JP2003101503A - Ofdm equalizer and equalization method for ofdm - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an equalization apparatus, having high estimation accuracy of transmission line response with a reduced number of arithmetic operations. SOLUTION: In an equalizer 1 for the OFDM, a divider 2 divides a received OFDM signal Y(m, k) into a received pilot signal Y(m, kp ) and a data signal Y(m, kd ). A pilot pattern generator 3 generates a known complex amplitude X(m, kp ), and a divider 4 divides the received pilot signal Y(m, kp ) by X(m, kp ), to output a transmission line response Y(m, kp )/X(m, kp ). An IFFT computing element 5 performs IFFT (Inverse Fast-Fourier Transform) on the response Y(m, kp )/X(m, kp ), which is converted into an impulse response h0 . Noise is eliminated from the response h0 by a noise elimination filter 6, and thereafter up-sampling processing is performed in an up-sampling portion 7. Thus the obtained signal is converted into an impulse response h2 . An FFT computing element 8 performs FFT (Fast-Fourier Transform) on the response h2 , to output an estimated transmission line response Hc (m, kd ). A divider 9 divides the data signal Y(m, kd ) by the estimated transmission line response Hc (m, kd ), to output an equalized data Y(m, kd )/Hc (m, kd ).

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to OFDM (Orthog
BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an equalizer for OFDM and an equalization method thereof for estimating a channel response of a transmission signal by an onal Frequency Divisoin Multiplex method and performing equalization processing.

[0002]

2. Description of the Related Art As a terrestrial digital broadcasting transmission system,
The OFDM system, which multiplex-transmits hundreds to thousands of carrier waves (subcarriers) within a band of one channel, has been adopted in Japan and Europe. The OFDM system is a multi-carrier modulation system in which transmission data is divided into a plurality of subcarriers and transmitted, and therefore has an advantage that frequency utilization efficiency is very high and it is resistant to frequency selective fading interference generated during mobile reception.

FIG. 15 shows a conventional OFDM receiver 10.
It is a functional block diagram which shows the outline of 0. RF (Radio Frequency) transmitted from an OFDM transmitter (not shown)
y) The signal 101 is received by the receiving antenna 102 through the transmission path, and then the IF (Intermediate) is received by the tuner 103.
Frequency) is converted into a signal. The IF signal is input to the mixer 102 via a BPF (bandpass filter) 104, is multiplied by the signal supplied from the carrier oscillator 103, and is then LPF (lowpass filter) 1.
It is output to 07. The LPF 107 outputs a signal obtained by removing unnecessary high-frequency components from the input signal to the A / D converter 108, and the A / D converter 108 directly converts the input signal into a digital signal (symbol signal) at a predetermined sampling frequency. Output to the parallel converter 109. And the serial-parallel converter 1
09 converts the input serial signal into a parallel signal and outputs the parallel signal to an FFT (Fast Fourier Transform) calculator 110.

The FFT calculator 110 Fourier transforms the input time domain symbol signal into a frequency domain received OFDM signal and outputs it to an equalizer 111 which performs waveform equalization on the received OFDM signal. The equalized data output from the equalizer 111 is converted from a parallel signal to a serial signal by the parallel-serial converter 112, and then the channel decoder 1
Viterbi decoding and Reed-Solomon decoding were applied in 13.
Next, in the source decoder 114, MPEG (Moving Pictu)
After being subjected to decoding such as re Experts Group) -2 system, it is converted to analog by the D / A converter 115 and output.

As described above, the OFDM system is a system for multiplex transmission of a large number of subcarriers, and a modulation system such as QPSK, DQPSK, or multilevel QAM can be adopted for each subcarrier, and different for each subcarrier. Modulation schemes can be used. According to the Japanese terrestrial digital broadcasting standard, one transmission symbol is divided into multiple segments.
A different modulation method is used for each segment. Also,
When multipath interference occurs on the transmission line, the amplitude and phase change for each subcarrier due to frequency selective fading. Generally, the subcarrier modulation method is DQPSK
If such a differential modulation method is used, the FFT calculator 11
Since the signal after FFT execution at 0 can be differentially demodulated, it can be demodulated without waveform equalization, but multi-valued QAM can be used when it is desired to improve frequency utilization efficiency and send a large amount of information in a narrow band.
Is adopted. Since multi-level QAM puts information on the amplitude,
The equalizer 111 must correct the phase and amplitude of the received OFDM signal (waveform equalization) for each carrier. On the transmitting side, a pilot signal for referring to the degree of phase and amplitude distortion is embedded in a predetermined carrier position in the OFDM signal to be transmitted at a constant rate.

