JP4740069B2 - Wraparound canceller - Google Patents

Description

  The present invention relates to a relay station or a relay apparatus in digital broadcasting or digital transmission using an OFDM (Orthogonal Frequency Division Multiplexing) system, and in particular, between transmitting and receiving antennas of a broadcast wave relay station in an SFN (Single Frequency Network). The present invention relates to a sneak canceller for removing a sneak current of radio waves in the network.

  The sneak canceller cancels the sneak wave component generated by the coupling between the transmitting and receiving antennas included in the received signal in the SFN broadcast wave relay station that retransmits at the same frequency as the frequency of the received signal, and retransmits only the upper station signal. It is a device for.

  FIG. 3 is a diagram showing a configuration of a conventional wraparound canceller. A conventional wraparound canceller 100 includes a frequency converter 101 that converts a received signal into an IF signal, an A / D converter 102 that converts an IF signal into a digital IF signal, and a digital IF signal orthogonal to an equivalent baseband signal. Quadrature demodulation unit 103 that demodulates, subtracting unit 104 that subtracts the replica signal of the sneak wave from the equivalent baseband signal, adaptive filter unit 105 that generates a replica signal of the sneak wave, and an equivalent baseband signal obtained by subtracting the replica signal of the sneak wave A quadrature modulation unit 106 that quadrature modulates the digital IF signal into a digital IF signal, a D / A conversion unit 107 that D / A converts the digital IF signal into an analog IF signal, and a frequency conversion unit 108 that converts the frequency of the analog IF signal into a transmission signal in the RF band. And a filter unit used in the adaptive filter unit 105 by estimating the sneak path. And a wraparound channel estimation unit 110 generates a.

  Under such a configuration, a sneak wave is generated between a transmission antenna that radiates a radio wave of a transmission signal and a reception antenna that receives a radio wave from an upper station as a reception signal. This sneak wave is generated when a part of the radio wave radiated from the transmitting antenna (the radio wave of the transmission signal) passes through the sneak path and is received by the receiving antenna that receives the upper-level local wave. It will occur.

  In order to cancel this sneak wave component from the received signal, a circuit having the same transmission characteristics as the sneak path (the propagation path characteristic of the sneak path) is generated inside the sneak canceller 100, and a sneak wave replica signal is generated. There is a need to. The wraparound canceller 100 can cancel the wraparound wave by subtracting the wraparound wave replica signal generated inside the wraparound canceller 100 from the received signal, extract only the signal from the upper station, and transmit it as a transmission signal. .

  Here, the adaptive filter unit 105 uses the filter coefficient generated based on the delay profile of the sneak path by the sneak path estimation unit 110 to realize a circuit of the sneak path propagation path characteristic, and Generate a replica signal.

  A sneak path estimation unit 110 that generates a filter coefficient extracts effective symbols based on symbol timing to generate an OFDM signal in the time domain of the effective symbol period, and an effective symbol extraction unit 111 that generates an OFDM signal in the time domain of the effective symbol period. FFT (Fast Fourier Transform) means 112 for performing fast Fourier transform on the carrier symbol, transmission path response calculating means 113 for calculating the transmission path response from the carrier symbol, and calculating the transmission path response of the cancellation residual from the transmission path response. IFFT (Inverse Fast Fourier Transform: Inverse Fast Fourier Transform) which performs inverse fast Fourier transform from the channel response of the cancellation residual to the impulse response of the wraparound cancellation residual. ) Means 115, multiplication means 116 for multiplying the impulse response of the wraparound cancellation residual by a predetermined constant, addition means 117 for adding the multiplied impulse response of the wraparound cancellation residual to generate a filter coefficient, and filter coefficients Delay means 118 is provided for outputting after a predetermined time delay.

  Examples of the wraparound canceller similar to the wraparound canceller 100 include those described in Patent Documents 1 to 6, for example.

JP-A-11-355160 JP 2000-341238 A JP 2000-349734 A JP 2001-94528 A JP 2000-295195 A Japanese Patent Laid-Open No. 2001-237749

  In the conventional wraparound canceller 100 shown in FIG. 3, when the reception state is poor, for example, when the C / N (Carrier to Noise Ratio) of the signal transmitted from the upper station is small, the signal at the reception unit is low because the reception power is low. When the C / N of the sneak wave is small, when the D / U (Desired to Undesired Ratio) of the sneak wave signal is small, or when it is under the same channel interference, the sneak path estimation unit 110 propagates the sneak path. When estimating the road characteristics, the noise component of the signal may be increased due to numerical instability.

  In this case, the adaptive filter unit 105 generates a replica signal of a sneak wave using the file coefficient generated from the signal having a large noise component by the sneaking propagation path estimation unit 110. That is, the adaptive filter unit 105 generates a replica signal including an erroneous sneak wave component that does not exist in the actual sneak wave propagation path. Therefore, the wraparound canceller 100 has a problem in that the signal oscillates as a result because the erroneous wraparound wave replica signal generated by the adaptive filter unit is not canceled out with the received signal.

