JP2003152672A - Method and device for demodulating multi-carrier - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a multi-carrier demodulator in which an error rate is not considerably deteriorated even when a delayed wave having a delay time difference exceeding a guard interval arrives. SOLUTION: In a use symbol extracting part 31, M' points (L<=M'<(NTGI/T) are extracted from the end of a valid symbol from N+(NTGI/T) compound digital signals. In such a case, T+TGI-(M'T/N) is greater than the time difference from the latest delayed wave. In a first matrix operation and linear operation part 321, A general inverse matrix (L rows and M' columns) of a matrix (L<M') of M' rows and L columns, in which (q) rows and (p) columns are exp(2πjkp nq /N), is found from a set kp } of numbers of L subcarriers and a set nq } of sampling number of M' points, and a column vector composed of a complex signal X(kp1 ) of L' subcarriers is provided by calculating the matrix of L' rows and M' columns extracting rows corresponding to the set kp1 } of numbers of L' pilot carriers. A waveform x'(nq ) of L' pilot carriers is generated and a complex signal X(kp2 ) of remaining L-L' subcarriers is found from M' complex numbers x(nq )-x'(nq ).

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a multicarrier demodulation method and a multicarrier demodulation device. The present invention is particularly effective for OFDM reception in a place where the influence of delayed waves is large.

[0002]

2. Description of the Related Art For example, in an OFDM modulation system, a waveform called a guard interval is added before an effective symbol in order to prevent collapse of orthogonality due to superposition of delayed waves. For this guard interval, for example, the trailing quarter of the effective symbol is added, and one symbol is multiplied by 5/4 so that delayed waves having a guard interval length or less are not affected during demodulation. At this time, by multiplying the "window" of the effective symbol length, the waveform for the effective symbol length is used for demodulation.

In recent years, the Multimedia Mobile Access Promotion Council was established to develop a broadband mobile communication system comparable to a broadband fiber to the home. MMAC (Multimedia Mob) advocated by this council
ile access communication system)
For band mobile access, we propose a carrier arrangement with 52 subcarriers (of which 4 are pilot carriers). This is shown in FIG.

FIG. 4 is a diagram showing the positions of subcarriers adopted in MMAC and pilot carriers in them. The carrier number k is an integer of −31 ≦ k ≦ 32, and the subcarrier with k = 0 is a so-called DC carrier. In MMAC, k = -21, -7, 7, 21 are pilot carriers (thick arrows in FIG. 4, P _{C1 to} P _C4 ).
Then, the receiving side sequentially monitors the amplitude and phase information. k = −31 to −27, 0, 1 of 27 to 32
The two carriers are null carriers, which are a null carrier N _GB1 , N _GB2 and a DC null carrier N _DC which form a guard band. These are indicated by dashed arrows in FIG. The remaining 48 solid-line arrows are subcarriers that are not pilot carriers and are actually used for data transmission.

[0005]

When a delayed wave having a delay time difference exceeding the guard interval arrives due to the influence of multipath, the error rate is greatly deteriorated. In such a case, it is necessary to increase the guard interval length, but this increases communication redundancy and results in deterioration of communication efficiency.

[0006] By the way, in OFDM, a carrier having a null carrier is often used even when a band of N carriers as shown in Fig. 4 is used. Further, since the pilot carrier can easily obtain the phase information, the inventor of the present application has conceived that the separation and demodulation of the subcarriers for the remaining actual data transmission can be executed more reliably by separating and demodulating only the pilot carrier first. .

Therefore, the present invention focuses on the fact that all the effective carriers can be demodulated from a shorter symbol length in a multicarrier communication having a null carrier and a pilot carrier. An object of the present invention is to provide a demodulation method and a demodulation device in which an error rate is not significantly deteriorated even under the influence of a multibus in which a delayed wave having a delay time difference exceeding a guard interval arrives by separating and demodulating an effective carrier without using it. To do.

