JP4780161B2 - Receiving device, receiving method, and program - Google Patents

Description

  The present invention relates to a receiving apparatus, a receiving method, and a program, and in particular, a receiving apparatus that can easily generate coefficients of an adaptive equalization filter used for demodulation of an OFDM signal from an OFDM signal in a time domain, The present invention relates to a receiving method and a program.

  As a modulation method for terrestrial digital broadcasting, a modulation method called an orthogonal frequency division multiplexing (OFDM) method is used.

  In the OFDM method, a large number of orthogonal subcarriers (subcarriers) are provided in the transmission band, data is allocated to the amplitude and phase of each subcarrier, and digital modulation is performed by PSK (Phase Shift Keying) or QAM (Quadrature Amplitude Modulation) It is a method to do.

  Since the OFDM system divides the entire transmission band by a large number of subcarriers, the band per subcarrier wave is narrowed and the transmission speed is slow, but the total transmission speed is the same as the conventional modulation system. Have. In addition, the OFDM scheme has a feature that multipath tolerance can be improved by providing a guard interval to be described later.

  Furthermore, since data is allocated to a plurality of subcarriers, the OFDM system can configure a transmission circuit by using an IFFT (Inverse Fast Fourier Transform) arithmetic circuit that performs inverse Fourier transform at the time of modulation. The reception circuit can be configured by using an FFT (Fast Fourier Transform) arithmetic circuit performed at the time of demodulation.

  Due to the above characteristics, the OFDM system is often applied to terrestrial digital broadcasting that is strongly affected by multipath interference. As terrestrial digital broadcasting standards adopting the OFDM system, for example, there are standards such as DVB-T (Digital Video Broadcasting-Terrestrial), ISDB-T (Integrated Services Digital Broadcasting-Terrestrial), and ISDB-TSB.

  FIG. 1 is a diagram illustrating OFDM symbols.

  In the OFDM system, signal transmission is performed in units called OFDM symbols.

  As shown in FIG. 1, one OFDM symbol is composed of an effective symbol that is a signal interval in which IFFT is performed at the time of transmission, and a guard interval (hereinafter referred to as GI) in which the waveform of the latter half of the effective symbol is copied. The The GI is inserted at a position before the valid symbol on the time axis.

  In the OFDM system, it is possible to prevent interference between OFDM symbols that occurs in a multipath environment by inserting a GI.

  A plurality of such OFDM symbols are collected to form one OFDM transmission frame. For example, in the ISDB-T standard, one OFDM transmission frame is formed from 204 OFDM symbols. The pilot signal insertion position is determined based on the unit of the OFDM transmission frame.

  In the OFDM system that uses the QAM modulation system as the modulation system for each subcarrier, the amplitude and phase of each subcarrier are different from those at the time of transmission and at the time of reception due to the influence of multipath, etc. during transmission. It will be different. Therefore, on the receiving side, it is necessary to equalize the signal so that the amplitude and phase of the received signal are equal to those transmitted.

  In the OFDM system, a pilot signal having a predetermined amplitude and a predetermined phase is discretely inserted in a transmission symbol on the transmission side, and the frequency characteristic of the transmission path is determined on the reception side based on the amplitude and phase of the pilot signal. The received signal is equalized according to the obtained characteristics of the transmission path.

  Thus, the pilot signal used to calculate the transmission path characteristics is referred to as a scattered pilot signal (hereinafter referred to as SP signal). FIG. 2 shows an arrangement pattern of the SP signal employed in the DVB-T standard or the ISDB-T standard in the OFDM symbol. In FIG. 2, the vertical direction is the time direction, and the horizontal direction is the frequency direction.

  FIG. 3 is a diagram illustrating a configuration example of a conventional OFDM receiver.

  As shown in FIG. 3, the OFDM receiver 1 includes a receiving antenna 11, a tuner 12, a BPF (Band Pass Filter) 13, an A / D (Analog / Digital) conversion circuit 14, an orthogonal demodulation circuit 15, an FFT circuit 16, An SP use equalization circuit 17 and an error correction circuit 18 are included.

  The receiving antenna 11 receives a broadcast wave broadcast from a broadcasting station and outputs an RF signal to the tuner 12.

  The tuner 12 includes a multiplication circuit 21 and a local oscillator 22, converts the frequency of the RF signal received by the receiving antenna 11 into an IF signal, and outputs the IF signal to the BPF 13.

  The BPF 13 performs filtering on the IF signal supplied from the tuner 12 and outputs a signal obtained by performing the filtering to the A / D conversion circuit 14.

  The A / D conversion circuit 14 performs A / D conversion on the signal supplied from the BPF 13 and outputs a digital IF signal to the quadrature demodulation circuit 15.

  The orthogonal demodulation circuit 15 obtains a baseband OFDM signal from the IF signal supplied from the A / D conversion circuit 14 by performing orthogonal demodulation using a carrier signal having a predetermined frequency (carrier frequency). This baseband OFDM signal is a so-called time-domain signal before the FFT operation.

  Hereinafter, the baseband OFDM signal before the FFT operation is referred to as an OFDM time domain signal. As a result of orthogonal demodulation, the OFDM time domain signal becomes a complex signal including a real axis component (I channel signal) and an imaginary axis component (Q channel signal). The orthogonal demodulation circuit 15 outputs the time domain OFDM signal to the FFT circuit 16.

  The FFT circuit 16 extracts a signal in the effective symbol length range by removing the signal in the GI range from the signal of one OFDM symbol in accordance with the symbol synchronization signal. The FFT circuit 16 performs an FFT operation on the extracted OFDM time domain signal to extract data orthogonally modulated on each subcarrier.

  The FFT circuit 16 outputs an OFDM signal representing the extracted data to the SP utilization equalization circuit 17. The OFDM signal output from the FFT circuit 16 is a so-called frequency domain signal after the FFT calculation. Hereinafter, the OFDM signal after the FFT operation is referred to as an OFDM frequency band signal.

