JP2001237749A - Wraparound canceller - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve the problem where an acquired impulse answer has an error and the wraparound elimination performance is deteriorated by a delay time due to the impulse answer being obtained from the signal of an observation point in an environment where the wraparound varies. SOLUTION: An estimation circuit 19, which estimates the current wraparound characteristic from the changed variables of impulse answers of past wraparound waves, is added to a coefficient updating circuit 18 in an FIR filter coefficient generating circuit 4. Thus, an estimation type coefficient updating circuit 17 is configured. In such a constitution, the deterioration of variance follow-up characteristic can be suppressed, even when the wraparound waves have time variances.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

TECHNICAL FIELD The present invention relates to OFDM (Orthog
The present invention relates to a relay station (specifically, a relay apparatus) in digital broadcasting and digital transmission by an onal Frequency Division Multiplexing (orthogonal frequency division multiplexing) method, and in particular, SFN
The present invention relates to a wraparound canceller for removing a wraparound of a radio wave (hereinafter, simply referred to as a wraparound) between a transmitting and receiving antenna of a broadcast wave relay broadcasting station in a (Single Frequency Network).

[0002]

2. Description of the Related Art As a conventional wraparound canceller, there are six patent applications (Japanese Patent Application Nos. 10-162189 and 11) for a wraparound canceller according to the present inventors' invention.
147885, Japanese Patent Application No. 11-156234, Japanese Patent Application No. 11-153430, Japanese Patent Application No. 11-266567, and Japanese Patent Application No. 11-988829).

The invention of each of the above patent applications will be described in the following two or three lines, respectively. 1. "Wraparound Canceller" (Japanese Patent Application No. 10-16218)
No. 9) Invention relating to basic configuration of wraparound canceller for BST (Band Segmented Transmission) -OFDM "Wraparound Canceller" (Japanese Patent Application No. 11-14788)
No. 5) Invention for adding a normalizing means by complex division to relax the requirement for the frequency synchronization circuit. "Wraparound Canceller" (Japanese Patent Application No. Hei 11-15623)
No. 4) Invention for observing a closed-loop transfer function by phase multiplication in a differential modulation scheme such as DQPSK-OFDM. "Wraparound Canceller" (Japanese Patent Application No. 11-15343)
No. 0) Invention in which nonlinear processing is performed on the estimated impulse response of the loop interference wave to extend the frequency band of the impulse response. "Wraparound Canceller" (Japanese Patent Application No. 11-26656)
No. 7) ISDB-T (Integrated Services Digital Broardca)
5. Invention relating to a method for estimating a closed-loop transfer function when the modulation scheme differs between segments in the sting-terrestrial scheme. “OFDM demodulator” (Japanese Patent Application No. 11-98829)
At the time of OFDM demodulation, an error occurs in the observed closed-loop transfer function due to the timing error of the rectangular window used before the FFT processing, and the invention for correcting this error

[0004] Two basic structural examples of the loop canceller applied to the broadcast relay of the direct relay system described in the specification of these conventional patent applications are shown and briefly described. Here, the direct relay system is a system in which the demodulation / determination / remodulation of the received signal is not performed, and the signal of the master station is directly amplified and re-radiated. FIG. 5 shows a configuration of a relay broadcasting apparatus using a loop-back canceller based on a direct relay system (SFN) that amplifies and re-radiates a master station signal without demodulation / determination / remodulation of a received signal. An example is shown.

In FIG. 5, a relay station for SFN re-radiates at the same frequency (the relay of the relay station is an amplifier 6).
), The transmission antenna 7 to the reception antenna 1
Radio waves wrap around. In the wraparound canceller surrounded by a broken line, a subtractor 2 and a FIR (Finite-duration Inpulse)
(Response) A replica of the wraparound wave is created using the filter 3 and the FIR filter coefficient generation circuit 4, and the wraparound wave is canceled by the subtractor 2. A point at which a signal supplied to the FIR filter 3 and the FIR filter coefficient control circuit 4 is extracted (hereinafter, referred to as an observation point) is an input of the bandpass filter 5. In FIG. 5, even if a frequency conversion circuit is inserted into the input / output signal of the loop canceller surrounded by a dotted line and the loop cancel is performed by changing the frequency band (for example, a high frequency signal, an intermediate frequency signal, an equivalent complex baseband). Signal) essentially unchanged.