[0008]

In order to solve the above-mentioned problems, according to the means of claim 1, the effective symbol length is T and the adjacent frequency interval of N subcarriers is 1 / T.
, N−L (L ≦ N) subcarriers are null carriers, and pilot carriers of L subcarriers are L ′ (L ′ <L). In a multi-carrier demodulation method that separates and demodulates into subcarriers, the delay time difference of the delay wave is estimated, and based on the estimated delay time difference, the effective symbol length T is adjusted so as not to include the portion where the waveform distortion due to the delay wave is generated. Length T
A part to be used symbols of M / N (L ≦ M ≦ N) is determined, and from the complex digital signal orthogonally demodulated at the sampling interval T / N, using the used symbol part M points, L ′
And demodulating the pilot pilot carriers of the present number, deleting the components of the L'number of pilot carriers from the part to be used symbols before demodulation using the L'number of pilot carriers that have been separated and demodulated, and From the part which becomes the used symbol before demodulation in which the component of
It is characterized in that L ′ subcarriers are separated and demodulated. Here, the separating and demodulating means may be one by the following complex matrix calculation, or may be Japanese Patent Application No. 2001-298078 which is another application of the inventor of the present application.
In addition, when L = N, it goes without saying that the communication uses all subcarriers without a null carrier. The same applies to the means of claim 2.

According to the second aspect of the invention, the effective symbol length is T, the guard interval length added before the effective symbol length is T _GI , and the adjacent frequency interval of N subcarriers is 1. / T, a multi-carrier modulated signal in which N−L (L ≦ N) subcarriers are null carriers and L ′ subcarriers (L ′ <L) are pilot carriers among the L subcarriers are received. In a multi-carrier demodulation method that separates and demodulates each subcarrier, the delay time difference of the delay wave is estimated, and based on the estimated delay time difference, the effective symbol length is adjusted so that the portion in which the waveform distortion due to the delay wave is generated is not included. After determining the use symbol portion of the length from the sum T + T _GI of the guard interval length TM '/ N (L ≦ M ' ≦ N + (NT GI / T)) and, orthogonally demodulated by the sampling interval T / N complex From the digital signal, L'pieces of pilot carriers are separated and demodulated by using the use symbol portion M'points, and L'is used for the use symbol portion M'points before demodulation using the separated and demodulated L'pieces of pilot carriers. The present invention is characterized in that the remaining LL subcarriers are separated and demodulated from the used symbol portion M'before demodulation in which the components of'the pilot carriers are deleted and the components of the L'carriers are deleted. To do.

According to the third aspect of the invention, the effective symbol length is T, the frequency interval between N subcarriers adjacent to each other is 1 / T, and NL (L≤N) subcarriers. Is a null carrier and a pilot carrier of L subcarriers has L ′ lines (L ′ <L), and a multicarrier demodulation device that separates and demodulates into each subcarrier is sampled at a sampling interval. T /
A quadrature demodulation and sampling unit for obtaining N complex digital signals quadrature-demodulated by N, a delay time difference estimation unit for estimating a delay time difference of a delay wave, and a waveform distortion due to the delay wave from the delay time difference of the delay time difference estimation unit. Of the complex digital signals of N (L ≦ M ≦ N) out of the N complex digital signals so as not to include the portion where And a first matrix operation unit for calculating a complex matrix of L ′ rows and M columns, which is a linear operation expression for separating and demodulating L ′ pilot carriers using M complex digital signals. A first linear calculation for separating and demodulating L ′ pilot carriers by multiplying the complex matrix of L ′ rows and M columns obtained by the matrix calculation unit and a column vector of length M by M complex digital signals Section, L ′ pilot carriers separated and demodulated by the first linear operation section, and M
Waveform component removing unit for generating new M complex digital signals composed of the waveforms of the remaining LL 'subcarriers in which the influence of the L'pilot carriers has been eliminated using the complex digital signals. And using the new M complex digital signals output from the waveform component removing unit, the remaining L-
A second matrix calculation unit that calculates a complex matrix of L−L ′ rows and M columns, which is a linear calculation formula for separating and demodulating L ′ subcarriers, and L− obtained by the second matrix calculation unit. L'row M
A second linear operation unit for separating and demodulating LL 'subcarriers by multiplying a complex matrix of columns and a new M-column vector of M complex digital signals. .