FIG. 16 is a diagram showing an arrangement example of data signals and pilot signals included in an OFDM signal, which is adopted in the European DVB-T system. In FIG.
The horizontal direction indicates the frequency direction (carrier direction), the vertical direction indicates the time direction (symbol direction), and each circle indicates a data signal to be placed on the kth subcarrier of the mth transmission symbol, CP (continuous pilot). ) Signal or SP (distributed pilot) signal. Since the receiving side has the arrangement information of each pilot signal, the CP signal and the SP signal can be extracted from the received OFDM signal. Also, as shown in FIG. 17, each transmission symbol of the OFDM signal is provided with a guard interval for reducing the influence of multipath between effective symbols for data transmission. Each guard interval is a copy of the last part of the effective symbol, and if the delay time τ of the reflected wave is within the guard interval period Tg, complete data for one transmission symbol that is not affected by inter-symbol interference (ISI) is extracted. be able to. Note that, normally, the guard interval period Tg is 1/4, 1/8, 1/16, 1 / of the effective symbol period Tu.
It is set to any of 32.

FIG. 18 shows an equalizer 11 for performing waveform equalization processing.
It is a functional block diagram which shows schematic structure of 1. The equalizer 111 receives OFDM from the FFT calculator 110.
The signal Y (m, k) (m: symbol number, k: carrier number) is input. Separator 111A is decomposed to the received OFDM signal Y (m, k) receive the data signal Y (m, k _d) and the received pilot signal Y (m, k _p) and, received pilot signal Y (m, k _p ) was output to the interpolation filter 111B, and outputs the received data signal Y (m, k _d) in the divider 111C. Since the interpolation filter 111B has a known complex amplitude X (m, k _p ) of the pilot signal to be embedded on the transmission side, the complex amplitude X (m, k _p ) and the received pilot signal Y (m, k _p ) are obtained. Based on this, the transmission path response H _c (m, k _d ) of the subcarrier transmitting the data signal is interpolated and estimated, and output to the divider 111C. The divider 111C receives the received data signal Y (m,
k _d) an estimated channel response H _c (m, k _d) divided by the equalized data _{Y (m, k d) /} H c (m, outputs the k _d).

The received pilot signal Y (m, k _p ) can be expressed by the following equation (1).

[0009]

[Equation 1]

In the above equation (1), H (m, k _p ) is the transmission path response of the subcarrier transmitting the pilot signal, X
(M, k _p ) is a known complex amplitude, and N (m, k _p ) is a noise signal mixed in the transmission path.

Pilot in the interpolation filter 111B
Signal transmission path response H (m, k_p) Is a simple public knowledge
The method is LS (Least Square) estimation.
It According to the LS estimation, the functional form of the received pilot signal Y
Is assumed to be Y = H · X + N (H_c: Transmission line response,
(X: known complex amplitude, N: noise signal). And pie
Estimated value H of lot signal transmission line response_c(M, k_p) Is
[H_c・ XY (m, k_p)] ²Determined to be the minimum
To be done. Therefore, the following equation (2) is established.

[0012]

[Equation 2]

Further, if the above equation (2) is further modified, the following equation (2A) is established.

[0014]

[Equation 3]

As can be seen from the above equation (2A), the estimated transmission line response H _c (m, k _p ) has a noise component N _c (m,
k _p ) = N (m, k _p ) / X (m, k _p ) is included, which is one of the causes for lowering the estimation accuracy of the channel response.

The transmission line response H calculated by the above equation (2)
_c (m, k _p ) is an estimated value of the channel response of the received pilot signal, not that of the data signal. Therefore, the transmission line response H _c (m, k _d ) of the data signal needs to be created by an interpolation filter based on the transmission line response H _c (m, k _p ) of the received pilot signal. In addition, since the transmission path response temporally changes with frequency, theoretically, there are two directions, the time direction (symbol direction) and the frequency direction (carrier direction).
Although a dimensional interpolation filter is required, it can be considered that the actual transmission line response fluctuates independently in the time direction and the frequency direction.Therefore, separate one-dimensional filters are used in the time direction and the frequency direction. Can be interpolated. The transmission line response in the time direction fluctuates relatively slowly, so a simple IIR (Infinite Impedance)
Sufficient estimation accuracy can be obtained by using the ulse response) type filter. On the other hand, the interpolation filter in the frequency direction is not so simple.