FIG. 6 shows another example of the configuration of a relay broadcast apparatus using a wraparound canceller of the direct relay system, similarly to FIG. Also in this drawing, each circuit portion is denoted by the same reference numeral as in FIG. The difference from the configuration of FIG. 5 is that the observation point P is the input of the bandpass filter 5 in FIG. 5, whereas the observation point P is the output of the amplifier 6 in FIG. In FIG. 6, as in FIG. 5, there is essentially no difference even if a frequency conversion circuit is inserted into the input / output signal of the loop canceller and the loop cancel is performed using any frequency band.

Here, a description will be given of the FIR filter coefficient generation circuit 4 used in the above-described wraparound canceller and composed of an impulse response calculation circuit and a coefficient update circuit. First, the impulse response calculation circuit will be described with reference to FIG. The signal at the observation point P (see FIGS. 5 and 6) is represented by s (t), and the signal obtained by performing FFT on the

[Outside 3] And Here, [3] is a K-dimensional vector at time n, where K is the total number of carriers per symbol of the OFDM signal. Further, n is an integer, and represents a time at which [outside 3] is observed discretely. The signals handled here are complex signals unless otherwise specified. The complex signal of the K-dimensional vector is input to a complex division circuit 9 and is converted into a pilot signal (vector) generated by a pilot signal generation circuit 10.

[Outside 4] Divided by the closed-loop transfer function (vector)

[Outside 5] Ask for.

(Equation 3)

However, a pilot signal (SP: Scattered Pi) in a terrestrial digital broadcasting system such as ISDB-T is used.
The lot) signal is intermittently inserted in the carrier direction and the symbol direction, and therefore, it is necessary to perform an interpolation process such as a linear interpolation on a portion having no SP signal. For details of the interpolation process, see Hamazumi et al., “Broadcast Loop Relay Canceller for Digital Terrestrial Broadcasting SFN”,
IEICE Technical Report EMCJ98-111 (1999-03),
Hayashi et al., "Study of Corresponding Equalization Method in DFDM Demodulation", TV Gakuhoho, Vol. 20, No. 53, pp55-
60, Oct. 1996. In the case of the differential modulation method in which the SP signal is not inserted, in the above-mentioned application entitled “Wraparound Canceller”,
Reference is made to Japanese Patent Application No. 11-156234, which has made it possible to estimate a closed-loop transfer function without using a reference signal such as this.

Next, from the closed-loop transfer function [Eq. 5], a cancellation residual operation circuit 11 cancels the residual (vector).

[Outside 6] Ask for.

(Equation 4) Where the scalar quantity

[Outside 7] Is defined by the following equation.

(Equation 5) Next, IFFT is performed by an inverse FFT (IFFT) circuit 12, and an impulse response (vector) is performed.

[Outside 8] Convert to

(Equation 6)

FIR coefficient generation circuit 4 (see FIGS. 5 and 6)
FIG. 8 shows a coefficient updating circuit which is a subsequent stage of FIG. As shown, the coefficient updating circuit includes a multiplier 13, an adder 14, and a delay circuit 15, and tap coefficients (vectors) of the FIR filter by the following successive updating formula.

[Outside 9] To update.

(Equation 7) Here, μ is an update coefficient (real number).

[0011]

As described above, the FIR filter coefficient generation circuit constituting the wraparound cancellation is described below.
The description has been made using equations (1) to (5). However, this method has a problem that, since a delay time exists in a series of FIR filter coefficient generation processes, when a wraparound wave has a temporal variation, a wraparound cancellation effect is reduced.