According to a fourth aspect of the present invention, in the multicarrier demodulating apparatus according to the third aspect, M is fixed and a complex matrix of L'rows and M columns is used instead of the first matrix operation unit. A first storage unit that stores a matrix, a second storage unit that stores a complex matrix of L−L ′ rows and M columns in place of the second matrix calculation unit, and the first linear calculation unit is the first storage unit. Is used, and the second linear operation unit uses the complex matrix of L−L ′ rows and M columns output from the second storage unit. And

Further, according to the means of claim 5, the effective symbol length is T, the guard interval length added before the effective symbol length is T _GI , and the adjacent frequency interval of N subcarriers is 1. / T, a multi-carrier modulated signal in which N−L (L ≦ N) subcarriers are null carriers and L ′ subcarriers (L ′ <L) are pilot carriers among the L subcarriers are received. In a multi-carrier demodulation device that separates and demodulates each subcarrier, a quadrature demodulation and sampling unit that obtains N + (NT _GI / T) complex digital signals that have been quadrature demodulated at a sampling interval T / N,
The delay time difference estimator that estimates the delay time difference between the delay waves, and the delay time difference of the delay time difference estimator is such that N + (NT _GI
/ T) of the complex digital signals, M '(L≤M'≤N + (NT _GI / T)) complex digital signals are extracted as used symbols, and an output of the used symbol extractor. Calculating a complex matrix of L ′ rows and M ′ columns, which is a linear arithmetic expression for separating and demodulating L ′ pilot carriers using M ′ complex digital signals
Matrix calculator, and the L'row M'column complex matrix obtained by the first matrix calculator and the column vector of length M'by M'complex digital signals are multiplied by L'pilots. A first linear operation unit that separates and demodulates the carrier, and L ′ pilot carrier signals that are separated and demodulated by the first linear operation unit and M ′ complex digital signals are used to generate L ′ pilot carrier signals. The waveform component removing unit that generates new M ′ complex digital signals composed of the waveforms of the remaining L−L ′ subcarriers whose influence has been deleted, and the new M ′ number of outputs of the waveform component removing unit. A second matrix operation unit for calculating a complex matrix of L-L 'rows and M'columns, which is a linear operation formula for separating and demodulating the remaining LL' subcarriers using a complex digital signal; L− obtained by the matrix operation unit of 2
A second linear operation unit for separating and demodulating LL 'subcarriers by multiplying a complex matrix of L'row M'columns and a column vector of length M'by new M'complex digital signals; It is characterized by having.

Further, according to the means of claim 6, in the multicarrier demodulating device of claim 5, M '
Is fixed, the first matrix operation unit is replaced with a first storage unit that stores a complex matrix of L ′ rows and M ′ columns, and the second matrix operation unit is replaced with LL ′ rows M ′. Second memorized complex matrix of columns
, The first linear operation unit uses a complex matrix of L ′ rows and M ′ columns output from the first storage unit, and the second linear operation unit outputs L−L output from the second storage unit. It is characterized in that a complex matrix of'row M'columns is used.

[0014]

In order to demodulate L effective carriers (L <N) of N OFDM-modulated subcarriers, a general inverse matrix to be described later should exist, and at least L points or more. Need sampling points of. However, it is not always necessary to use N points. Then, if a delayed wave larger than the guard interval arrives,
The influence of the waveform distortion caused by the delay time is, for example, M points from the end of the effective symbol (L ≦ M <N).
It can be said that it is possible to demodulate L effective carriers (L <N) by using the M sampling points if the sampling points are not reached. That is, it is possible to remove the influence of the waveform distortion due to the delayed wave before separating and demodulating the effective carrier. This is ideal because it is also affected by noise. Thus, it is possible to provide a demodulation method and a demodulation device in which the error rate does not significantly deteriorate even when a delayed wave that exceeds the guard interval arrives. At this time, by obtaining the subcarriers called pilot carriers (L 'lines) together with the information such as the phase by using M sampling points first, the separation and demodulation of the remaining LL' subcarriers is more accurate. become. That is, according to the present invention, by separating and demodulating in two steps from M points (L ≦ M <N) sampling points, L ′ pilot carriers having high reliability due to equalization are first separated and demodulated, and L ′ pilot carriers are then demodulated. After removing the component of the pilot carrier, the information (phase etc.) of the L ′ pilot carriers is used in the separation and demodulation of the remaining LL ′ subcarriers (claims 1, 2).
3, 4, 5, 6). When only delayed waves smaller than the guard interval arrive, the influence of noise can be suppressed by using more sampling points than N sampling points of the effective symbol (claims 2, 5). , 6). Of these, if the M point or M'point is fixed during a period in which the arrival state of the delayed wave does not change significantly, such as in the same packet in packet transmission, the complex matrix should be stored in advance instead of matrix calculation. Then, the calculation time in the demodulator (time lag from reception to monotone)
Can be shortened. Since the complex matrix is uniquely determined by the values of N and M, it is stored in advance,
It can be read out and used according to the value of M.