  In addition, the cause of such a problem is when the transmission line response calculation unit 113 of the sneak path estimation unit 110 calculates the transmission line response and when the cancellation residual calculation unit 114 calculates the cancellation residual. The reason is that the noise component of the signal is increased due to numerical instability. That is, when division is performed with zero or a very small numerical value, the result of the division becomes indefinite or a numerically very large error occurs.

  Generally, when there is a possibility of dividing by zero, this is avoided by adding a small amount in advance to the denominator. However, this avoidance method is not a fundamental solution because the division result includes a very large error numerically.

  Therefore, the present invention has been made to solve such a problem, and an object of the present invention is to provide a wraparound canceller that can operate stably even when the reception state is poor or under the same channel interference. It is to provide.

  In order to solve the above-described problem, the wraparound canceller according to the present invention synthesizes the input OFDM signal and the wraparound replica signal generated by the adaptive filter in reverse phase, thereby canceling the wraparound canceller. FFT means for fast Fourier transforming a signal into frequency domain carrier symbols, transmission path response calculation means for calculating a transmission line response from the carrier symbols output by the FFT means, and transmission path output by the transmission path response calculation means From the response, a cancellation residual calculation means for calculating a wraparound cancellation residual in the frequency domain, and a component equal to or higher than a preset threshold value among the frequency domain wraparound cancellation residual output by the cancellation residual calculation means. Threshold processing means to be removed and the threshold processing IFFT means for inverse fast Fourier transforming the wraparound cancellation residual output from the stage into a time domain wraparound cancellation residual, and multiplication means for multiplying the time domain wraparound cancellation residual output from the IFFT means by a preset constant. And an addition means for adding the update of the time domain sneak cancellation residual output by the multiplication means to the filter coefficient before the unit update time used by the adaptive filter, and outputting to the adaptive filter as a new filter coefficient. It is characterized by providing.

  Further, the wraparound canceller according to the present invention synthesizes the input OFDM signal and the wraparound replica signal generated by the adaptive filter in reverse phase, thereby canceling the wraparound canceller. Of the FFT means for performing a fast Fourier transform on carrier symbols, the transmission line response calculation means for calculating a transmission line response from the carrier symbols output by the FFT means, and the transmission line response output by the transmission line response calculation means, A threshold processing unit that replaces a component that is equal to or less than a set threshold with a preset value, and a cancellation residual that calculates a wraparound cancellation residual in the frequency domain from the transmission path response output by the threshold processing unit. A frequency region output by the difference calculating means and the cancellation residual calculating means; IFFT means for inverse fast Fourier transforming the wraparound cancellation residual of time to a wraparound cancellation residual in the time domain, multiplication means for multiplying the time domain wraparound cancellation residual output by the IFFT means by a preset constant, Addition means for adding an update of the time domain sneak cancellation residual output by the multiplication means to the filter coefficient before the unit update time used by the adaptive filter and outputting the result to the adaptive filter as a new filter coefficient It is characterized by.

  As described above, according to the present invention, the cancellation residual equal to or higher than the threshold value is removed from the cancellation residual of the sneak wave calculated from the transmission path response of the signal after the sneak cancellation. Further, according to the present invention, among the transmission line responses of the signal after the sneak cancel, the transmission line response below the threshold value is replaced with a preset value. Thereby, in the calculation of the cancellation residual of the sneak wave, the division result does not become unstable and a large numerical error does not occur. Therefore, it is possible to realize a wraparound canceller that can operate stably even when the reception state is poor or under the same channel interference.

  The best mode for carrying out the present invention will be described below in detail with reference to the drawings. This best mode will be described with reference to the wraparound canceller 1 of the first embodiment shown in FIG. 1 and the wraparound canceller 2 of the second embodiment shown in FIG. In both the first and second embodiments, signal noise is caused by numerical instability so that the receiver can operate stably even when the reception state is poor or under the same channel interference. The ingredients are kept from becoming large. Specifically, in the wraparound canceller 1 according to the first embodiment, regarding the calculation of the cancellation residual by the cancellation residual calculation unit 14, the division result does not become indefinite when the division is performed with zero or a very small numerical value. A threshold value processing means 19 provided in the subsequent stage of the cancellation residual calculation means 14 forcibly sets a cancellation residual equal to or greater than a predetermined value to zero. In addition, in the wraparound canceller 2 according to the second embodiment, the calculation of the cancellation residual by the cancellation residual calculation unit 14 is performed so that the division result does not become indefinite when the division is performed with zero or a very small numerical value. The threshold value processing means 29 provided in the subsequent stage of the calculating means 13 forcibly sets the calculation result of the transmission line response below a predetermined value to a predetermined sufficiently large value.