In consideration of actual hardware, the signal input to the coefficient updating circuit ([8]) has an impulse response of one sample past, which will be described later, because of the delay time required for the processing of equations (1) to (5). (vector)

[Outside 10] Will be input. Therefore

(Equation 8) The coefficient (vector) written to the FIR filter is

[Outside 11] In other words, the coefficient at time n-1 is obtained. In an environment in which the wraparound varies, the signal at the observation point is received, the wraparound elimination performance is reduced due to a time delay until the wraparound characteristic is calculated and the coefficient of the FIR filter is updated.

SUMMARY OF THE INVENTION It is an object of the present invention to provide a wraparound canceller configured to suppress a decrease in wraparound performance and improve a fluctuation follow-up characteristic even when a wraparound wave has a temporal variation.

[0014]

In order to achieve the above object, a wraparound canceller according to the present invention comprises a subtractor and a wraparound substantially connected so that an output signal is supplied to a subtraction terminal of the subtractor. And a signal processing unit for generating a signal copy. The subtracted terminal of the subtractor is supplied with a received OFDM signal including the wraparound signal, and an output terminal of the subtractor. An input terminal of the relay broadcaster is substantially connected, and an input terminal of the signal processing unit is such that any one of input and output signals of the relay broadcaster is branched and substantially supplied. The signal processing unit is configured to include a transversal filter and a filter coefficient generation circuit for controlling a tap coefficient of the filter. An impulse response calculation circuit that is supplied with one of the input / output signals of the relay broadcaster and calculates an impulse response of the signal, an output of the impulse response calculation circuit is supplied, and a tap coefficient of the filter is calculated. A coefficient updating circuit to be updated and an output of the impulse response calculation circuit are supplied, a tap coefficient of the current filter is predicted from a change amount of an impulse response of a past wraparound wave, and the coefficient update is performed based on the prediction result. And a prediction circuit for controlling a tap coefficient output of the circuit.

Further, the wraparound canceller of the present invention provides a circuit portion of a predictive coefficient updating circuit composed of the coefficient updating circuit and the predicting circuit so as to perform the calculation of [Equation 1].
It is characterized by comprising a plurality of delay circuits, multipliers and adders.

Further, the wraparound canceller of the present invention comprises a circuit portion of a predictive coefficient updating circuit comprising the coefficient updating circuit and the predicting circuit, wherein at least a predictive coefficient vector Using the difference from [Eq. 2] as input, the prediction interpolation coefficient [Equation 2] at each time n, l (integer) obtained by equally dividing the interval of the discrete coefficient update time n (integer) into L is calculated. And so on, subtractors, multipliers,
It is characterized by comprising a divider and an adder.

[0017]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below based on embodiments of the present invention with reference to the accompanying drawings. As described above, the wraparound canceller according to the present invention provides a transversal filter such as an FI that constitutes the wraparound canceller.
By adding a prediction circuit for predicting the current FIR filter coefficient from the change amount of the impulse response of the past wraparound wave to the coefficient update circuit in the R filter coefficient generation circuit, and forming a prediction type coefficient update circuit, In this case, even if there is a temporal variation, the degradation of the variation following characteristic is suppressed.

FIG. 1 shows a first embodiment of the wraparound canceller according to the present invention.
Is shown in a block diagram. FIG. 2, including FIG. 1, showing a second embodiment of the present invention, which will be described next, is the same as FIG. 5 and FIG. 6 (a configuration example of a relay broadcast apparatus using a conventional loop canceller). Circuit portions are denoted by the same reference numerals.