[0015]

DESCRIPTION OF THE PREFERRED EMBODIMENTS First, a general inverse matrix will be described. In the OFDM system carrier, the nth point (n is 0
Consider that the waveform x (n) with the integers from 1 to N-1 satisfies the following expression (1). Note that X (k) is the k-th (k
Is a symbol included in subcarriers of 0 to N−1). Further, both x (n) and X (k) are complex numbers.

[Equation 1]

^Equation 1 is, so to speak, a column vector consisting of N complex numbers x (n), an N ^× N matrix whose n + 1 × k + 1 column component is W _N ^-kn , and N complex numbers X (k). It is a product of column vectors consisting of. Considering the left side of ^Equation 1 as a vector and the right side as a product of a matrix and a vector, an inverse matrix of an N-by-N matrix in which an ^element at n + 1 row and k + 1 column is W _N ^-kn, that is, k +
If an N-row N-column matrix whose 1-row n + 1-column component is W _N ^kn is multiplied from both sides of Equation 1 from the left, the matrix and N complex numbers x
The product of column vectors of (n) yields a column vector of N complex numbers X (k). This is an N-point inverse discrete Fourier transform (IDF) on the modulation side that is usually used in OFDM.
T) and the N-point discrete Fourier transform (DFT) on the demodulation side.

Now, assuming that there are N−L (L <N) null symbols (where X (k) = 0) out of N subcarriers, the right side of Equation 1 is a matrix of N rows and L columns. And X (k) is 0
Is a product of column vectors consisting of L complex numbers X (k _p ). Therefore, further focusing on only M (L ≦ M <N, for example, an integer n _q from N−M to N−1) out of N complex numbers x (n) showing a waveform, the M number of complex numbers x (n _q )
^Will be the product of a column vector consisting of L complex numbers X (k _p ) and a matrix of M rows and L columns of q _{N and} p columns of W _N ^-kpnq . That is, it is as follows.

[Equation 2]

The q-th row and the p-th column are W_N ^-kpnqM-by-L matrix of
If the rank of A is L, then (A^*A)^-1A^*Obtained as
There is a matrix of L rows and M columns (general inverse matrix). However, A
^*Is a conjugate transposed matrix of A, has L rows and M columns, and A^*A is
It is a matrix of L rows and L columns. The left side of Equation 2 is a vector, and the right side is
Considering the product of matrix and vector, the general inverse matrix (A^*A)
^-1A^*Multiplying both sides of Equation 2 from the left, the matrix and M
Complex number x (n_q), The product of the column vectors
Prime X (k_p) Is obtained. This is the application
It is a linear operation on the demodulation side used in the invention. Also, general retrograde
Row (A^*A)^-1A^*Is the matrix operation of the present invention.
It That is, (A^*A)^-1A^*Seeking
Therefore, M complex numbers x (n_qColumn vector consisting of
To Le (A ^*A)^-1A^*Take from the left.

Even when the rank (rank) of the matrix A of M rows and L columns of W _N ^-kpnq whose q rows and p columns is smaller than L, the matrix of L rows and M columns is uniformly set as follows. And the matrix and M
The product of column vectors of x complex numbers x (n _q ) yields a column vector of L complex numbers. Assuming that the rank (rank) of the matrix A having M rows and L columns is r (r ≦ L), the matrix A is calculated by using an appropriate unitary matrix U (where M rows and M columns) and V (where L rows and L columns). It is possible to perform "diagonalization" such as (singular value decomposition theorem or Autonne-Eckart-Young theorem).

[Equation 3]

Where σ ₁ , σ ₂ , ..., σ _r are positive and M
The non-zero singular values of the matrix L of row L columns and the non-zero eigenvalues of the matrix A ^* A of L rows and L columns are σ ₁ ² , σ ₂ ² , ..., σ _r ² . Therefore, the general inverse matrix A ⁺ is obtained as follows. A ⁺ is a matrix of L rows and M columns.