  First, Example 1 will be described. FIG. 1 is a block diagram showing a first configuration of a wraparound canceller according to the present invention. The wraparound canceller 1 includes a frequency conversion unit 101, an A / D conversion unit 102, an orthogonal demodulation unit 103, a subtraction unit 104, an adaptive filter unit 105, an orthogonal modulation unit 106, a D / A conversion unit 107, a frequency conversion unit 108, and A sneak path estimation unit 10 is provided. The sneak path estimation unit 10 includes an effective symbol extraction unit 11, an FFT unit 12, a transmission path response calculation unit 13, a cancellation residual calculation unit 14, a threshold processing unit 19, an IFFT unit 15, a multiplication unit 16, and an addition. Means 17 and delay means 18 are provided. However, in FIG. 1, the same reference numerals as those in FIG. 3 are assigned to portions common to the conventional wraparound canceller 100 shown in FIG.

  The frequency conversion unit 101 converts the RF band received signal received via a receiving antenna (not shown) into an IF band signal, and outputs the IF signal to the A / D conversion unit 102. The A / D converter 102 receives the IF signal frequency-converted by the frequency converter 101, A / D-converts the IF signal into a digital IF signal using a sampling clock supplied from a synchronous reproduction unit (not shown), The digital IF signal is output to the orthogonal demodulation unit 103. The quadrature demodulator 103 receives the digital IF signal A / D converted by the A / D converter 102 and converts the digital IF signal into an I component (In-phase component) and a Q component (Quadrature component). ) To the equivalent baseband signal and output the equivalent baseband signal to the subtractor 104.

  Note that the A / D conversion unit 102 and the orthogonal demodulation unit 103 may be configured in reverse order. That is, the orthogonal demodulator 103 is provided after the frequency converter 101, and the A / D converter 102 is provided after the orthogonal demodulator 103. In this case, the quadrature demodulating unit 103 receives the IF signal frequency-converted by the frequency converting unit 101, performs analog quadrature demodulation on the IF signal to an equivalent baseband signal of I component and Q component, and performs an equivalent baseband signal of I component. And Q component equivalent baseband signals are output to the A / D converter 102, respectively. The A / D conversion unit 102 includes two A / D conversion units, which respectively input I-component and Q-component equivalent baseband signals and A / D-convert them into digital signals.

  The subtracting unit 104 inputs the equivalent baseband signal quadrature demodulated by the quadrature demodulating unit 103 and the wraparound wave replica signal from the adaptive filter unit 105, and subtracts the wrapping wave replica signal from the equivalent baseband signal and outputs it. To do. Here, the equivalent baseband signal (equivalent baseband signal obtained by subtracting the replica signal of the sneak wave) from the subtracting unit 104 is divided into three, and the quadrature modulation unit 106, the adaptive filter unit 105, and the sneak path estimation unit 10 Each is output to the effective symbol extraction means 11.

  The first equivalent baseband signal among the three divided equivalent baseband signals is output to quadrature modulation section 106. The quadrature modulation unit 106 receives the equivalent baseband signal from the subtraction unit 104, performs quadrature modulation on the equivalent baseband signal to a digital IF signal, and outputs the digital IF signal to the D / A conversion unit 107. The D / A converter 107 receives the digital IF signal orthogonally modulated by the orthogonal modulator 106, D / A converts the digital IF signal into an analog IF signal, and outputs the IF signal to the frequency converter 108. . Frequency converter 108 receives the IF signal D / A converted by D / A converter 107, converts the frequency of the IF signal to an RF band transmission signal, and transmits the transmission signal to the outside via a transmission antenna (not shown). Output.

  Further, the second equivalent baseband signal among the equivalent baseband signals distributed in the subsequent stage of the subtracting unit 104 is output to the effective symbol extraction means 11. The effective symbol extraction unit 11 receives the equivalent baseband signal from the subtracting unit 104, extracts a signal corresponding to the effective symbol period in one OFDM transmission symbol period based on the symbol timing, and performs OFDM in the time domain of the effective symbol period. The signal is output to the FFT means 12. The FFT means 12 inputs the time-domain OFDM signal of the effective symbol period from the effective symbol extraction means 11, performs fast Fourier transform on the time-domain OFDM signal to a carrier symbol which is a frequency-domain OFDM signal, and transmits the carrier symbol Output to the road response calculation means 13.