In the wraparound canceler of the present invention, the parts of the subtractor 2 and the FIR filter 3 (including the circuit connection) are the same as those of the conventional one (see FIGS. 5 and 6). The filter coefficient generation circuit 4 will be described. In FIG. 1, the FIR filter coefficient generation circuit 4 includes an impulse response calculation circuit 16 for calculating an impulse response at the observation point P, and a prediction type coefficient update circuit 17
It is constituted by. An important point of the present invention is that a prediction type coefficient updating circuit 17 is provided in a conventional coefficient updating circuit 18 (see FIG. 8) for predicting a current FIR filter coefficient based on a change amount of an impulse response of a past wraparound wave. This is realized by adding a circuit 19.

FIG. 2 shows a second embodiment of the wraparound canceller according to the present invention.
Is shown in a block diagram. This embodiment corresponds to the conventional wraparound canceller shown in FIG. 5 in the first embodiment, and corresponds to the observation point P of the impulse response.
Is the point on the input side of the amplifier 6, whereas the point on the output side of the amplifier 6 is the observation point P (corresponding to the conventional wraparound canceller shown in FIG. 6). It is different from the form.

FIG. 3 shows a third embodiment of the wraparound canceller according to the present invention.
Is shown in a block diagram. The present embodiment and the fourth embodiment which is an extension thereof include the first and second embodiments.
In the embodiment, the prediction type coefficient updating circuit 17 is composed of a coefficient updating circuit 18 and a prediction circuit 19 as shown in FIGS.
This is not a separate circuit configuration but an integrated predictive coefficient updating circuit.

That is, in the predictive coefficient updating circuit shown in FIG. 3, three delay circuits 20 and four multipliers 21
And two adders 22 are interconnected as shown,
The impulse response (vector) of equation (6) above
0] and past impulse response (vector)

[Outside 12] By observing both of the above, the fluctuation component of the impulse response is predicted and corrected.

The principle of operation of the predictive coefficient updating circuit will be described using mathematical expressions. The coefficient (vector) of the FIR filter corresponding to the characteristic of the true loop propagation path at the coefficient update time n is

[Outside 13] And the prediction coefficient vector of the coefficient of the FIR filter is [1]. Now, in calculating [1], the FIR corresponding to the characteristic of the true wraparound propagation path one step past.
The coefficient vector of the filter

[Outside 14] Is considered, and the difference is corrected using the difference. Assuming that the prediction correction coefficient is μ _p , the prediction coefficient vector [1] is given by the following equation.

(Equation 9) Here, the equation for updating in an ideal state where there is no time delay (coefficient calculation delay) from receiving the signal of the observation point, calculating the wraparound characteristic and updating the coefficient of the FIR filter is represented by an impulse response (vector).

[Outside 15] Where μ is the update coefficient for

(Equation 10) Can be represented by

The following equations are obtained by substituting the equations (8) and (9) into the equation (7).

[Equation 11] here,

[Outside 16] And the following equation is obtained.

(Equation 12) This equation is calculated by the predictive coefficient updating circuit shown in FIG.
(Second order case).

Equation (11) indicates that the prediction coefficient vector [〔1] of the filter coefficient at time n is the prediction coefficient vector [外 2] and the prediction coefficient vector at time n−1 and n−2 before time n, respectively.

[Outside 17] To the sum total is multiplied by the coefficient a _1, a _2, respectively, the impulse response vector [outer 10] in each of the previous time from the time n n-1 and n-2 and [External 1
2] is expressed by the sum of the sum of the products multiplied by b ₁ and b ₂ , respectively. Therefore, as shown in FIG. 3, by a plurality of delay circuits, multipliers and adders, ( It can be easily understood that the calculation of the expression 11) is performed.

The above equation (11) can be generally extended to the M-th order, and the coefficient updating equation in this case is as follows.

(Equation 13) FIG. 4 shows an example of the configuration of a predictive coefficient updating circuit when the M-order is extended (fourth embodiment of the wraparound canceller of the present invention).

The prediction coefficient vector [1] given by the equation (12) is a value obtained at every discrete time n, and if the calculation time of the impulse response vector [15] becomes longer, the interval of n becomes longer. It becomes longer, and the follow-up characteristic to the fluctuation of the wraparound decreases.