[Equation 4]

As described above, if the general inverse matrix A ⁺ of the equation 4 is used, even if the rank (r) r of the matrix A of M rows and L columns is not equal to L, the general inverse matrix A ⁺ and M complex numbers are used. The product of column vectors of x (n _q ) yields a column vector of L complex numbers. This can be thought of as obtaining an approximate solution of L complex numbers X (k _p ) by the method of least squares. Further, in the same manner as the above description, when the L complex numbers X (k _p ) are obtained from the M ′ waveforms exceeding N, the general inverse matrix can be obtained.

[First Embodiment] FIG. 1 shows a concrete first embodiment of the present invention.
3 is a block diagram showing a configuration of a multicarrier demodulation device 100 according to the embodiment of FIG. Multicarrier demodulator 100
Is an orthogonal demodulation / sampling unit 10, a preamble extraction unit 21, a synchronization establishment unit 22, a delay time difference estimation unit 23, a subcarrier phase / amplitude estimation unit 24, a used symbol extraction unit 31, a first matrix calculation / linear calculation unit 321. , A pilot carrier waveform generator 33, a pilot carrier waveform remover 34, a second matrix calculator / linear calculator 322, a channel characteristic equalizer 25, and a symbol determiner 26. Hereinafter, in this embodiment, the preamble (pilot symbol)
2 illustrates a demodulation device that demodulates data from an OFDM modulated wave having a guard interval. The number of carriers is N
The number of valid carriers is L (L <N) and the number of pilot carriers is L ′ (L ′ <L).

This embodiment corresponds to a concrete embodiment of claims 1 to 3 and 5. Correspondence between the configuration of the multicarrier demodulation device 100 and the configurations of claims 3 and 5 is as follows. That is, the first matrix calculation and linear calculation unit 321
Is the first matrix operation unit and the first linear operation unit, the pilot carrier waveform generation unit 33 and the pilot carrier waveform removal unit 34 are the waveform component removal units, and the second matrix operation and linear operation unit 322 is the second It corresponds to the matrix operation unit and the second linear operation unit, respectively.

The quadrature demodulation and sampling section 10 forms a so-called in-phase component I and quadrature component Q digital signal sequence. That is, the effective symbol length is T, and the guard interval length added before the effective symbol length is T _GI , and orthogonal demodulation is performed at the sampling interval T / N N + (N
The real and imaginary parts of the T _GI / T) complex digital signals. Of this output, the preamble (pilot signal)
Is detected by the preamble extraction unit 21. In this way, the synchronization establishing unit 22 synchronizes the entire demodulator from the information of the symbol section including the preamble (pilot signal). Further, the delay time difference estimation unit 23 detects the time difference with the latest delayed wave. Further, the subcarrier phase / amplitude estimation unit 24 detects the phase / amplitude information of each subcarrier and outputs the information for equalization.

According to the information on the time difference between the synchronization signal of the synchronization establishing section 22 and the latest delayed wave of the delay time difference estimating section 23,
The used symbol extraction unit 31 extracts a used symbol from the digital signal sequence of the in-phase component I and the quadrature component Q as follows, for example. That is, M ′ points (L ≦ M ′ <N + (NT _GI / T)) are extracted from the end of the effective symbol from the real and imaginary parts of the N + (NT _GI / T) complex digital signals. Here, T + T _GI − (M′T / N) is larger than the time difference from the latest delayed wave.

Next, the M'pieces of the used symbol extractor 31 are
Complex number x (n_q) And the slowest delay of the delay time difference estimation unit 23
Information on the time difference from the wave is the first matrix calculation and linear calculation unit 3
21 is output. First matrix calculation and linear calculation unit 32
In 1, a set of numbers of L effective carriers {k_p}set
And the set of sampling numbers of M'points {n_q} To line q
p columns are exp (2πjk_pn_q/ N) M'row L matrix A (L ≤
Compute the generalized inverse matrix (L row, M'column) of M '). L <M '
Therefore, A^*Is the conjugate transposed matrix of A, and (A^*A) ^-1
A^*Obtained as. In addition, the rank of the matrix A of M ′ rows and L columns is L
The following cases are based on the above singular value decomposition theorem. Got next
Number of L'pilot carriers in the generalized inverse matrix
Set of {k_p1}, And extract the row corresponding to
Find the matrix. Thus, the matrix of the L'row M'column is M '
Complex numbers x (n_q), L'pilot carrier
A complex signal X (k_p1) Is obtained. Thus the slowest
From the section that is not affected by the delayed wave,
Carrier (carrier number {k_p1}) Can be demodulated
It will be possible.