  The transmission line response calculation means 13 receives the carrier symbol from the FFT means 12 and calculates the transmission line response after the wraparound cancellation (the transmission line response from the upper station to the wraparound canceller 1 after the wraparound cancellation) from the carrier symbol. And output to the cancellation residual calculation means 14. Details of the transmission path response calculation process will be described later. The cancellation residual calculation means 14 receives the transmission path response after the sneak cancellation from the transmission path response calculation means 13, and the propagation path characteristics of the actual sneak propagation path (from the transmission antenna to the reception antenna). The cancellation residual (frequency domain signal), which is the difference between the sneaking propagation path characteristic) and the propagation path characteristic realized by the adaptive filter 105, is calculated and output to the threshold processing means 19. Details of the cancellation residual calculation process will be described later.

  The threshold processing means 19 receives the cancellation residual from the cancellation residual calculation means 14, compares the amplitude of the cancellation residual with a preset threshold, and cancels the residual that is equal to or greater than the threshold. Are output to the IFFT means 15. Details of the threshold processing will be described later. The IFFT means 15 receives the cancellation residual subjected to the threshold processing by the threshold processing means 19, performs inverse fast Fourier transform on the frequency domain cancellation residual to the impulse response of the time domain cancellation residual, and performs multiplication. Output to means 16.

  The multiplication means 16 receives the cancellation residual impulse response from the IFFT means 15, multiplies the cancellation residual impulse response by a preset value μ, and outputs the result to the addition means 17. Here, the preset value μ is a value in the range of 0 <μ <1, and the fluctuation followability of the propagation path characteristic of the sneak path realized by the adaptive filter unit 105 and its estimation accuracy It is an adaptive parameter for adjusting the trade-off between the two. When μ is set to a large value, the fluctuation followability of the propagation path characteristics in the fluctuation environment is improved. On the other hand, when μ is set to a small value, the estimation accuracy of propagation path characteristics in an environment with little fluctuation is improved.

  The adding means 17 receives the impulse response of the cancellation residual in the wraparound propagation path from the multiplication means 16 and adds the impulse response of the cancellation residual in the wraparound propagation path to the filter coefficient before the unit coefficient update time supplied from the delay means 18. Are added and output as new filter coefficients. Here, the filter coefficient from the adding unit 17 is divided into two, one is output to the adaptive filter unit 105, and the other is output to the delay unit 18. The delay unit 18 receives the filter coefficient from the addition unit 17, holds the filter coefficient, delays it until the next coefficient update (during the unit coefficient update time), and outputs it to the addition unit 17 after the delay time elapses. .

  Also, the third equivalent baseband signal among the equivalent baseband signals distributed in three at the subsequent stage of the subtracting unit 104 is output to the adaptive filter unit 105. In addition, one of the two filter coefficients distributed in the subsequent stage of the adding means 17 is output to the adaptive filter unit 105. The adaptive filter unit 105 receives an equivalent baseband signal and a filter coefficient, filters the equivalent baseband signal using the filter coefficient, realizes a circuit of propagation path characteristics of the wraparound propagation path, and creates a wraparound wave replica A signal is generated and output to the subtracting unit 104.

  Next, when the wraparound canceller 1 shown in FIG. 1 is applied to terrestrial digital broadcasting of an ISDB-T (Integrated Services Digital Broadcasting-Terrestrial) transmission system, the transmission path response calculation means 13 of the wrapping propagation path estimation unit 10, cancel The operations of the residual calculation unit 14 and the threshold processing unit 19 will be described in detail.

[Calculation of transmission line response]
First, details of the operation of the transmission line response calculation means 13 will be described. As described above, the transmission line response calculation means 13 receives a carrier symbol and calculates the transmission line response after the wraparound cancellation from the carrier symbol.

FIG. 7 shows an SP when a specific subcarrier of a specific symbol is assigned to a pilot signal in the ISDB-T system and DVB-T (Digital Video Broadcasting Terrestrial) system, which are broadcasting systems for digital terrestrial television broadcasting. It is a figure which shows arrangement | positioning of (Scattered Pilot: scattered pilot). In FIG. 7, SP is indicated by a black circle, and other carrier symbols such as data symbols are indicated by white circles. Hereinafter, regarding the SP arrangement, the interval in the subcarrier direction in consecutive symbols is N _f , and the interval in the symbol direction in the same subcarrier is N _t . Since the amplitude and phase of the SP are determined in advance, the sneak canceller 1 on the receiving side can generate the same signal.

を満足する。但し、ｍｏｄは剰余を示す。以下、式（1）を満足するｉ，ｋをそれぞれｉ_ｐ，ｋ_ｐとする。 FIG. 4 is a block diagram showing a first configuration of the transmission path response calculation means 13 shown in FIG. The transmission line response calculation means 13-1 includes an SP extraction stage 31, a reference SP generation stage 32, a division stage 33, and an interpolation stage 34. For a subcarrier assigned to an SP, if the symbol number is i and the subcarrier number is k,

Satisfied. However, mod indicates a remainder. Hereinafter, i and k satisfying the expression (1) are assumed to be i _p and k _p , respectively.