  The SP utilization equalization circuit 17 calculates the transmission path characteristics of all subcarriers using the SP signal arranged as shown in FIG. 2, and based on the calculated transmission path characteristics, the OFDM frequency domain signal To compensate for distortion caused by the transmission path. The SP utilization equalization circuit 17 outputs a signal obtained by compensating for distortion due to the transmission path to the error correction circuit 18 as an equalization signal.

  The error correction circuit 18 performs deinterleave processing on the signal interleaved on the transmission side, and further performs processing such as puncture, Viterbi decoding, spread signal removal, and RS decoding. The error correction circuit 18 outputs data obtained by performing various processes as decoded data to a subsequent circuit.

  The OFDM system has a feature that demodulation processing can be performed without causing inter-symbol interference even in a multipath environment in which delay spread falls within GI by inserting GI before an effective symbol. Yes.

  However, in an environment where a long delay multipath may occur, such as a single frequency network (SFN), the delay spread may exceed the GI. In this case, inter-symbol interference and inter-carrier interference occur, and reception performance is greatly degraded.

  In order to solve this problem, an OFDM receiver 2 having a configuration as shown in FIG. 4 and an OFDM receiver 3 having a configuration as shown in FIG. 5 have been proposed. 4 and 5, the same reference numerals are given to the same components as those of the OFDM receiver 1 of FIG. 3.

  4 includes an adaptive equalization filter 31 in front of the FFT circuit 16 in addition to the configuration of the OFDM receiver 1 shown in FIG. In the OFDM receiver 2, multipath components included in the OFDM time domain signal are removed by adaptively controlling the coefficients of the adaptive equalization filter 31. The configuration of the adaptive equalization filter 31 will be described later.

  On the other hand, the OFDM receiver 3 of FIG. 5 includes an interference canceling circuit 41 in front of the FFT circuit 16 in addition to the configuration of the OFDM receiver 1 shown in FIG.

  The interference removal circuit 41 includes an adaptive equalization filter 51, a replica generation circuit 52, and a synthesis circuit 53. The OFDM time domain signal output from the orthogonal demodulation circuit 15 is input to the adaptive equalization filter 51 and the synthesis circuit 53.

  The adaptive equalization filter 51 removes the multipath component by filtering the OFDM time domain signal supplied from the orthogonal demodulation circuit 15 and outputs the OFDM time domain signal from which the multipath component has been removed to the replica generation circuit 52. .

  The replica generation circuit 52 reproduces the removed multipath component based on the OFDM time domain signal supplied from the adaptive equalization filter 51, and outputs the reproduced multipath component signal to the synthesis circuit 53.

  The combining circuit 53 uses the multipath component reproduced by the replica generation circuit 52 for the inter-symbol interference component and the inter-carrier interference component included in the FFT section of the OFDM time domain signal supplied from the orthogonal demodulation circuit 15. Remove. The combining circuit 53 outputs an OFDM time domain signal from which the inter-symbol interference component and the inter-carrier interference component included in the FFT section are removed to the FFT circuit 16.

  As described above, Patent Document 1 discloses a technique for reproducing the replica and removing the interference component.

  Here, the adaptive equalization filter 31 of FIG. 4 will be described.

  FIG. 6 is a diagram illustrating a configuration example of the adaptive equalization filter 31.

  As shown in FIG. 6, the adaptive equalization filter 31 includes a variable coefficient filter 61, an SP extraction circuit 62, an IFFT circuit 63, and a main wave component removal circuit 64. The OFDM time domain signal output from the quadrature demodulation circuit 15 is input to the variable coefficient filter 61, and the OFDM frequency domain signal output from the FFT circuit 16 is input to the SP extraction circuit 62.

  The variable coefficient filter 61 filters the OFDM time domain signal supplied from the quadrature demodulation circuit 15 using a coefficient set based on the signal supplied from the main wave component removal circuit 64, and provides the OFDM time domain signal. Multipath components contained in the signal are removed. The variable coefficient filter 61 outputs the OFDM time domain signal from which the multipath component is removed to the FFT circuit 16.

  The SP extraction circuit 62 extracts the SP signal inserted at the position shown in FIG. 2 from the frequency domain OFDM signal supplied from the FFT circuit 16 and removes the modulation component, thereby transmitting the frequency domain transmission line. Calculate the characteristics. The SP extraction circuit 62 outputs the calculated transmission path characteristic to the IFFT circuit 63.

  The IFFT circuit 63 performs IFFT calculation to convert the transmission path characteristic in the frequency domain into the impulse response characteristic of the transmission path in the time domain. The IFFT circuit 63 outputs the impulse response characteristic of the transmission line in the time domain to the main wave component removal circuit 64.

  The main wave component removal circuit 64 removes the main wave component from the time domain impulse response calculated by the IFFT circuit 63, leaves only the multipath component, and outputs a multipath component signal to the variable coefficient filter 61. In the variable coefficient filter 61, a coefficient corresponding to the amplitude and phase of the multipath component is set in the tap corresponding to the delay time of the multipath component acquired by the main wave component removal circuit 64, and the multipath component is filtered out. Removing is done.

  FIG. 7 is a diagram illustrating a configuration example of the variable coefficient filter 61 of FIG.

  As shown in FIG. 7, the variable coefficient filter 61 includes a variable coefficient FIR filter 71 and a variable coefficient IIR filter 72. A coefficient update circuit (not shown) is also provided in the variable coefficient filter 61. The OFDM time domain signal is input to the variable coefficient FIR filter 71.

  The variable coefficient FIR filter 71 removes or suppresses a multipath (hereinafter referred to as pre-echo) component that arrives earlier than the main wave by performing filtering using a coefficient generated by a coefficient update circuit (not shown).