The L ′ pilot carrier complex signals X (k _p1 ) are output to the pilot carrier waveform generator 33. The pilot carrier waveform generation unit 33 generates x ′ (n _q ) which is the waveform of the pilot carrier at the M ′ point from the complex signal X (k _p1 ) of the pilot carrier. next,
In the pilot carrier waveform removal unit 34, the used symbol extraction unit 31 outputs an M′-order complex vector indicating x ′ (n _q ), which is the waveform of the pilot carrier at the point M ′ output from the pilot carrier waveform generation unit 33. By subtracting from the M′-th order complex vector indicating M ′ complex numbers x (n _q ), the waveform of the pilot carrier at the M ′ point is removed and the waveform of the LL ′ subcarriers is removed. A certain number of M ′ complex numbers x (n _q ) −x ′ (n _q ) are obtained.

Next, the pilot carrier waveform removing unit 34
Information of the time difference between the M ′ complex numbers x (n _q ) −x ′ (n _q ) output from the delay time difference estimation unit 23 and the slowest delayed wave is
Is output to the matrix operation and linear operation unit 322. In the second matrix operation / linear operation unit 322, q rows and p2 columns are exp from the set {k _p2 } of the numbers of the L−L ′ subcarriers and the set {n _q } of the sampling numbers of M ′ points. (2πjk _p2 n _q /
The general inverse matrix (LL ′ row × M ′ column) of the matrix A (LL ′ <M ′) of M ′ row and LL ′ column of N) is calculated. L-L '<because by M a A ^* as a conjugate transpose matrix of A, obtained as ^{* (A * A) -1 A} . In this way, the inverse matrix is multiplied by M ′ complex numbers x (n _q ) to obtain a complex signal X (k _p2 ) of LL ′ subcarriers. In this way, it is possible to demodulate LL 'effective carriers from the section that is not affected by the slowest delayed wave. The complex signal X (k _p2 ) of the L−L ′ subcarriers is output to the channel characteristic equalization unit 25, and the subcarrier phase / amplitude estimation unit 24 outputs the phase / amplitude information of each subcarrier. After the equalization process is performed, the symbol determiner 26 outputs the demodulated data signal sequence.

FIG. 2 shows the operation of the multicarrier demodulation device 100. As shown in (a), when a delayed wave that exceeds the guard interval arrives with respect to the preceding wave, it is possible to demodulate L subcarriers by using symbols that are shorter than the effective symbol length. Further, as shown in (b), when a delayed wave that does not exceed the guard interval arrives with respect to the preceding wave, a used symbol longer than the effective symbol length causes
It becomes possible to further improve the S / N ratio.

FIG. 3 shows the effect of the simulation of the multicarrier demodulation device 100. The subcarriers of L = 52 in the arrangement of FIG. 4 in which the DC carrier is null (of which,
L '= 4 lines are pilot carriers) and T _GI = T /
4. QPSK for subcarrier modulation, equal power for delayed wave and desired wave, power per bit / noise power of 30d
B, the product of Doppler frequency and symbol length due to movement is 0.0
[0004] From FIG. 3, according to the present invention, a delayed wave that exceeds the guard interval by about T / 8 is generated (total of 3T / 8).
It can be seen that a demodulation device can be provided in which the error rate does not significantly deteriorate even when it arrives.

[Modification] The first matrix operation / linear operation unit 321 and the second matrix operation / linear operation unit 32 having the above configuration
Instead of 2, a set of sampling numbers {n _q } at _M'points is fixed, and a memory for storing the required complex matrix of L'row M'columns and complex matrix of LL'row M'columns is stored. It may be configured to include a first arithmetic unit and a second arithmetic unit that each have. This corresponds to the embodiments of claims 4 and 6 of the present application. The first arithmetic unit corresponds to the first storage unit and the first linear arithmetic unit, and the second arithmetic unit corresponds to the second storage unit and the second linear arithmetic unit.