  The SP extraction stage 31 receives a carrier symbol, extracts an SP from the carrier symbol, and outputs it to the division stage 33. The reference SP generation stage 32 generates a reference SP having the same amplitude and phase as the SP generated by the upper station and outputs the reference SP to the division stage 33. The division stage 33 divides the SP extracted by the SP extraction stage 31 by the reference SP generated by the reference SP generation stage 32 and outputs the result to the interpolation stage 34. In addition, the interpolation stage 34 interpolates the transmission line response generated by the division stage 33.

Here, assuming that the SP extracted in the SP extraction stage 31 is Y _{ip, kp} and the reference SP generated in the reference SP generation stage 32 is X _{ip, kp} , the transmission path in the symbol number i _p and the subcarrier number k _p The response F _{ip, kp} is expressed by the following equation.

  The transmission line response calculation means 13-1 uses the SP adopted in the ISDB-T method as a reference signal for calculating the transmission line response, but is not limited to this. A symbol that has a known amplitude and phase and can be generated by the wraparound canceller 1 on the receiving side may be used as the reference signal.

  FIG. 5 is a block diagram showing a second configuration of the transmission path response calculation means 13 shown in FIG. The transmission line response calculation means 13-2 includes a channel equalization stage 40, a determination stage 36, and a division stage 37. The channel equalization stage 40 includes an SP extraction stage 31, a reference SP generation stage 32, a division stage 33, an interpolation stage 34, and a division stage 35. 4 is compared with the transmission line response calculation means 13-2 shown in FIG. 5, the SP extraction stage 31, the reference SP generation stage 32, the division stage 33, and the interpolation stage 34 are compared. The transmission line response calculation means 13-2 is different in that it includes a division stage 35, a determination stage 36, and a division stage 37 in addition to the transmission line response calculation means 13-1. To do. In FIG. 5, the same reference numerals as those in FIG. 4 are given to the portions common to FIG. 4, and detailed description thereof is omitted. Here, the symbol number i is omitted.

  The channel equalization stage 40 receives a carrier symbol from the FFT means 12 and the division stage 35 outputs a carrier symbol other than SP (hereinafter referred to as a data symbol) of the carrier symbol by the interpolation stage 34. Channel equalization is performed by dividing by the transmission path response F. The channel equalization stage 40 outputs the data symbol after channel equalization to the determination stage 36.

ここで、Ｄ_ｋはサブキャリヤ番号ｋのサブキャリヤの周波数における送信シンボル、Ｎ_ｋは雑音を示す。式（３）によれば、Ｚ_ｋは、第２項により信号点Ｄ_ｋを中心に分散することがわかる。 Data symbol Z _k after channel equalization can be expressed as:

Here, D _k is a transmission symbol at the frequency of the subcarrier of subcarrier number k, and N _k is noise. According to Equation (3), it can be seen that Z _k is distributed around the signal point D _k by the second term.

  Next, the decision stage 36 inputs the data symbol after channel equalization from the channel equalization stage 40, performs threshold judgment, and is a known value having the smallest Euclidean distance in the signal space from the carrier symbol after equalization. The transmission symbol is output to the division stage 37 as an estimated value of the transmission symbol.

ｄｅｃ（ｙ）は、しきい値判定の関数であり、信号空間においてデータシンボルＹに最も近い送信信号を返す。 Here, the estimated value of the transmission symbol can be expressed by Equation (4).

Dec (y) is a threshold determination function and returns the transmission signal closest to the data symbol Y in the signal space.

つまり、除算段３７は、キャリヤシンボルを、式（４）に示す送信シンボルの推定値である基準信号で除算し、その結果の伝送路応答を出力する。 Assuming that there is no determination error in equation (4), the estimated value of the transmission symbol shown in equation (4) is equal to the transmission symbol D _{k at} the frequency of the subcarrier of subcarrier number k. Therefore, the estimated value of the transmission symbol shown in Equation (4) can be used as a reference signal in the same manner as SP, and the transmission line response can be obtained as shown in Equation (5).

That is, the division stage 37 divides the carrier symbol by the reference signal that is the estimated value of the transmission symbol shown in Equation (4), and outputs the resultant transmission path response.

[Calculation of wraparound cancellation residual]
Next, details of the operation of the cancellation residual calculation means 14 will be described. As described above, the cancellation residual calculation means 14 inputs the transmission line response, and from the transmission line response, the propagation path characteristic characteristic of the actual wraparound propagation path and the propagation path characteristic realized by the adaptive filter 105 are obtained. The cancellation residual which is a difference between them is calculated.

FIG. 6 is a diagram showing a model of the wraparound canceller 1 shown in FIG. Here, X _k , Y _k , and C _k are the transmission signal at the frequency of the subcarrier of subcarrier number k, the retransmission signal after the wraparound cancellation (the transmission signal output by the wraparound canceller 1), and the propagation path of the wraparound propagation path Each characteristic is shown.