  The variable coefficient FIR filter 71 outputs a pre-echo equalized signal, which is an OFDM time domain signal with the pre-echo component removed or suppressed, to the variable coefficient IIR filter 72. Since it is difficult to completely remove the pre-echo component, not only the signal from which the pre-echo component has been removed but also the signal in which the pre-echo component has been suppressed is regarded as a signal after pre-echo equalization.

  As shown in FIG. 7, the variable coefficient IIR filter 72 includes a variable coefficient FIR filter 81 and a subtraction circuit 82. The pre-echo equalized signal supplied from the variable coefficient FIR filter 71 is input to the subtraction circuit 82.

  The variable coefficient FIR filter 81 filters the signal output from the subtraction circuit 82 using a coefficient generated by a coefficient update circuit (not shown), and subtracts the signal obtained by performing the filtering. Output to.

  The subtracting circuit 82 subtracts the signal supplied from the variable coefficient FIR filter 81 from the pre-echo equalized signal to remove a multipath (hereinafter referred to as post-echo) component that arrives later than the main wave, An equalized time domain signal which is a signal obtained by removing the post-echo component is output. The equalization time domain signal output from the subtraction circuit 82 is input to the variable coefficient FIR filter 81 and input to the FFT circuit 16.

  The adaptive equalization filter 51 in FIG. 5 also has a configuration similar to that of the variable coefficient filter 61.

  A device that removes multipath components in the time domain, such as the OFDM receiver 2 in FIG. 4 and the OFDM receiver 3 in FIG. 5, is provided with a variable coefficient FIR filter, and controls the coefficient of the variable coefficient FIR filter. By doing so, the multipath component is removed.

  Therefore, if the coefficient is not appropriate, the multipath component cannot be completely removed, and a multipath component having a time that is an integral multiple of the delay time of the existing multipath is added. . This becomes particularly remarkable when the update of the coefficient cannot follow the fluctuation due to Doppler or the like.

  Since the SP signal is extracted from the OFDM frequency band signal and the transmission path characteristic is estimated based on the SP signal, the above algorithm is an algorithm to which the delay profile is applied.

  On the other hand, a least mean square (LMS) algorithm that minimizes a mean square error (MSE (Mean Square Error)) is known as an adaptive algorithm that does not apply a delay profile. The LMS algorithm is a method using a known reference signal, and is most widely used because of features such as good adaptive performance and low calculation amount.

  FIG. 8 is a diagram illustrating an example of a circuit configuration to which the LMS algorithm is applied.

  The circuit in FIG. 8 includes a variable coefficient FIR filter 91 and a coefficient calculation circuit 92. When the input signal is x [k], the LMS algorithm aims to reproduce the desired signal d [k] by filtering the input signal x [k] in the variable coefficient FIR filter 91.

  The variable coefficient FIR filter 91 filters the input signal x [k] using the coefficient calculated by the coefficient calculation circuit 92, and uses the signal obtained by the filtering as the output signal y [k]. Output. The output signal y [k] is output to the outside and input to the coefficient calculation circuit 92.

  The coefficient calculation circuit 92 calculates a coefficient using the LMS algorithm, and outputs the calculated coefficient to the variable coefficient FIR filter 91. An input signal x [k] is also input to the coefficient calculation circuit 92.

  In the example of FIG. 8, the coefficient calculation circuit 92 includes a subtraction circuit 101, a multiplication circuit 102, a multiplication circuit 103, an integration circuit 104, and a shift register 105.

  The subtraction circuit 101 subtracts the desired signal d [k] from the output signal y [k] that is the output of the variable coefficient FIR filter 91 to generate an error signal e [k]. The desired signal d [k] is a known signal, and the LMS algorithm is an algorithm applicable when there is such a known signal.

  The multiplier circuit 102 includes a plurality of multipliers. In each multiplier, an error signal e [k] and a signal x [k−i] obtained by delaying the input signal x [k] in each delay element of the shift register 105. ]. The multiplication circuit 102 outputs a sample correlation value, which is a multiplication result of each multiplier, to the multiplication circuit 103.

  The multiplier circuit 103 includes a plurality of multipliers. In each multiplier, the sample correlation value calculated by the multiplier circuit 102 is multiplied by the step size μ, and the multiplication result is output to the integration circuit 104.

  The integration circuit 104 is composed of a plurality of integrators, and in each integrator, a coefficient is generated by integrating the multiplication result of the multiplication circuit 103. In the integration circuit 104, for example, a coefficient that cancels a sample correlation value representing the correlation between the error signal e [k] and the input signal x [k] is generated. The integration circuit 104 sets the generated coefficient to each tap of the variable coefficient FIR filter 91.

The above processing performed in the coefficient calculation circuit 92 is expressed as follows. In the following equation (1), γj [k] represents a coefficient set for each tap of the variable coefficient FIR filter 91.

As a variation of the LMS algorithm as described above, Leaky- is an algorithm for performing integration processing in the form including a Leak component when performing integration processing in the integration circuit 104 for the purpose of suppressing coefficient divergence or the like. The LMS algorithm is also known. The Leaky-LMS algorithm is expressed by the following equation (4) using the parameter λ.

For the purpose of improving the tracking performance, a VSS (Variable Step Size) -LMS algorithm, which is an algorithm that makes the step size μ variable according to an error signal, is also known. As an update algorithm for the step size μ, algorithms represented by the following expressions (5) and (6) are known.

  As described above, in the circuit using the LMS algorithm, the sample correlation value of the error signal e [k] and the input signal x [k] is obtained and updated in the direction to cancel the sample correlation value. Eventually, the error signal e [k] and the input signal x [k] converge to a value that is uncorrelated.

JP 2007-6067 A

  Since the accuracy of the coefficient of the variable coefficient FIR filter has a large effect on the reception performance, the algorithm beyond the generation of the equalized signal by removing the multipath component in the time domain using the adaptive equalization filter has an error in the coefficient. If this occurs, the reception performance deteriorates.