FIG. 19 is a schematic block diagram showing a conventional example of the frequency domain equalizer 111. As mentioned above, FFT
The arithmetic unit 110 transforms the symbol signal in the time domain into the received OFDM signal Y (m, k) in the frequency domain to equalize the signal.
Output to 1. In the equalizer 111, the serial / parallel circuit 12
0 converts the input reception OFDM signal Y (m, k) into a parallel signal and outputs the parallel signal, and the separator 121 converts the input signal into the reception pilot signal Y (m, k _p ) and the reception data signal Y.
(M, k _d ), where k _p ≠ k _d . The pilot pattern generator 122 has a known complex amplitude X (m, k
_p ) and output. Divider 123, received pilot signal Y (m, k _p) a is divided by its complex amplitude X (m, k _p), the estimated value H _c of the channel response of the received pilot signal (m, k _p) = Y (m, k _p ) / X (m, k _p ). Next, the symbol filter 124 smoothes (interpolates) the estimated value H _c (m, k _p ) in the time direction (symbol direction), and then the interpolation filter 125 performs interpolation processing in the frequency direction (carrier direction). Execution is performed, and the estimated value H _c (m, k _d ) of the transmission path response of the data signal is output. Then, the divider 126 divides the received data signal Y (m, k _d ) by the estimated transmission line response H _c (m, k _d ) to obtain equalized data Y.
Output (m, k _d ) / H _c (m, k _d ).

The estimation accuracy of the channel response mainly depends on the interpolation filter 125. As the interpolation filter 125, F
Although an IR (Finite Impulse Response) type low-pass filter can be used, there is a problem that high estimation accuracy cannot be obtained because the noise removal performance of this type of low-pass filter is low. Regarding the FIR type low-pass filter, for example, refer to “A Study on Performance Improvement of OFDM Equalizer in Fading Transmission Line” (Akio Yamamoto, Keizo Nishimura,
Journal of the Institute of Image Information and Television Engineers Vol.53, No.11, pp.1557-156
5, 1999).

Further, when the estimation accuracy is improved by taking noise into consideration, a Wiener filter having excellent noise removing performance can be adopted. Wiener
For the filter, see, for example, "Statistical Digi".
tal Signal Processing and Modeling "(Monson H. Haye
s, p.335, ISBN: 0-471-59431-8). However, the design of the Wiener filter requires parameters such as the autocorrelation function of the transmission path response and the variance value of noise, and the amount of calculation increases, which is currently difficult in practice.

[0020]

SUMMARY OF THE INVENTION In view of the above problems, the present invention is to solve the problems that an equalizer for OFDM and a method for equalizing the same are used, which have a high estimation accuracy of a channel response and a small amount of calculation. It is in the point of providing.

[0021]

In order to solve the above problems, the invention according to claim 1 is transmitted by an OFDM (Orthogonal Frequency Division Multiplexing) system for multiplex transmission of a plurality of subcarriers satisfying an orthogonal relationship with each other. An equalizer for OFDM for equalizing a received signal, the received OF obtained by transforming the received signal from a time domain to a frequency domain by Fourier transform.
Extracting means for extracting a received pilot signal and a data signal embedded in a known pattern from the DM signal;
Pilot pattern generating means for generating a known complex amplitude of the received pilot signal, and division means for dividing the received pilot signal by the complex amplitude to calculate the channel response of the subcarrier transmitting the received pilot signal, Inverse Fourier transform means for calculating an impulse response in the time domain by applying an inverse Fourier transform to the transmission path response output from the division means, and an up for adding a zero-value data string to the impulse response An up-sampling means for performing a sampling process, and a Fourier transform means for performing an Fourier transform on the impulse response output from the up-sampling means to calculate an estimated value of the channel response interpolated in the frequency direction,
Second dividing means for dividing the data signal by the estimated value of the transmission path response.

The invention according to claim 2 is the O according to claim 1.
An equalizer for FDM, further comprising noise removing means for removing a noise component from the impulse response calculated by the inverse Fourier transforming means, wherein the up-sampling means outputs the impulse output from the noise removing means. The up-sampling process is performed on the response.

The invention according to claim 3 is the invention according to claim 1 or 2.
In the OFDM equalizer described, the Fourier transform means has a fast Fourier transform function, and the inverse Fourier transform means has an inverse fast Fourier transform function.

