearrangement unit and the transmission path matrix;
The demodulation method characterized by including. A transmission path calculation step of calculating a transmission path matrix by performing transmission path estimation based on the received signal;
The rearrangement of the transmission path matrix is performed on the received signal by rearranging the divided matrix, which is a square partial matrix excluding the end area of the transmission path matrix, in the upper left area of the rearranged matrix. A reordering step for performing reordering corresponding to
A division decomposition matrix which is a matrix obtained by Cholesky decomposition of the division matrix is obtained, a Cholesky decomposition result of the transmission line matrix is obtained based on the division decomposition matrix, and the received signal rearranged by the rearrangement unit; A demodulation step for performing a demodulation process based on the Cholesky decomposition result,
The demodulation method characterized by including. A transmission path calculation step of calculating a transmission path matrix by performing transmission path estimation based on the received signal;
Obtaining a reduced matrix obtained by reducing the transmission line matrix; obtaining a reduced decomposition matrix that is a matrix obtained by performing Cholesky decomposition on the reduced matrix; obtaining a Cholesky decomposition result of the transmission line matrix based on the division decomposition matrix; A demodulation step for performing a demodulation process based on the signal and the Cholesky decomposition result;
The demodulation method characterized by including.

Images