  Also, in order to cope with fast fluctuations in the transmission path, it is preferable to follow the coefficient as quickly as possible to the fluctuation in the transmission path, but in general, there is a trade-off relationship between the followability and stability of the coefficient. If the coefficient followability is improved, the stability is lowered.

  By the way, the algorithm that estimates the delay profile and applies the delay profile to the coefficient of the variable coefficient filter 61 as employed in the OFDM receiver 2 of FIG. 6 depends greatly on the estimation accuracy of the delay profile. Have the problem of being.

  The LMS algorithm can be considered as a method for solving this problem. As described above, the LMS algorithm is an algorithm that can be applied only when there is a known signal.

  Therefore, in the terrestrial digital broadcasting standards such as DVB-T and ISDB-T, since the OFDM time domain signal does not include a known signal, the LMS algorithm cannot be applied as it is.

  The present invention has been made in view of such a situation.For example, even when a known signal is not included, the coefficient of the adaptive equalization filter used for demodulation of the OFDM signal is calculated from the OFDM signal in the time domain. It can be easily generated.

A receiving apparatus according to an aspect of the present invention includes an OFDM signal receiving unit that receives an OFDM signal, and a predetermined number of taps in which a variable first coefficient is set, and the time received by the OFDM signal receiving unit. A first filter means for removing pre-echo components by applying an adaptive filter to the OFDM signal in the region and generating a pre-echo equalized signal; and a post-echo equalization signal generated by the first filter means Subtracting means for subtracting the echo component signal to generate an equalized signal, and a predetermined number of taps in which a variable second coefficient is set, and the equalized signal generated by the subtracting means By applying an adaptive filter to the second filter means for generating the post-echo component signal, the time-domain OFDM signal received by the OFDM signal receiving means, and the first filter means. Based on the pre-echo equalized signal generated by the above, a coefficient generating means for generating the first and second coefficients, and an FFT operation on the equalized signal generated by the subtracting means, FFT calculation means for generating an OFDM signal in the frequency domain. The coefficient generation means includes a first extraction means for extracting a signal in a section including a GI of the main wave from the pre-echo equalized signal generated by the first filter means, and the first extraction means. From the pre-echo equalization signal, third filter means for generating a dummy signal of a post-echo component by applying an adaptive filter to the extracted signal using the same coefficient as the second coefficient, An error signal generating means for subtracting the dummy signal generated by the third filter means to generate an error signal representing an error of the second coefficient; and a signal in a section including a copy source of the main wave GI. A second extraction unit that extracts the pre-echo equalized signal and outputs it as a reference signal; the error signal generated by the error signal generation unit; and the second extraction unit Post echo coefficient updating means for updating the second coefficient based on the correlation value of the reference signal that has been output can be provided.

  The coefficient generation means extracts a signal in a section including the GI copy source of the main wave from the time-domain OFDM signal and outputs it as a reference signal, and extracts by the third extraction means Pre-echo coefficient updating means for updating the first coefficient based on a correlation value between the reference signal and the pre-echo equalized signal in the section including the main wave GI. Can be provided.

A receiving method or program according to one aspect of the present invention receives an OFDM signal, has a predetermined number of taps in which a variable first coefficient is set, and applies an adaptive filter to the received time-domain OFDM signal To remove the pre-echo component, generate a pre-echo equalized signal, subtract the post-echo component signal from the generated pre-echo equalized signal, generate an equalized signal, and set a variable second coefficient The post-echo component signal is generated by applying an adaptive filter to the generated post-equalization signal, and the received time-domain OFDM signal and the pre-echo equalization signal Based on the above, the first and second coefficients are generated, an FFT operation is performed on the generated equalized signal, an OFDM signal in the frequency domain is generated, and a signal in a section including the GI of the main wave is generated. Pre-echo equalization A dummy signal of a post-echo component is generated by applying an adaptive filter to the extracted signal using the same coefficient as the second coefficient, and the dummy signal is generated from the pre-echo equalized signal. Is generated, an error signal representing the error of the second coefficient is generated, a signal in a section including the copy source of the main wave GI is extracted from the pre-echo equalized signal, output as a reference signal, and the error Updating the second coefficient based on a correlation value between the signal and the reference signal .

In one aspect of the present invention, a pre-echo component is received by receiving an OFDM signal, having a predetermined number of taps for which a variable first coefficient is set, and applying an adaptive filter to the received time-domain OFDM signal Are removed, and a signal after pre-echo equalization is generated. In addition, a post-echo component signal is subtracted from the generated pre-echo equalized signal to generate an equalized signal. The post-echo component signal is generated by applying an adaptive filter to the generated equalized signal having a predetermined number of taps in which a variable second coefficient is set, and the received time domain OFDM Based on the signal and the pre-echo equalized signal, the first and second coefficients are generated, and an FFT operation is performed on the generated equalized signal to generate a frequency-domain OFDM signal. . Further, a signal in a section including the main wave GI is extracted from the pre-echo equalized signal, and an adaptive filter is applied to the extracted signal using the same coefficient as the second coefficient, thereby performing post-echo. A component dummy signal is generated, the dummy signal is subtracted from the pre-echo equalized signal, an error signal representing the error of the second coefficient is generated, and a signal in a section including the copy source of the main wave GI is obtained. Extracted from the pre-echo equalized signal and output as a reference signal, and the second coefficient is updated based on the correlation value between the error signal and the reference signal.

  According to one aspect of the present invention, coefficients of an adaptive equalization filter used for demodulation of an OFDM signal can be easily generated from a time-domain OFDM signal.

  FIG. 9 is a diagram illustrating a configuration example of an OFDM receiving apparatus according to an embodiment of the present invention.

  As shown in FIG. 9, the OFDM receiving apparatus 201 includes a receiving antenna 211, a tuner 212, a BPF 213, an A / D conversion circuit 214, an orthogonal demodulation circuit 215, an adaptive equalization filter 231, an FFT circuit 216, and an SP utilization equalization circuit. 217 and an error correction circuit 218.