The invention according to claim 4 is an equalization method for OFDM, which equalizes a reception signal transmitted by an OFDM (Orthogonal Frequency Division Multiplexing) method for multiplex transmission of a plurality of subcarriers satisfying an orthogonal relationship with each other. And (a) extracting a received pilot signal and a data signal embedded with a known pattern from the received OFDM signal obtained by transforming the received signal from the time domain to the frequency domain by Fourier transform, and (b) Generating a known complex amplitude of the received pilot signal; and (c) dividing the received pilot signal extracted in step (a) by the complex amplitude generated in step (b). A step of calculating a transmission line response of the subcarrier transmitting the pilot signal, and (d) an inverse Fourier transform with respect to the transmission line response calculated in the step (c). Calculating the impulse response in the time domain by performing the conversion, (e) performing up-sampling processing for adding a zero-value data sequence to the impulse response, and (f) the step (e) At (g) calculating an estimated value of the transmission path response interpolated in the frequency direction by applying a Fourier transform to the impulse response subjected to the up-sampling processing, Dividing the data signal by the estimated value.

The invention according to claim 5 is the O according to claim 4.
The FDM equalization method further comprises: (h) removing a noise component from the impulse response calculated in the step (d), wherein the step (e) includes the step (h).
Is a step of performing the up-sampling process on the impulse response from which the noise component has been removed.

The invention according to claim 6 is the invention according to claim 4 or 5.
In the described equalization method for OFDM, the inverse Fourier transform in the step (d) is an inverse fast Fourier transform, and the Fourier transform in the step (f) is a fast Fourier transform.

[0027]

BEST MODE FOR CARRYING OUT THE INVENTION An equalizer according to an embodiment of the present invention will be described below.

FIG. 1 is a block diagram showing a schematic configuration of the equalization apparatus 1 according to this embodiment. This equalizer 1 is
The integrated circuit is incorporated in the equalizer 111 of the OFDM receiver 100 as shown in FIG.
As shown in FIG. 1, the FFT calculator 110 converts the symbol signal in the time domain into the received OFDM signal Y (m,
k) (m: symbol number, k: carrier number) is Fourier-transformed and output to the equalization device 1 according to the present embodiment.
In this equalizer 1, reference numeral 2 is a separator, 3 is a pilot pattern generator, 4 is a divider, and 5 is an IFFT for performing an inverse transform (IFFT) of FFT (Fast Fourier Transform).
An arithmetic unit, 6 is a noise removal filter, 7 is an up-sampling unit, 8 is an FFT arithmetic unit that executes FFT, and 9 is a divider.

The separator 2 inputs the received OFDM signal Y
Separating the (m, k) receive the pilot signal Y (m, k _p) and the data signal Y (m, k _d) and (where, k _d ≠ k _p),
The separated received pilot signal Y (m, k _p ) is divided by a divider 4
And outputs the data signal Y (m, k _d ) to the divider 9. The pilot pattern generator 3 generates a known complex amplitude X (m, k _p ) of the pilot signal and outputs it to the divider 4. The divider 4 inputs from the separator 2 according to the following equation (3). The pilot signal Y (m, k _p ) is given a complex amplitude X
(M, k _p) estimated channel response of the pilot signal by dividing by H _c (m, k _p) and outputs a.

[0030]

[Equation 4]

Next, the IFFT calculator 5 executes IFFT on the transmission path response H _c (m, k _p ) input from the divider 4 to generate and output a time domain impulse response h ₀ . FIG. 2 is a diagram showing a graph of this impulse response h ₀ . The vertical axis of FIG. 2 corresponds to the amplitude of the impulse response h ₀ , and the horizontal axis corresponds to the time axis. The impulse response h ₀ is the case where noise is mixed in the transmission symbol on the transmission path.

Next, the noise removal filter 6 operates in the IFFT.
2 for the impulse response h ₀ output from the calculator 5
The impulse response h ₁ from which the noise component is removed is output by performing the binarization process. FIG. 3 is a diagram showing a graph of the impulse response h ₁ obtained by removing the noise component from the impulse response h ₀ shown in FIG. As shown in FIGS. 2 and 3, the impulse responses h ₀ and h ₁ both have two peaks near the times t = 0 and t = 10. Generally, in a transmission path, when the carrier amplitude is Rayleigh-distributed and Rayleigh fading in which the phase is uniformly distributed occurs, multiple waves including direct waves and reflected waves are received by the receiver. Path interruption occurs. According to the two-wave Rayleigh fading model that considers only two waves, the direct wave and the reflected wave, the impulse response h (t) of the transmission path can be approximately expressed by the following equation (4).