Consider this. FIG. 9 shows that when calculating the prediction coefficient vector given by the equation (12),
4 shows a timing relationship between an OFDM signal and a prediction coefficient. In FIG. 9, the arrow line shown below indicates the timing on the time axis at which the coefficient is discretely updated. FIG.
(A) is an OFD transmitted for each symbol number i.
The M signal (OFDM signal at the observation point) is shown.
In the example of FIG. 9, the impulse response vector [1
5] is expressed as two symbols. Therefore, the time interval of the coefficient update time n is three times the time interval of the symbol time i of the OFDM signal.

In this case, as shown in FIGS. 9B and 9C, an impulse response vector is calculated from one symbol of the OFDM signal, and a prediction coefficient vector is calculated from the calculation result. However, since the prediction coefficient vector reflected on the FIR filter at the time of the OFDM signal Si is mainly a delay time required for calculating the impulse response vector, the OFF in the past by the calculated delay time is used.
It has to be obtained from the DM signal. In FIG. 9, it is assumed that this situation is taken at time n-1 at the OFDM signal Si-3, and the calculated impulse response vector is
0].

As described above, F updated at discrete time n
The coefficients of the IR filter are not updated until the next time n + 1, and the previous prediction coefficient vector [1] is retained (see FIG. 9C). That is, the calculation of the impulse response vector is performed once for every two symbols of the OFDM signal, and the prediction coefficient vector of one step past is held for the symbol section that is not calculated. However, holding the prediction coefficient vector one step past reduces the follow-up characteristic to the wraparound fluctuation.

To solve this problem, the following three methods can be considered. 1) Speed up the calculation and minimize the delay time required to calculate the prediction coefficient vector. 2) Making full use of parallel processing to minimize the delay time required to calculate the prediction coefficient vector. 3) An interpolated coefficient is created from the coefficient obtained at each discrete coefficient update time n, and the coefficient is updated little by little.

The method of the above 1) is a DSP (Digital Sign)
al Processor), which can be realized in the age when devices capable of extremely high-speed calculations are available.
However, currently no such DSP is available.
The method 2) uses a plurality of DSP devices,
There is nothing that cannot be achieved by devising a circuit, but there is a disadvantage that hardware is complicated.

On the other hand, the method 3) has a discrete time n such as the prediction coefficient vector given by the equation (12).
Instead of updating the prediction coefficient every time, the interpolation coefficient is created and the prediction coefficient is updated little by little, and as described below, it can be realized only by adding a simple calculation process (This is referred to as a fifth embodiment).

In the methods 1) and 2), the time required for calculating the prediction coefficient vector cannot be reduced to less than one symbol in principle. The reason is that the FFT is used for the calculation, so that the calculation cannot be started until one symbol period of the OFDM signal ends. On the other hand, in the fifth embodiment, although there is a problem of the prediction error, the time for calculating the coefficient can be made shorter than one symbol period.

In this method (method of the above 3)),
First, the interval of the discrete coefficient update time n is equally divided into L. Here, it is assumed that each time equally divided into L is represented by n and l. Here, l is 0, 1,..., L,.
Is an integer in the range When interpolating from the prediction coefficient vector, it is considered to obtain a prediction interpolation coefficient vector at times n and l from the difference between the prediction coefficient vector [1] and the prediction coefficient vector [2]. Now, the initial value l = 0
Are the values of the prediction coefficient vector [1] and the prediction coefficient vector [2], respectively.

[Equation 14] When

(Equation 15) , And the prediction interpolation weighting coefficient is μ _pi , the prediction interpolation coefficient vector

[Outside 18] Becomes as follows.

(Equation 16)

The calculation of the equation (13) takes the prediction coefficient vector [1] and the prediction coefficient vector [2] as inputs, and is shown in FIG. 10 composed of a subtractor, a multiplier, a divider and an adder. It is realized by a circuit.