  The reception antenna 211 receives a broadcast wave broadcast from a broadcast station and outputs an RF signal to the tuner 212.

  The tuner 212 includes a multiplication circuit 221 and a local oscillator 222, converts the frequency of the RF signal received by the reception antenna 211 into an IF signal, and outputs the IF signal to the BPF 213.

  The BPF 213 performs filtering on the IF signal supplied from the tuner 212 and outputs a signal obtained by performing the filtering to the A / D conversion circuit 214.

  The A / D conversion circuit 214 performs A / D conversion on the signal supplied from the BPF 213 and outputs a digital IF signal to the quadrature demodulation circuit 215.

  The orthogonal demodulation circuit 215 obtains a baseband OFDM signal from the IF signal supplied from the A / D conversion circuit 214 by performing orthogonal demodulation using a carrier signal having a predetermined frequency (carrier frequency). The orthogonal demodulation circuit 215 outputs the time domain OFDM signal to the adaptive equalization filter 231.

  The adaptive equalization filter 231 includes variable coefficient FIR filters 241 and 242, a subtraction circuit 243, and a filter coefficient generation circuit 244. The OFDM time domain signal output from the orthogonal demodulation circuit 215 is input to the variable coefficient FIR filter 241 and the filter coefficient generation circuit 244.

  Note that when the profile assumed by the OFDM receiver 201 is only post-echo, for example, the variable coefficient FIR filter 241 may not be provided.

  The variable coefficient FIR filter 241 performs filtering on the OFDM time domain signal supplied from the orthogonal demodulation circuit 215 using a pre-coefficient that is a pre-echo equalization coefficient generated by the filter coefficient generation circuit 244, Remove or suppress the pre-echo component contained in the OFDM time domain signal. The variable coefficient FIR filter 241 outputs the pre-echo equalized signal to the subtraction circuit 243 and the filter coefficient generation circuit 244.

  The variable coefficient FIR filter 242 filters the OFDM time domain signal supplied from the subtraction circuit 243 using a post coefficient that is a post-echo equalization coefficient generated by the filter coefficient generation circuit 244. The variable coefficient FIR filter 242 outputs a post-echo component signal obtained by filtering to the subtraction circuit 243.

  The subtraction circuit 243 subtracts the signal supplied from the variable coefficient FIR filter 242 from the pre-echo equalized signal to remove the post-echo component included in the pre-echo equalized signal and remove the post-echo component OFDM Output time domain signal. The signal output from the subtraction circuit 243 is input to the FFT circuit 216 and simultaneously input to the variable coefficient FIR filter 242 in order to generate a signal used to remove the post-echo component at the next time.

  The filter coefficient generation circuit 244 generates pre-coefficients and post-coefficients based on the pre-echo removal OFDM time domain signal supplied from the quadrature demodulation circuit 215 and the pre-echo equalization signal supplied from the variable coefficient FIR filter 241. Generate. The filter coefficient generation circuit 244 outputs the generated pre coefficient to the variable coefficient FIR filter 241 and outputs the post coefficient to the variable coefficient FIR filter 242. Details of the filter coefficient generation circuit 244 will be described later.

  The FFT circuit 216 extracts an effective symbol length range signal by excluding a signal in the GI range from one OFDM symbol signal in accordance with the symbol synchronization signal, and performs an FFT operation on the extracted OFDM time domain signal. The FFT circuit 216 outputs the OFDM frequency band signal obtained by performing the FFT operation to the SP use equalization circuit 217.

  The SP utilization equalization circuit 217 calculates the transmission path characteristics of all subcarriers using the SP signal, and compensates for distortion of the OFDM frequency band signal due to the transmission path based on the calculated transmission path characteristics. The SP utilization equalization circuit 217 outputs a signal obtained by compensating for distortion due to the transmission path to the error correction circuit 218 as an equalization signal.

  The error correction circuit 218 performs deinterleaving processing on the signal interleaved on the transmission side, and further performs processing such as puncture, Viterbi decoding, spread signal removal, and RS decoding. The error correction circuit 218 outputs data obtained by performing various kinds of processing to the subsequent circuit as decoded data.

  In the example of FIG. 9, only the adaptive equalization filter 231 is provided in front of the FFT circuit 216. However, a removal circuit that includes an adaptive equalization filter and removes multipath interference is shown in FIG. May be provided.

  FIG. 10 is a diagram illustrating a configuration example of the adaptive equalization filter 231 in FIG. The same components as those shown in FIG. 9 are denoted by the same reference numerals. The overlapping description will be omitted as appropriate.

  As shown in FIG. 10, the filter coefficient generation circuit 244 of the adaptive equalization filter 231 includes a main wave position detection circuit 251, a signal extraction circuit 252, a variable coefficient FIR filter 253, a subtraction circuit 254, a delay circuit 255, and a coefficient update circuit. 256, a signal extraction circuit 257, a delay circuit 258, a coefficient update circuit 259, and a signal extraction circuit 260. The OFDM time domain signal output from the quadrature demodulation circuit 215 is input to the signal extraction circuit 260, and the pre-echo equalized signal output from the variable coefficient FIR filter 241 is subtracted from the main wave position detection circuit 251, the signal extraction circuit 252, and the subtraction. The data is input to the circuit 254, the signal extraction circuit 257, and the delay circuit 258.

  Here, generation of coefficients performed by the filter coefficient generation circuit 244 will be described with reference to FIGS. 11 to 14 as appropriate. For convenience of explanation, first, generation of a post coefficient will be described, and then generation of a pre coefficient will be described.

  Generation of the post coefficient is basically performed by the main wave position detection circuit 251, the signal extraction circuit 252, the variable coefficient FIR filter 253, the subtraction circuit 254, the delay circuit 255, the coefficient update circuit 256, and the signal extraction circuit 257. .

  As described with reference to FIG. 7, the processing for removing the post-echo is often performed using an IIR filter. However, when a configuration for performing coefficient calculation processing is incorporated on the feedback path of the IIR filter, since there is a possibility of falling into a local optimum point, care in consideration of this is necessary.