[0033]

[Equation 5]

In the above equation (4), α ₁ (t) is the Rayleigh amplitude of the direct wave, α ₂ (t) is the Rayleigh amplitude of the reflected wave, τ is the delay time of the reflected wave with respect to the direct wave, and δ (x) is Is the delta function for the variable x. Two impulse signals corresponding to the direct wave and the reflected wave respectively appear in the impulse responses h ₀ and h ₁ shown in FIGS. 2 and 3, and the transmission path can be represented by the two-wave Rayleigh fading model. I understand.

After calculating the impulse response h ₁ with the noise removed as described above, the up-sampling unit 7
The impulse response h ₁ is scaled by dividing it by the interpolation magnification L, and the up-sampling process of adding the data D ₀ consisting of a number of zero values according to the interpolation magnification L is executed to obtain the impulse response h ₂ . Output. Specifically, the data of the impulse response h ₁ is assumed to be composed of n (n ≧ 1) components as shown in the following equation (5).

[0036]

[Equation 6]

In the above equation (5), h ₁ (t) represents the value of the impulse response h ₁ at time t. Further, when the number of zero values forming the added data D ₀ = {0, 0, ..., 0} is represented by q, the impulse response h ₂ is calculated from n + q data as shown in the following equation (6). Composed.

[0038]

[Equation 7]

In the above equation (6), h ₂ (t) is the value of the impulse response h ₂ at time t. In FIG. 4, the up-sampling unit 7 has the impulse response h shown in FIG.
Is a diagram showing a graph of the impulse response h ₂ which has output then process _1.

Next, the FFT calculator 8 Fourier transforms the impulse response h ₂ output from the up-sampling unit 7 into a signal in the frequency domain, thereby interpolating the transmission line response in the frequency direction (carrier direction). Estimated value H _c (m,
output k _d ). Then, the divider 9 receives the data signal Y (m, k _d ) input from the separator 2 and the estimated transmission line response H _c.
The equalized data Y is obtained by multiplying the reciprocal of (m, k _d ) by
Output (m, k _d ) / H _c (m, k _d ).

FIG. 5 shows the impulse response h ₂ shown in FIG.
Estimated channel response H _c (m,
It is a graph of k _d). The vertical axis of FIG. 5 corresponds to the amplitude of the estimated transmission path response, and the horizontal axis corresponds to the carrier number k. Circles in the figure show data before interpolation, and stars in the figure show data after interpolation. On the other hand, FIG. 6 is a graph of the transmission line response H _c ′ (m, k _d ) obtained by performing FFT on the impulse response h ₁ shown in FIG. 3 and then interpolating it twice by an FIR low pass filter. FIG. Figure 6
The vertical axis of corresponds to the amplitude of the transmission line response H _c '(m, k _d ),
The horizontal axis corresponds to the carrier number k. The circles in the figure show the data before the interpolation, and the crosses in the figure show the data after the interpolation. Thus, it can be seen that FIGS. 5 and 6 show substantially the same interpolation result. Therefore,
(1) In order to obtain the result of FIG.
And interpolation method to execute the FFT after scaling by a zero value corresponding to the interpolation factor L with respect to _1, in order to obtain a (2) results in FIG. 6, F with respect to the impulse response h ₁
The method of performing FT and interpolating L times is equivalent. The interpolation according to (1) is called so-called "DFT interpolation". Its theoretical proof can be found in the article "Multirate Digital S
ignal Processing "(Ronald E. Crochiere, Lawrence
R. Rabiner, pp.35-39, ISBN: 0-13-605162-6).

According to the terrestrial digital broadcasting standard in Japan, as shown in FIG. 7, each transmission symbol of an OFDM signal has a hierarchical structure in which it is divided into a plurality of segments in the carrier direction (frequency direction), and synchronous modulation is performed. Segment (synchronous modulator) SM ₁ , ..., Differential modulation segment (differential modulator) DM ₁ , ..., Partial receiver PR ₁ ,. The partial reception section PR ₁ has a hierarchical structure including a single or a plurality of differential modulation sections DM ₁ and a synchronous modulation section SM ₁ . Further, FIG. 8 is a diagram showing a signal arrangement of the synchronous modulator. In FIG. 8, as in FIG. 16, the horizontal direction is the frequency direction (carrier direction), the vertical direction is the time direction (symbol direction), and each circle represents a data signal, SP.
Signal, TMCC (Transmission and Multiplexing Conf
igurationControl) signal and AC (Auxiliary Channel)
l) shows the signal. The TMCC signal is a signal for transmitting control information, and the AC signal is an extension signal for transmitting additional information. The arrangement of these TMCC signals and AC signals reduces the influence of multipath on the transmission line response. In order to do so, they are randomly arranged in the frequency direction (carrier direction). In the illustrated example, the carrier number k of one segment is 0 to 107, but the mode in which the carrier number k is 0 to 215 or the carrier number k is 0 to 43.
There is also a mode that takes 1.