ここで、回り込みキャンセル後の再送信信号の伝送路応答Ｆ_ｋは、

と表すことができる。 The model of FIG. 6 can be expressed by Equation (6).

Here, the transmission path response F _k of the retransmission signal after the wraparound cancellation is

It can be expressed as.

ここで、Ｆ_ｋは伝送路応答である。 Therefore, the cancellation residual E _k that is the difference between the propagation path characteristic of the actual sneak propagation path and the propagation path characteristic realized by the adaptive filter 105 can be obtained using Expression (8).

Here, F _k is a transmission line response.

As described above, the IFFT unit 15 inputs the cancellation residual E _k (cancellation residual subjected to threshold processing by the threshold processing unit 19), and the residual is subjected to inverse fast Fourier transform processing in the time domain. In this case, the impulse response of the wraparound cancellation residual can be obtained as shown in Equation (9).

[Threshold processing]
Next, details of the operation of the threshold processing means 19 will be described. As described above, the threshold processing means 19 inputs the cancellation residual, compares the amplitude of the cancellation residual with a preset threshold, and determines the amplitude component of the cancellation residual equal to or greater than the threshold. Is output.

As shown in equation (8), when calculating the coupling loop cancellation residuals, it is necessary to divide the processing performed by the channel response F _k retransmission signal after echo cancellation. In this case, when the amplitude of the transmission line response F _k that is the denominator is zero or a very small value, the division result becomes indefinite or the division result includes a numerically very large error. . Using this indefinite division result or a division result including a large error, if the impulse response of the wraparound cancellation residual is obtained by the equation (9), the noise level becomes large.

FIG. 8 is a diagram illustrating an example of signal point arrangement of transmission signals on a complex plane. However, here, scaling is performed with the following numerical values so that the signal point of the data symbol becomes an integer value.

このため、受信した信号のキャリヤシンボルが原点付近に分布する可能性は小さい。しかし、雑音成分が大きい場合、上位局から中継局である回り込みキャンセラ１までの間の伝送路にマルチパスがある場合、回り込み波成分が十分にキャンセルされていない場合等においては、受信した信号のキャリヤシンボルが振幅及び位相歪みを受けているため、その信号点が原点付近になる可能性がある。 When the transmission path response is calculated by the configuration of the transmission path response calculation means 13-1 shown in FIG. 4, only the carrier symbol of the transmission signal having the following amplitude in FIG. 8 is used.

For this reason, the possibility that the carrier symbols of the received signal are distributed near the origin is small. However, if the noise component is large, if there is a multipath on the transmission path from the upper station to the sneak canceller 1 as the relay station, or if the sneak wave component is not sufficiently canceled, the received signal Since the carrier symbol is subjected to amplitude and phase distortion, the signal point may be near the origin.

  Further, when the transmission line response is calculated by the configuration of the transmission line response calculation means 13-2 shown in FIG. 5, the signal point of the carrier symbol of the transmission signal used for calculating the transmission line response is shown in FIG. Can be all points. For example, if the amplitude of the signal point of the carrier symbol of the transmission signal is small as in (± 1, ± 1), the possibility that the signal point of the carrier symbol of the received signal is near the origin is increased.

ここで、この式（１０）の左辺は、しきい値処理後のキャンセル残差Ｅ_ｋを示す。 Therefore, the threshold processing means 19 applies a preset threshold value to the wraparound cancellation residual obtained by the cancellation residual calculation means 14 by the expression (8) as shown in the following expression (10). A process of removing components equal to or more than th is performed. Thereby, the impulse response of the wraparound cancellation residual can be accurately obtained.

Here, the left side of the equation (10) indicates the cancellation residual E _k after the threshold processing.

Here, the threshold is preferably th = 2. This is a value obtained by an experiment by the inventors of the present application. When | E _k | in Equation (10) is 1 or more, this means that the cancellation residual is greater than or equal to the desired wave component, and the subcarrier alone is in an oscillation state. From this state, it is necessary to set the threshold value to th = 1 or less. However, since the entire band is not necessarily in the oscillation state, the threshold value is set to 2, which is a value slightly larger than th = 1. In general, when the threshold value is set to be small, it becomes difficult to obtain the wraparound canceling effect that is the original purpose of the wraparound canceller 1. On the other hand, when the threshold value is set to be large, it is difficult to stably operate the wraparound canceller 1 when the reception state is poor or under the same channel interference. By setting the threshold value to th = 2, it is possible to obtain the original effect of the wraparound canceller 1 described above and to operate the wraparound canceller 1 stably.