  In the configuration of FIG. 10, by using the variable coefficient FIR filter 253 in addition to the variable coefficient FIR filter 242, when the line from the variable coefficient FIR filter 241 to the FFT circuit 216 via the subtraction circuit 243 is brought online, the line is offline. A coefficient is calculated. The configuration for performing the coefficient calculation processing as described below can be incorporated into a variable coefficient IIR filter including a variable coefficient FIR filter and a subtraction circuit.

  The main wave position detection circuit 251 detects a predetermined position (time) of the reference main wave, such as the start position of the main wave GI, from the pre-echo equalized signal supplied from the variable coefficient FIR filter 241. The configuration and position detection of the main wave position detection circuit 251 will be described later with reference to FIGS. Main wave position detection circuit 251 outputs a signal representing the detected position to signal extraction circuit 252, signal extraction circuit 257, and signal extraction circuit 260.

  The signal extraction circuit 252 extracts a signal in a section including the main wave GI from the pre-echo equalized signal supplied from the variable coefficient FIR filter 241 in accordance with the position detected by the main wave position detection circuit 251. The signal extraction circuit 252 outputs the extracted signal of the section including the main wave GI to the variable coefficient FIR filter 253.

  FIG. 11 is a diagram illustrating an example of signals generated in each unit.

The signal S ₁ shown at the top of FIG. 11 is a signal after pre-echo equalization in an initial state (a state where no post coefficient is generated) obtained in a two-wave environment. The pre-echo equalized signal S ₁ includes a main wave component and a post-echo component. The upper band of the pre-echo equalized signal S ₁ represents the main wave, and the lower band represents the post-echo.

In FIG. 11, the fact that the width of the band representing the post echo is narrower than the width of the band representing the main wave indicates that the amplitude of the post echo is smaller than the amplitude of the main wave. The horizontal direction in FIG. 11 represents the time direction. In this example, a delay occurs for a time corresponding to the interval from the arrow A ₁ to the arrow A ₂ . Focusing on one OFDM symbol indicating a section by an arrow in FIG. 11, the start position of the OFDM symbol of interest transmitted by the main wave becomes the position of the arrow A _1, the start position of the OFDM symbol of interest transmitted by the post-echo Is the position of arrow A ₂ .

In the signal S ₂ shown in the second stage from the top in FIG. 11, the signal extraction circuit 252 has a predetermined number of 0s in the signal in the section W ₁ including the main wave GI extracted from the pre-echo equalized signal S _1. It is a signal obtained by adding only. For example, the signal extraction circuit 252 adds 0 to the series of the interval W ₁ . This signal S ₂ is supplied from the signal extraction circuit 252 to the variable coefficient FIR filter 253. By adding 0, signal S ₂ will longer signal than GI length on the time axis.

  Returning to the description of FIG. 10, the variable coefficient FIR filter 253 performs filtering on the signal supplied from the signal extraction circuit 252. In the initial state, a predetermined coefficient generated by the coefficient update circuit 256 is used for filtering. The variable coefficient FIR filter 253 outputs a signal obtained by performing filtering to the subtraction circuit 254 as a multipath dummy signal.

A signal S ₃ shown in the third row from the top in FIG. 11 is a multipath dummy signal generated by the variable coefficient FIR filter 253. As shown in FIG. 11, the multipath dummy signal S ₃ is converted in amplitude and phase by multiplying the signal S ₂ whose position is sequentially shifted in the time direction by the coefficient of the tap corresponding to each position. The signal obtained by converting the amplitude and phase is expressed as a signal obtained by adding the signals.

  The signal shown as “tap 0 output” in FIG. 11 represents the signal output from the tap of tap number 0 of the variable coefficient FIR filter 253, and the signal shown as “tap 1 output” is the tap of the variable coefficient FIR filter 253. The signal output from the tap of number 1 is represented. The variable coefficient FIR filter 253 has the same configuration as that of the variable coefficient FIR filter 91 shown in FIG. If a set of one delay element and one multiplier is one tap, the tap number 0, the tap number 1, the tap number 2, the tap number 2, and so on in order from the left tap in FIG. .

The multipath dummy signal S ₃ expressed in this way is supplied from the variable coefficient FIR filter 253 to the subtraction circuit 254.

  The subtracting circuit 254 subtracts the multipath dummy signal from the pre-echo equalized signal to generate an error signal. The subtraction circuit 254 outputs the generated error signal to the delay circuit 255.

The signal S ₄ shown in the fourth stage from the top in FIG. 11 is an error signal obtained by subtracting the multipath dummy signal S ₃ from the signal for the range of ₁ OFDM symbol in the pre-echo equalized signal S _1. Represents. Error signal S ₄ is mainly composed of three signal components of claim d0 to d2.

The term 0 is a term obtained by removing the post-echo GI from the post-echo equalization signal S ₁ .

  The term d1 is a term including a main wave GI and a post-echo component generated when the coefficient of the tap having no multipath (post-echo) is non-zero.

  Here, a tap having a multipath is a tap for multiplying a signal delayed in a delay element of the FIR filter by a time corresponding to a multipath delay time. A tap without a multipath is a tap other than a tap with a multipath among taps of the FIR filter.

The term d2 is obtained by multiplying the difference between the tap coefficient having a multipath and the optimum coefficient corresponding to the actual multipath amplitude and phase by the difference between the coefficient and the GI of the main wave. Term. Since contains this section d2, the error signal S ₄ is a coefficient of the tap there is multipath, a signal representative of the difference between the actual optimal coefficient corresponding to the amplitude and phase of the multipath.

The error signal S ₄ including such a component is supplied from the subtraction circuit 254 to the delay circuit 255.