In order to clarify the difference between the prediction interpolation coefficient vector (Equation (13)) obtained as described above and the prediction coefficient vector given by Expression (12), the time of both coefficient vectors is calculated. FIG. 11 shows the relationship. FIG. 11 is the same as FIG. 9 except that FIG. 11D is shown in FIG.

The above is the prediction interpolation coefficient vector [18]
Is calculated from the difference between the prediction coefficient vector [1] and the prediction coefficient vector [2]. This is the same as the prediction coefficient vector [1] and the prediction coefficient vector [2]. Coefficient vector

[Outside 19] And prediction coefficient vector

[Outside 20] It is also possible to obtain a prediction interpolation coefficient vector [Eq.

Next, in order to evaluate the follow-up characteristics of the wraparound canceller according to the wraparound fluctuation, the characteristics are compared using graphs shown in FIGS. 12 to 14 created by simulation. First, FIG. 12 shows the result of calculating the relationship between the wraparound fluctuation and the cancellation error when the prediction algorithm according to the present invention is not used. In this case, ISDB-
Assuming the T method, mode 3, guard interval ratio 1 /
8. It is assumed that the delay time for calculating the coefficient is 15 ms, the wraparound frequency is 2 Hz, and the waveform changes sinusoidally.

In this case (FIG. 12), the value of μ of the wraparound cancel algorithm is 1.0
Was used. In the figure, the horizontal axis is time, and the vertical axis is the amplitude value of each characteristic. From FIG. 12, it can be understood that the output characteristic of the transversal (FIR) filter has a step-like shape, a phase delay occurs, and a cancellation residual occurs with respect to the wraparound fluctuation characteristic.

Next, FIG. 13 shows the result of calculating the relationship between the wraparound fluctuation and the cancellation error when the prediction algorithm for the prediction coefficient vector given by the above equation (12) is used. The calculation parameters in this case are also shown in FIG.
2, the present embodiment (fourth embodiment) uses a prediction algorithm. Therefore, a prediction weighting coefficient μ _p = 0.8 of this algorithm is added. Compared to FIG. 12, the phase delay of the output characteristic of the transversal (FIR) filter is improved by prediction, but the time interval of the coefficient control is 15 m.
Since it is large as s, it can be understood that the characteristic becomes a step-like characteristic and the cancellation residual does not become very small.

Finally, FIG. 14 shows the result of calculating the relationship between the wraparound fluctuation and the cancellation error when the prediction algorithm for the prediction interpolation coefficient vector given by the above equation (13) is used. The calculation parameters in this case are the same as those in FIG. As can be seen from FIG. 14, in the present embodiment (fifth embodiment), since the interpolation processing is performed in addition to the prediction processing, it can be understood that the cancellation residual is significantly reduced.

[0043]

According to the present invention, even in the case where there is a temporal variation in the wraparound wave, the error of the coefficient of the transversal filter, for example, the FIR filter can be reduced.
The fluctuation follow-up characteristics can be improved. As a result, it is possible to provide a wraparound canceller that operates stably even in an environment where the wraparound wave fluctuates greatly due to a typhoon or the like.

Also, as in the fifth embodiment described above,
When the interpolation processing is performed in addition to the prediction processing, it is possible to further improve the follow-up characteristics with respect to the wraparound fluctuation.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a first embodiment of a wraparound canceller according to the present invention.

FIG. 2 is a block diagram showing a second embodiment of the wraparound canceler of the present invention.

FIG. 3 is a block diagram showing a third embodiment of the wraparound canceller of the present invention.

FIG. 4 is a block diagram showing a wraparound canceller according to a fourth embodiment of the present invention.

FIG. 5 shows an example of a configuration of a relay broadcast device using a conventional loop canceller of a direct relay system.

FIG. 6 shows another configuration example of a relay broadcast device using a conventional loop canceller of a direct relay system.

FIG. 7 shows a conventional impulse response calculation circuit constituting an FIR filter coefficient generation circuit.