  As described above, according to the sneak canceller 1 of the first embodiment shown in FIG. 1, the threshold processing unit 19 includes the sneak wave cancellation residual calculated from the propagation path response of the signal after the sneak cancellation. The cancellation residual exceeding the preset threshold value is removed. As a result, even if the cancellation residual calculation means 14 divides by zero or a very small numerical value of the propagation path response, even if the cancellation residual which is the calculation result becomes indefinite or large, the threshold value The processing means 19 can remove the indefinite or large value cancellation residual. As a result, the impulse response of the sneak cancellation residual can be accurately obtained by the IFFT means 15, and the filter coefficient can be accurately obtained by the sneaking propagation path estimation unit 10, so that the reception condition is poor or the same. Even in the case of channel interference, it is possible to realize a wraparound canceller that can operate stably.

  FIG. 9 is a diagram showing the estimation result of the sneak propagation path by the conventional sneak canceller 100, and FIG. 10 is a diagram showing the estimation result of the sneak propagation path by the sneak canceller 1 of the first embodiment shown in FIG. . However, here, for simplicity of explanation, the adaptive filter unit 105 that generates a wraparound replica signal is not operated, and wraparound channel characteristics obtained by one coefficient update are shown. FIG. 9 and FIG. 10A are diagrams representing on the complex plane the propagation path characteristics of the retransmitted signal after the wraparound cancellation and the wraparound cancellation residual calculated using Expression (8). Moreover, (b) is a figure which shows the impulse response of the wraparound cancellation residual calculated using Formula (9).

  Here, since the adaptive filter 105 that generates the wraparound replica signal does not operate, (b) corresponds to the delay profile of the wraparound propagation path.

  In the conventional wraparound canceller 100, the noise of the delay profile of the estimated wraparound propagation path is determined by a component whose cancellation residual values I and Q are (−7, 2.5) (see FIG. 9A). It can be seen that the level is increased (see FIG. 9B). On the other hand, in the wraparound canceller 1 according to the first embodiment, a component (see FIG. 10A) in which the values of the cancellation residuals I and Q are (−7, 2.5) (see FIG. 10A) is processed by the threshold processing unit 19. It can be seen that the delay profile of the sneak path can be accurately obtained by removing at (see FIG. 10B).

  Next, Example 2 will be described. FIG. 2 is a block diagram showing a second configuration of the wraparound canceller according to the present invention. The wraparound canceller 2 includes a frequency conversion unit 101, an A / D conversion unit 102, an orthogonal demodulation unit 103, a subtraction unit 104, an adaptive filter unit 105, an orthogonal modulation unit 106, a D / A conversion unit 107, a frequency conversion unit 108, and A sneak path estimation unit 20 is provided. Further, the sneak path estimation unit 20 includes an effective symbol extraction unit 11, an FFT unit 12, a transmission path response calculation unit 13, a threshold processing unit 29, a cancellation residual calculation unit 14, an IFFT unit 15, a multiplication unit 16, and an addition. Means 17 and delay means 18 are provided.

  When the wraparound canceller 1 shown in FIG. 1 and the wraparound canceller 2 shown in FIG. 2 are compared, the frequency conversion unit 101, the A / D conversion unit 102, the orthogonal demodulation unit 103, the subtraction unit 104, the adaptive filter unit 105, and the orthogonal modulation unit 106, the D / A conversion unit 107 and the frequency conversion unit 108 are the same, but the sneak canceller 2 includes a sneak path estimation unit 20 different from the sneak path estimation unit 10 included in the sneak canceller 1. It differs in that it has. The threshold processing means 19 in the sneak path estimation unit 10 of the sneak canceller 1 performs threshold processing on the cancellation residual, whereas the threshold processing means in the sneak propagation path estimation unit 20 of the sneak canceller 2. 29 differs from the transmission line response calculation means 13 in that it performs threshold processing on the transmission line response. However, in FIG. 2, the same reference numerals as those in FIG. 3 or FIG. 1 are attached to the portions common to the conventional wraparound canceller 100 shown in FIG. 3 or the wraparound canceller 1 shown in FIG. A description of portions common to those in FIG. 1 is omitted.

  The threshold processing means 29 of the sneak path estimation unit 20 inputs the transmission path response calculated by the transmission path response calculation means 13, and compares the amplitude of the transmission path response with a preset threshold value. Then, the transmission line response larger than the threshold value is output to the cancellation residual calculation means 14 as it is, the transmission line response component equal to or less than the threshold value is set to a preset transmission line response value, and the cancellation residual value is set. It outputs to the difference calculation means 14.

  The above-described equation (10) performs threshold processing for the wraparound cancellation residual obtained by equation (8). However, threshold processing is performed on the transmission line response before calculating the cancellation residual. Even if it performs, the same result can be obtained.

ここで、式（１１）の左辺は、しきい値処理後の伝送路応答Ｆ_ｋを示す。 Therefore, as shown in the following equation (11), the threshold processing unit 29 converts a channel response component equal to or less than a preset threshold th to a positive value of a sufficiently large propagation path response. Replace with a constant value. As a result, the impulse response of the wraparound cancellation residual can be accurately obtained without performing threshold processing for the wraparound cancellation residual.