In addition to the configurations shown in the above-mentioned embodiment and modification,
The delay time difference estimation unit 23 may output time difference information (delay dispersion) considering the strength of the delayed wave as information on the time difference from the latest delayed wave. Subcarrier modulation is QP
The present invention is not limited to SK, and the present invention can be applied to those using other modulation methods such as PSK or QAM.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a multicarrier demodulation device 100 according to a specific embodiment of the present invention.

FIG. 2 is a conceptual diagram showing the operation of the multicarrier demodulation device 100, (a) shows a case where a delayed wave exceeding the guard interval arrives, and (b) shows a case where a delayed wave does not exceed the guard interval arrives.

FIG. 3 is a simulation result diagram showing a relationship between a delayed wave and an error rate of the multicarrier demodulation device 100.

FIG. 4 is a conceptual diagram showing an example of an arrangement of pilot carriers.

[Explanation of symbols]

10 Quadrature demodulation and sampling section 21 Preamble extractor 22 Synchronization establishment section 23 Delay time difference estimator 24 Subcarrier phase / amplitude estimation unit 25 Channel characteristic equalizer 26 symbol determination unit 31 Used Symbol Extraction Unit 321 First Matrix Calculation and Linear Calculation Unit 322 Second matrix calculation and linear calculation unit 33 Pilot carrier waveform generator 34 Pilot Carrier Waveform Removal Unit

Claims (6)