Although the signal arrangement of the differential modulation section DM ₁ shown in FIG. 7 will not be described, the C arrangement shown in FIG. 16 is used.
The P signal is arranged in the differential modulator DM ₁ .

In this embodiment, it is used to estimate the transmission line response.
Used pilot signal Y (m, k _p) Is SP (Distributed
Illot) signal. The SP signal is PRBS
BPSK modulation using output bit string of random code sequence)
Is generated by. As shown in FIG. 8, the SP signal is
Once every 12 carriers in the carrier direction and 4 in the symbol direction
It is inserted cyclically once in NMBOL. others
Therefore, the delay of the reflected wave of the two-wave Rayleigh fading model
If the time τ is within 1/3 of the effective symbol length Tu,
It is possible to prevent characteristic deterioration due to occurrence of inter-signal interference.

For the synchronous modulator SM _{1 having} such a signal arrangement, the conventional equalizer 111 shown in FIG.
The symbol filter 124 interpolates the transmission path response in the symbol direction every two carriers. As a result, the transmission path response of the data signal of each column of carrier number k = 0, 3, 6, ... Is interpolated and estimated. Then the equalizer 111
The interpolation filter 125 interpolates and estimates the transmission path response in the carrier direction. As a result, carrier numbers k = 1, 2,
Transmission path response H of the data signal of each column of 4, 5, 7, ...
_c (m, k _d ) is interpolated. Therefore, the conventional equalizer 11
In No. 1, the serial / parallel circuit 120 buffers the input received OFDM signal Y (m, k) in order to perform the interpolation in the symbol direction by the symbol filter 124, converts the received OFDM signal Y (m, k) in parallel in the symbol direction, and outputs Had to do.

On the other hand, the equalizer 1 according to the present embodiment is
Only the interpolation process in the carrier direction (frequency direction) is performed on the synchronous modulator SM ₁ . The IFFT calculator 5, the noise removal filter 6, the up-sampling unit 7, and the FFT calculator 8 are provided for each symbol of the received OFDM signal Y (m, k) for one symbol whose carrier number k is 0 to 107. Since the interpolation processing in the frequency direction (carrier direction) is executed, the serial-parallel circuit 120 unlike the conventional equalizer 111 is unnecessary. Further, only interpolation processing in the carrier direction is required without performing interpolation processing in the symbol direction. Therefore, there are advantages that the equalization processing becomes faster and the memory can be saved.

However, since the SP signal is cyclically inserted once every 12 carriers in the carrier direction, 2
Wave delay time τ of Rayleigh fading model
Is within 1/12 of the effective symbol length Tu, characteristic deterioration due to occurrence of intersymbol interference or the like can be dealt with, but characteristic delay occurs if the delay time τ of the reflected wave is longer than Tu / 12. There is a possibility. In the Rayleigh fading transmission line, it is known that the power of the reflected wave has an exponentially decaying characteristic with respect to the delay time τ. This fact is reflected in the literature "A comparative study of pilot-
based channel estimators for wireless OFDM "(Magnu
s Sandell, Ove Edfors, Research ReportLULEA 1996: 1
9, Lulea University of Technology, September 199
6). Therefore, the electric power of the reflected wave having a long delay time τ becomes an extremely small amount, which is considered to be practically negligible. In addition, wireless LAN (Local Area N
In the IEEE 802.11a standard of etwork), a signal arrangement in which a pilot signal is cyclically inserted at a rate of one in 16 carriers in the frequency direction, and a transmission path response is interpolated and estimated for each symbol in the frequency direction for synchronous detection. The above fact is realistic because it is stipulated that Therefore, even if the delay time τ of the reflected wave exceeds Tu / 12 and is long, it can be considered that the influence of the characteristic deterioration is small.