Here, the left side of the equation (11) indicates the transmission line response F _k after threshold processing.

  As described above, according to the sneak canceller 2 of the second embodiment shown in FIG. 2, the threshold value processing unit 29 has a value equal to or lower than a preset threshold value in the transmission path response of the signal after the sneak cancel. The transmission line response is replaced with a sufficiently large transmission line response value set in advance. As a result, the calculation result of the wraparound cancellation residual can be prevented from becoming indefinite or numerically very large. That is, the impulse response of the wraparound cancellation residual can be accurately obtained by the IFFT means 15 without performing threshold processing for the wraparound cancellation residual. Further, since the filter coefficient can be accurately obtained by the wraparound channel estimation unit 20, a wraparound canceler that can operate stably even when the reception state is poor or under the same channel interference is realized. be able to.

It is a block diagram which shows the 1st structure of the wraparound canceller by this invention. It is a block diagram which shows the 2nd structure of the wraparound canceller by this invention. It is a block diagram which shows the structure of the conventional wraparound canceller. It is a block diagram which shows the 1st structure of a transmission-line response calculation part. It is a block diagram which shows the 2nd structure of a transmission-line response calculation part. It is a block diagram which shows the model of a wraparound canceller. It is a figure which shows arrangement | positioning of SP. It is a figure which shows arrangement | positioning of a signal point. It is a figure which shows the estimation result of the roundabout propagation path in the conventional roundabout canceller. It is a figure which shows the estimation result of the roundabout propagation path in the roundabout canceller (Example 1) by this invention.

Explanation of symbols

1,2,100 wraparound canceller 10,20,110 wraparound propagation path estimator 11,111 effective symbol extraction means 12,112 FFT means 13,113 transmission path response calculation means 14,114 cancellation residual calculation means 15,115 IFFT means 16, 116 Multiplication means 17, 117 Addition means 18, 118 Delay means 19, 29 Threshold processing means 30, 113 Transmission path response calculation means 31 SP extraction stage 32 Reference SP generation stage 33, 35, 37 Division stage 34 Interpolation stage 36 decision stage 40 channel equalization stage 101 frequency conversion unit 102 A / D conversion unit 103 quadrature demodulation unit 104 subtraction unit 105 adaptive filter unit 106 orthogonal modulation unit 107 D / A conversion unit 108 frequency conversion unit

Claims (2)

In the wraparound canceller that cancels the wraparound by combining the input OFDM signal and the wraparound replica signal generated by the adaptive filter in reverse phase,
FFT (Fast Fourier Transform) means for fast Fourier transforming the signal after the wraparound cancellation to a frequency domain carrier symbol;
Transmission path response calculating means for calculating a transmission path response from the carrier symbol output by the FFT means;
Cancellation residual calculation means for calculating a frequency domain wraparound cancellation residual from the transmission line response output by the transmission line response calculation means,
Threshold processing means for removing a component equal to or higher than a preset threshold value from the frequency domain wraparound cancellation residual output by the cancellation residual calculation means;
IFFT (Inverse Fast Fourier Transform) means for performing an inverse fast Fourier transform on the wraparound cancellation residual output from the threshold processing means into a wraparound cancellation residual in the time domain;
Multiplication means for multiplying a time domain wraparound cancellation residual output by the IFFT means by a preset constant;
Addition means for adding an update of the time domain sneak cancellation residual output by the multiplication means to the filter coefficient before the unit update time used by the adaptive filter, and outputting to the adaptive filter as a new filter coefficient A wraparound canceller characterized by that. In the wraparound canceller that cancels the wraparound by combining the input OFDM signal and the wraparound replica signal generated by the adaptive filter in reverse phase,
FFT means for fast Fourier transforming the signal after the wraparound cancellation to a carrier symbol in the frequency domain;
Transmission path response calculating means for calculating a transmission path response from the carrier symbol output by the FFT means;
Threshold value processing means for replacing a component below a preset threshold value with a preset value in the transmission path response output by the transmission path response calculating means;
Cancellation residual calculation means for calculating a wraparound cancellation residual in the frequency domain from the transmission line response output by the threshold processing means;
IFFT means for inverse fast Fourier transforming the frequency domain wraparound cancellation residual output by the cancellation residual calculation means into a time domain wraparound cancellation residual;
Multiplication means for multiplying a time domain wraparound cancellation residual output by the IFFT means by a preset constant;
Addition means for adding an update of the time domain sneak cancellation residual output by the multiplication means to the filter coefficient before the unit update time used by the adaptive filter, and outputting to the adaptive filter as a new filter coefficient A wraparound canceller characterized by that.

Images