  The delay circuit 255 delays the error signal supplied from the subtraction circuit 254 by a time corresponding to an interval one sample less than the effective symbol length with reference to a position one sample after the start position of the main OFDM symbol. The delayed error signal is output to the coefficient update circuit 256. The position serving as a reference for delay and the time corresponding to one sample are appropriately set according to the characteristics of the transmission line.

A signal S ₅ shown in the fifth stage from the top in FIG. 11 represents the error signal S ₄ delayed in the delay circuit 255. For the term d0, a range having the same length as the term d1 indicated by a bold line is extracted. The start position of the signal S ₅ coincides with the start position of the section from which the main wave GI is copied.

  On the other hand, the signal extraction circuit 257 to which the pre-echo equalized signal output from the variable coefficient FIR filter 241 is input, from the pre-echo equalized signal according to the position detected by the main wave position detection circuit 251, The signal of the section including the copy source of is extracted. The signal extraction circuit 257 outputs the signal in the section including the copy source of the main wave GI to the coefficient update circuit 256 as a reference signal in accordance with the output of the signal by the delay circuit 255.

A signal S ₆ shown in the sixth stage from the top in FIG. 11 is a reference signal extracted by the signal extraction circuit 257. In the example of FIG. 11, the signal in the section W ₂ including the copy source of the main wave GI is extracted from the post-echo equalization signal S ₁ .

  The coefficient update circuit 256 generates a post coefficient based on the error signal supplied from the delay circuit 255 and the reference signal supplied from the signal extraction circuit 257, and the generated post coefficient is variable with the variable coefficient FIR filter 242. The coefficient FIR filter 253 is output.

  FIG. 12 is a diagram illustrating a configuration example of the coefficient update circuit 256.

  As shown in FIG. 12, the coefficient update circuit 256 includes a selector 271, a shift register 272, a multiplier circuit 273, an integrator circuit 274, a multiplier circuit 275, and an integrator circuit 276. The error signal output from the delay circuit 255 is input to each multiplier of the multiplication circuit 273, and the reference signal output from the signal extraction circuit 257 is input to the selector 271.

  The selector 271 adds a predetermined number of 0s to the second half of the reference signal so as to have the same length as the delay signal. The selector 271 outputs a reference signal added with 0 to the shift register 272.

  The shift register 272 includes a plurality of delay elements, and sequentially delays the reference signal supplied from the selector 271 in each delay element. Data stored in each delay element is used to update the coefficient of the corresponding tap.

  The multiplier circuit 273 is composed of a plurality of multipliers, and each multiplier multiplies the error signal and the reference signal delayed by each delay element of the shift register 272 to calculate a sample correlation value. The multiplication circuit 273 outputs the calculated sample correlation value to the integration circuit 274.

  The integrating circuit 274 includes a plurality of integrators, and each of the integrators integrates the sample correlation value calculated by the multiplying circuit 273 for the reference signal interval. The integration circuit 274 outputs the integration result of the sample correlation value to the multiplication circuit 275. By performing the integration process in this manner, the accuracy of the sample correlation value can be improved. The integration process by the integration circuit 274 is reset every time the target OFDM symbol is switched.

  The multiplier circuit 275 includes a plurality of multipliers. In each multiplier, the integration result of the sample correlation value calculated by the integration circuit 274 is multiplied by the step size μ, and the multiplication result is output to the integration circuit 276 as a coefficient update value. To do.

  The integrating circuit 276 is composed of a plurality of integrators. Each of the integrators integrates the coefficient update value calculated by the multiplying circuit 275, and outputs the integration result as a post coefficient.

  For example, the multiplier at the left end of the multiplier circuit 273, the integrator at the left end of the integrator circuit 274, the multiplier at the left end of the multiplier circuit 275, and the integrator at the left end of the integrator circuit 276 have the tap number 0 of the variable coefficient FIR filters 242 and 253. The tap coefficient is generated. The second multiplier from the left of the multiplier circuit 273, the second integrator from the left of the integrator circuit 274, the second multiplier from the left of the multiplier circuit 275, and the second integrator from the left of the integrator circuit 276 are: The tap coefficient of tap number 1 of the variable coefficient FIR filters 242 and 253 is generated. Multipliers and integrators are provided according to the number of taps.

  The principle of coefficient generation (updating) performed by the coefficient updating circuit 256 having the above configuration will be described with reference to FIGS.

  Coefficients are generated by assuming that a given interval of the OFDM time domain signal is one sample, but only the sample in the interval in the GI and the sample in the interval in the GI copy source have a large correlation, but the correlation between the other samples is extremely high. This is done using the characteristic of the OFDM time domain signal that it is small. Since the GI and its copy source are the same signal, the correlation is obtained in this way.

  FIG. 13 is a diagram illustrating an example of a signal used for generating a coefficient of a tap having no multipath.

When there is no multipath in the tap with tap number 0 and the coefficient set for the tap is non-zero, a point between the error signal S ₅ and the reference signal S ₆ is indicated at the tip of the arrow A _11. as such, the correlation results in accordance with the component section d1 of the error signal S _5.

Terms other than the term d1 of the error signal S ₅ can be regarded as noise terms whose correlation with the reference signal S ₆ is 0 on average. Term d11 represents the correlation between the noise terms and the reference signal S _6.

  The sample correlation values thus obtained are processed in the multiplier circuit 273, the integrator circuit 274, the multiplier circuit 275, and the integrator circuit 276, respectively, and a tap coefficient with a tap number of 0 is generated.

  Since there is no multipath in tap number 0, the coefficient is updated in the direction to cancel the sample correlation value, as in the LMS algorithm. That is, the tap coefficient of tap number 0 is always controlled to go to 0.

  Not only the tap with tap number 0, but the coefficients of all taps having no multipath are updated in the same manner.

  FIG. 14 is a diagram illustrating an example of a signal used for generating a coefficient of a tap having a multipath.

The difference from FIG. 13 is set between the error signal S ₅ and the reference signal S ₆ , as shown at the tip of the arrow A _{21 in} FIG. 14, and a coefficient corresponding to the amplitude and phase of the actual multipath. That is, a correlation corresponding to the component of the term d2 depending on the difference between the coefficients is generated.