[Claims] 1. An effective symbol length of T, N adjacent subcarriers having a frequency interval of 1 / T, and N−L (L ≦ L
N) subcarriers are null carriers, and among L subcarriers, L ′ pilot carriers (L ′ <L).
In a multi-carrier demodulation method that receives a multi-carrier modulated signal and separates and demodulates it into each sub-carrier, the delay time difference of the delayed wave is estimated, and the waveform distortion due to the delayed wave is generated based on the estimated delay time difference. Effective symbol length T to length TM / N (L ≦ M ≦
N), the portion to be the used symbol is determined, and L'is used from the complex digital signal orthogonally demodulated at the sampling interval T / N by using the use symbol portion M points.
And demodulate the pilot pilot carriers of the present number, and remove the components of the L ′ pilot carriers from the portion to be the used symbols before demodulation by using the L ′ pilot carriers that have been separated and demodulated to obtain L ′ pilots. A multi-carrier demodulation method, characterized in that the remaining LL 'subcarriers are separated and demodulated from a portion which becomes the use symbol before demodulation in which carrier components are deleted. 2. The effective symbol length is T, the guard interval length added before the effective symbol length is T _GI , the adjacent frequency interval of N subcarriers is 1 / T, and NL (L ≦ L). N) subcarriers are null carriers, and among L subcarriers, pilot carriers are L ′ (L ′).
In a multi-carrier demodulation method of receiving a multi-carrier modulated signal of <L) and separating and demodulating each sub-carrier, a delay time difference of a delayed wave is estimated, and waveform distortion due to the delayed wave is estimated based on the estimated delay time difference. The sum T of the effective symbol length and the guard interval length so that the generated portion is not included
+ T _GI to length TM '/ N (L≤M'≤N + (NT _GI /
T)), the used symbol portion is determined, and then, from the complex digital signal orthogonally demodulated at the sampling interval T / N, using the used symbol portion M ′ point, L ′
And demodulating the pilot pilot carriers of the present number, and deleting the components of the pilot symbol carriers L'of the used symbol portion M'before demodulation using the demodulated and demodulated pilot carriers of the L'number of pilot carriers. A multicarrier demodulation method, characterized in that the remaining LL 'subcarriers are separated and demodulated from the used symbol portion M'before demodulation in which carrier components are deleted. 3. The effective symbol length is T, the adjacent frequency interval of N subcarriers is 1 / T, and NL (L ≦ L)
N) subcarriers are null carriers, and among L subcarriers, L ′ pilot carriers (L ′ <L).
In a multi-carrier demodulation device that receives a multi-carrier modulated signal that is and is separated and demodulated into sub-carriers, an orthogonal demodulation and sampling unit that obtains N complex digital signals that are orthogonally demodulated at a sampling interval T / N, and a delayed wave And a delay time difference estimator for estimating the delay time difference of M, and the delay time difference of the delay time difference estimator is used as M to be used symbols among the N complex digital signals so as not to include a portion in which the waveform distortion due to the delay wave is generated. Pieces (L
≦ M ≦ N) used symbol extraction unit for extracting a complex digital signal, and a linear operation for separating and demodulating L ′ pilot carriers using the M complex digital signals output by the used symbol extraction unit A first matrix operation unit for calculating a complex matrix of L ′ rows and M columns, which is an expression, a complex matrix of L ′ rows and M columns obtained by the first matrix operation unit, and a length of the M complex digital signals. A column vector of size M for separating and demodulating L ′ pilot carriers, and L ′ pilot carriers separated and demodulated by the first linear calculator and the M number of pilot carriers. A waveform component removing unit that uses the complex digital signal to generate new M complex digital signals composed of the waveforms of the remaining L−L ′ subcarriers in which the influence of the L ′ pilot carriers has been removed; Wave formation A complex matrix of LL ′ rows and M columns, which is a linear arithmetic expression for separating and demodulating the remaining LL ′ subcarriers using the new M complex digital signals output from the removing unit, is calculated. A second matrix operation unit, a complex matrix of L−L ′ rows and M columns obtained by the second matrix operation unit, and a length M based on the new M complex digital signals.
And a second linear operation unit for separating and demodulating LL ′ subcarriers by multiplying the column vector of 4. The multicarrier demodulator according to claim 3, wherein M is fixed, and the first storage unit stores a complex matrix of L ′ rows and M columns instead of the first matrix operation unit. , A second storage unit that stores a complex matrix of L−L ′ rows and M columns in place of the second matrix operation unit, and the first linear operation unit outputs L output from the first storage unit. 'A complex matrix of M rows is used, and the second linear operation unit outputs L output from the second storage unit.
A multicarrier demodulation device characterized by using a complex matrix of L ′ rows and M columns. 5. The effective symbol length is T, the guard interval length added before the effective symbol length is T _GI , the adjacent frequency interval of N subcarriers is 1 / T, and N−L (L ≦ L). N) subcarriers are null carriers, and among L subcarriers, pilot carriers are L ′ (L ′).
In a multi-carrier demodulation device that receives a multi-carrier modulated signal of <L) and separates and demodulates it into each sub-carrier, orthogonally demodulated at a sampling interval T / N of N + (NT _GI
/ T) Quadrature demodulation and sampling section for obtaining complex digital signals, delay time difference estimation section for estimating delay time difference of delay wave, and delay time difference of the delay time difference estimation section cause waveform distortion due to delay wave N + so as not to include the part
And use the symbol extractor for extracting complex digital signal (NT _GI / T) M 'number (L ≦ M' as used symbols of the number of complex digital signal _{≦ N + (NT GI / T} )), the use symbol extractor A matrix calculation unit for calculating a complex matrix of L'row M'columns, which is a linear calculation formula for separating and demodulating L'pilot carriers using the M'complex digital signals output by the unit And L ′ rows and M ′ columns of the complex matrix obtained by the first matrix operation section and the M ′ column vector of length M ′ by the M ′ complex digital signals are multiplied to separate L ′ pilot carriers. By using the first linear operation unit for demodulation, the L ′ pilot carriers separated and demodulated by the first linear operation unit, and the M ′ complex digital signals, the influence of the L ′ pilot carriers is obtained. Deleted and left LL 'subkeys A waveform component removing unit for generating new M'complex digital signals composed of a rear waveform, and the remaining LL 'lines using the new M'complex digital signals output from the waveform component removing unit. Second matrix operation unit for calculating a complex matrix of L-L 'rows and M'columns, which is a linear operation formula for separating and demodulating the subcarriers of, and L-L' obtained by the second matrix operation unit. A second linear operation unit for separating and demodulating LL 'subcarriers by multiplying the complex matrix of row M'columns and the column vector of length M'by the new M'complex digital signals. A multi-carrier demodulation device having. 6. The multi-carrier demodulator according to claim 5, wherein M is fixed, and a first storage unit that stores a complex matrix of L ′ rows and M ′ columns instead of the first matrix calculation unit is stored. And a second storage unit that stores a complex matrix of L−L ′ rows and M ′ columns in place of the second matrix calculation unit, and the first linear calculation unit outputs the first storage unit. Using a complex matrix of L ′ rows and M ′ columns, the second linear operation unit outputs L output from the second storage unit.
A multicarrier demodulation device characterized by using a complex matrix of L ′ rows and M ′ columns.