In the noise removing filter 6, I
Impulse response h output from the FFT calculator 5₀Against
Since the noise is removed by the transmission line response H
_c(M, k _d) Estimation accuracy can be significantly improved
Becomes

The simulation results (FIGS. 9 to 14) of the equalization processing according to this embodiment will be described below.
In the simulation, an OFDM signal having a segment structure having the synchronous modulator SM ₁ is generated, modulated into a transmission signal, passed through a transmission path for simulation, and then demodulated. In addition, additive Gaussian noise (A
WGN = 10 dB) and two-wave Rayleigh fading (D / U = 5 dB) of the direct wave (D) and the reflected wave (U) were prepared.

FIG. 9 is a graph showing the transmission line response H _p of the SP signal when Gaussian noise is not mixed in the transmission line on which the two-wave Rayleigh fading is applied. The vertical axis of FIG. 9 corresponds to the amplitude of the transmission path response H _p , and the horizontal axis corresponds to the carrier number k. Further, FIG. 10 shows an impulse response h _{0 obtained} by applying IFFT to the transmission line response H _p shown in FIG.
It is a figure which shows the graph of. The vertical axis of FIG. 10 corresponds to the amplitude of the impulse response h ₀ , and the horizontal axis corresponds to the time t.
As shown in the above equation (4) of the two-wave Rayleigh fading model, it can be seen that the graph of FIG. 10 is an impulse response composed of two pulses.

Next, FIG. 11 is a diagram showing a graph of the transmission path response H _p of the SP signal when Gaussian noise is mixed in the transmission path on which the two-wave Rayleigh fading is applied.
When noise is mixed in the transmission line, it is difficult to directly remove the noise from the transmission line response H _{p in the} frequency domain. Further, the IFFT calculator 5 causes the transmission line response H shown in FIG.
_{When p} is converted into a signal in the time domain, an impulse response h ₀ as shown in the graph of FIG. 12 is obtained. As shown in FIG. 12, since the signal level of the noise is smaller than the original signal level included in the impulse response h ₀ , it can be seen that the noise can be easily removed by the binarization process. FIG.
FIG. 13 is a diagram showing a graph of an impulse response h ₁ obtained by removing noise from the impulse response h ₀ shown in FIG. 12 by the noise removal filter 6.

Then, the up-sampling unit 7
The impulse response h ₂ by executing the above upsampling process on the impulse response h ₁ shown in FIG. 13
And the FFT calculator 8 outputs the impulse response h ₂
Is subjected to FFT to output the estimated transmission path response H _c of the data signal. FIG. 14 is a diagram showing a graph of the estimated transmission line response H _c . As shown in FIG. 14, it can be seen that the transmission line response H _c in which the noise component is largely removed is obtained.

[0053]

As described above, the O according to claim 1 of the present invention
According to the FDM equalization apparatus and the OFDM equalization method according to claim 4, it is sufficient to perform only interpolation processing in the frequency direction (carrier direction) on a received OFDM signal in a symbol unit. There is no need to perform interpolation processing in the carrier direction after performing interpolation processing in the symbol direction, unlike the device and the equalization method. Therefore, the serial-parallel conversion of the received OFDM signal as in the related art is not necessary, the memory can be saved, and the equalization processing speed can be improved because the amount of calculation is reduced.

According to the second and fifth aspects, noise is not removed from the transmission path response in the frequency domain, but noise is removed from the impulse response in the time domain output from the inverse Fourier transforming means. Therefore, the noise removal performance can be improved, and the estimation accuracy of the transmission path response of the data signal can be improved.

According to claims 3 and 6, the amount of calculation is reduced, and the equalization processing speed can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing a schematic configuration of an OFDM equalizer according to an embodiment of the present invention.

FIG. 2 is a diagram showing a graph of an impulse response when noise is mixed in a signal.

FIG. 3 is a diagram showing a graph of an impulse response obtained by removing a noise component from the impulse response shown in FIG.

FIG. 4 is a block diagram of an equalizer for OFDM according to the present embodiment.
It is a figure which shows the graph of the impulse response output from a sampling part.

5 is a diagram showing a graph of a transmission line response obtained by applying FFT to the impulse response shown in FIG.

FIG. 6 is a diagram showing a graph of a transmission line response obtained by interpolating a transmission line response obtained by applying FFT to the impulse response shown in FIG. 3 by a conventional method.

FIG. 7 is a diagram showing a segment structure of a transmission symbol of an OFDM signal.