FIG. 8 shows a conventional coefficient updating circuit constituting an FIR filter coefficient generating circuit.

FIG. 9 shows a timing relationship between an OFDM signal and a prediction coefficient in calculating a prediction coefficient vector given in the fourth embodiment.

FIG. 10 is a block diagram showing a wraparound canceller according to a fifth embodiment of the present invention.

FIG. 11 Predicted interpolation coefficient vector (fifth embodiment)
14 shows the time relationship between both coefficient vectors of the prediction coefficient vector and the prediction coefficient vector (the fourth embodiment).

FIG. 12 shows a result of calculating a relationship between a wraparound fluctuation and a cancellation error when the prediction algorithm according to the present invention is not used.

FIG. 13 shows a result of calculating a relationship between a wraparound fluctuation and a cancellation error when a prediction algorithm of a prediction coefficient vector given in the fourth embodiment is used.

FIG. 14 illustrates a result of calculating a relationship between a wraparound fluctuation and a cancellation error when a prediction algorithm of a prediction interpolation coefficient vector given in a fifth embodiment is used.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Receiving antenna 2 Subtractor 3 FIR filter 4 FIR filter coefficient generation circuit 5 Bandpass filter 6 Amplifier 7 Transmission antenna 8 FFT circuit 9 Complex division circuit 10 Pilot signal generation circuit 11 Cancel residual calculation circuit 12 IFFT circuit 13, 21, 24 Multipliers 14, 22, 26 Adders 15, 20 Delay circuit 16 Impulse response calculation circuit 17 Predictive coefficient updating circuit 18 Coefficient updating circuit 19 Predicting circuit 23 Subtractor 25 Divider

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Kazuhiko Shibuya 1-10-11 Kinuta, Setagaya-ku, Tokyo Japan Broadcasting Corporation Japan Broadcasting Research Institute F-term (reference) 5K022 DD01 DD13 DD19 DD21 DD31 DD34 5K046 AA09 EE06 EE57 EF11 EF21 EF23 EF46 5K072 AA04 BB14 BB27 GG01 GG11

Claims (3)

[Claims] And a signal processing unit for generating a copy of the wraparound signal, the signal processing unit being substantially connected so that an output signal thereof is supplied to a subtraction terminal of the subtractor. The received OFDM signal including the wraparound signal is substantially supplied to the subtracted terminal of the subtractor, the input terminal of the relay broadcaster is substantially connected to the output terminal of the subtracter, and the signal processing unit The input terminal is configured so that either one of the input and output signals of the relay broadcaster is branched and substantially supplied, and the signal processing unit controls the transversal filter and the tap coefficient control of the filter. And a filter coefficient generating circuit for receiving the input signal of the relay broadcaster. An impulse response calculation circuit for calculating a pulse response; an output of the impulse response calculation circuit being supplied to update a tap coefficient of the filter; and an output of the impulse response calculation circuit being supplied and a past wraparound. A prediction circuit that predicts the current tap coefficient of the filter from the amount of change in the impulse response of the wave, and controls a tap coefficient output of the coefficient update circuit based on the prediction result. A wraparound canceller. 2. The wraparound canceller according to claim 1, wherein a circuit portion of the prediction-type coefficient updating circuit constituted by said coefficient updating circuit and said prediction circuit is expressed by: A wraparound canceller comprising a plurality of delay circuits, multipliers, and adders so as to perform the calculation of. 3. The wraparound canceller according to claim 1, wherein a circuit portion of a prediction-type coefficient update circuit constituted by said coefficient update circuit and said prediction circuit is at least a prediction coefficient vector according to claim 2. ] And prediction coefficient vector And the prediction interpolation coefficient at each time n, l (integer) obtained by equally dividing the interval between discrete coefficient update times n (integer) by L A wraparound canceller characterized by using a subtractor, a multiplier, a divider, and an adder so as to perform the calculation.