8 is a diagram showing a signal arrangement of synchronous modulation segments forming the transmission symbol shown in FIG. 7.

FIG. 9 is a diagram showing a simulation result of equalization processing according to the present embodiment.

FIG. 10 is a diagram showing a simulation result of equalization processing according to the present embodiment.

FIG. 11 is a diagram showing a simulation result of equalization processing according to the present embodiment.

FIG. 12 is a diagram showing a simulation result of equalization processing according to the present embodiment.

FIG. 13 is a diagram showing a simulation result of equalization processing according to the present embodiment.

FIG. 14 is a diagram showing a simulation result of equalization processing according to the present embodiment.

FIG. 15 is a functional block diagram showing a schematic configuration of a conventional OFDM receiver.

FIG. 16: O adopted in the European DVB-T system
It is a figure which shows the data signal of an FDM signal, and arrangement | positioning of a pilot signal.

FIG. 17 is an explanatory diagram showing a structure of a transmission symbol.

FIG. 18 is a functional block diagram showing a schematic configuration of a general equalizer.

FIG. 19 is a functional block diagram showing a conventional example of a frequency domain equalizer.

[Explanation of symbols]

1 Equalizer 2 separator 3 Pilot pattern generator 4 divider 5 IFFT calculator 6 Noise removal filter 7 Up Sampling Section 8 FFT calculator 9 divider

Claims (6)

[Claims] 1. OFDM (Orthogonal Frequency Division Multiplexing) for multiplex transmission of a plurality of subcarriers satisfying an orthogonal relationship with each other
An OFDM equalizer for equalizing a received signal transmitted by a method, comprising: a received pilot signal embedded in a known pattern from a received OFDM signal obtained by transforming the received signal from a time domain to a frequency domain by Fourier transform. Extraction means for extracting a signal and a data signal, pilot pattern generation means for generating a known complex amplitude of the received pilot signal, and a sub-transmitting the received pilot signal by dividing the received pilot signal by the complex amplitude Division means for calculating a transmission path response of a carrier, inverse Fourier transform means for calculating an impulse response in a time domain by applying an inverse Fourier transform to the transmission path response output from the division means, and the impulse response Up-sampling processing that adds a zero-valued data string to Pulling means, Fourier transform means for performing Fourier transform on the impulse response output from the up-sampling means to calculate an estimated value of the transmission line response interpolated in the frequency direction, and the data signal for the transmission line. An equalizer for OFDM, comprising: a second division unit that divides by an estimated value of a response. 2. The equalizer for OFDM according to claim 1, further comprising noise removing means for removing a noise component from the impulse response calculated by the inverse Fourier transforming means, and the up-sampling means. An equalizer for OFDM, which performs the up-sampling process on the impulse response output from the noise removing means. 3. The equalizer for OFDM according to claim 1 or 2, wherein the Fourier transform means has a fast Fourier transform function and the inverse Fourier transform means has an inverse fast Fourier transform function. Equalizer. 4. OFDM (Orthogonal Frequency Division Multiplexing) for multiplex transmission of a plurality of subcarriers satisfying an orthogonal relationship with each other
An equalization method for OFDM for equalizing a received signal transmitted by a method, comprising: (a) embedding a known pattern from a received OFDM signal obtained by transforming the received signal from a time domain to a frequency domain by Fourier transform. The received pilot signal and the data signal that are present, (b) generating a known complex amplitude of the received pilot signal, (c) the received pilot signal extracted in step (a), Dividing by the complex amplitude generated in step (b) to calculate the channel response of the subcarrier transmitting the received pilot signal; and (d) the channel calculated in step (c). Calculating an impulse response in the time domain by applying an inverse Fourier transform to the response, and (e) adding a zero-value data string to the impulse response. And (f) performing an Fourier transform on the impulse response that has been subjected to the up-sampling process in step (e) to obtain an estimated value of the transmission line response interpolated in the frequency direction. An equalization method for OFDM, comprising: a calculating step; and (g) dividing the data signal extracted in the step (a) by the estimated value. 5. The OFDM equalization method according to claim 4, further comprising: (h) removing a noise component from the impulse response calculated in the step (d). ) Is an equalization method for OFDM, which is a step of performing the up-sampling process on the impulse response from which the noise component has been removed in the step (h). 6. The equalization method for OFDM according to claim 4, wherein the inverse Fourier transform in step (d) is an inverse fast Fourier transform, and the Fourier transform in step (f) is a fast Fourier transform. The equalization method for OFDM.