In this way, the multiplication circuit 273 to the sample correlation values obtained, the integrating circuit 274, multiplication circuit 275, are subjected to respective processing in the integrator circuit 276, the coefficient in a direction to cancel the correlation term d2 and the reference signal S ₆ Updated. Eventually, coefficients corresponding to the amplitude and phase of the actual multipath are generated.

  Coefficients that are sequentially generated as described above are supplied to both the variable coefficient FIR filter 242 and the variable coefficient FIR filter 253 and updated, thereby performing post-echo removal processing and gradually increasing the coefficient accuracy. Such an operation is realized.

  Next, generation of the pre coefficient will be described.

  The pre-coefficient generation is basically performed by the main wave position detection circuit 251, the delay circuit 258, the coefficient update circuit 259, and the signal extraction circuit 260.

  The pre coefficient is also generated in the same procedure as the post coefficient. The configuration related to the post coefficient is provided with two variable coefficient FIR filters, a variable coefficient FIR filter 253 that is a coefficient generation filter and a variable coefficient FIR filter 242 that is a multipath removal filter. Since the configuration relating to the pre-use coefficient is merely provided with the variable coefficient FIR filter 241, there is a difference in the procedure.

  Since the output signal of the variable coefficient FIR filter 241 is used for generating a coefficient, the variable coefficient FIR filter 241 in FIG. 10 can be regarded as sharing the FIR filter for coefficient generation and multipath (pre-echo) removal. it can. Similarly to the post coefficient, it is also possible to generate the pre coefficient off-line by separately using a variable coefficient FIR filter for coefficient generation.

  The variable coefficient FIR filter 241 has a configuration similar to that of the variable coefficient FIR filter 71 shown in FIG. The coefficient of the final tap is 1, and a coefficient that removes the pre-echo component included in the signal output from the final tap is generated.

  In order to remove or suppress the pre-echo component contained in the signal output from the final tap, the amplitude of the main signal of the future signal is set to the multipath amplitude for each signal in order from the final tap signal. Is synthesized with the same amplitude as. Then, instead of removing or suppressing the pre-echo component, an attenuated multipath component having a delay time that is twice the actual multipath delay time is generated.

  Therefore, this newly generated multipath component must be removed at the tap of the variable coefficient FIR filter 241 that targets the signal delayed in the delay element by twice the actual multipath delay time.

  Based on the above, the output of the variable coefficient FIR filter 241 will be described with reference to FIG.

A signal S ₁₁ shown at the top of FIG. 15 is an OFDM time domain signal including a pre-echo component. This OFDM time domain signal S ₁₁ is input to the variable coefficient FIR filter 241. The upper band represents the pre-echo and the lower band represents the main wave.

The signal S ₁₂ represents the output of a tap intended for a signal delayed in the delay element by a time twice as long as the multipath delay time, and the signal S ₁₃ represents the output of a tap with multipath.

Signal S ₁₄ represents the output signal of the final tap of the variable coefficient FIR filter 241. Since the coefficient of the final tap of the variable coefficient FIR filter 241 is 1, its output is simply a delayed version of the OFDM time domain signal S ₁₁ . An arrow A _{31 in} FIG. 15 represents that the OFDM time domain signal S ₁₁ is delayed in all the delay elements of the variable coefficient FIR filter 241.

Signal S ₁₅ represents the output when the coefficient of the tap without multipath is non-zero.

The signal shown at the bottom of FIG. 15 represents the post-echo equalization signal S ₁ . The pre-echo equalized signal S ₁ is obtained by synthesizing each signal in order from the signal of the final tap with the same amplitude as the multipath amplitude of the future signal for the same time as the delay time. Arrows A ₃₂ and A _{33 in} FIG. 15 represent the combination of the main wave and the multipath.

The post-echo equalization signal S ₁ has a delay time of the main wave component, a multipath component that remains without being removed, and a delay time that is twice the delay time of the multipath component that cannot be removed. It consists of components and components that occur when the coefficients of taps without multipath are non-zero.

The pre-echo equalized signal S ₁ including such components is input to the main wave position detection circuit 251, the delay circuit 258, and the signal extraction circuit 260 as a signal used for generating the pre coefficient. The signal extraction circuit 260 also receives the OFDM time domain signal output from the quadrature demodulation circuit 215.

  The main wave position detection circuit 251 detects a predetermined position of the main wave as described above. A signal representing the position detected by the main wave position detection circuit 251 is input to the signal extraction circuit 260.

  The delay circuit 258 extracts the signal corresponding to the number of taps of the variable coefficient FIR filter 241 (data of the same number of samples as the number of taps) from the pre-echo equalized signal based on the GI of the main wave. The pre-echo equalized signal supplied from the variable coefficient FIR filter 241 is delayed so as to coincide with the start point of the signal in the section including the GI copy source of the main wave extracted by the circuit 260. The delay circuit 258 outputs the delayed pre-echo equalized signal to the coefficient update circuit 259.

  In the coefficient update circuit 259, the data range corresponding to the number of taps of the variable coefficient FIR filter 241 determined with reference to the start position of the main wave GI among the delayed pre-echo equalization signals is treated as an error signal. Since the main wave GI start position is detected by the main wave position detection circuit 251 using the pre-echo equalized signal, in order to extract the error signal range from the pre-echo equalized signal, such a variable coefficient is used. Adjustment for the delay time of the FIR filter 241 is required.

  The signal extraction circuit 260 extracts, from the OFDM time domain signal supplied from the quadrature demodulation circuit 215, a signal in a section including the GI copy source of the main wave according to the position detected by the main wave position detection circuit 251. The signal extraction circuit 260 outputs the signal in the section including the copy source of the main wave GI to the coefficient update circuit 259 as a reference signal.

  The coefficient update circuit 259 has a configuration similar to that of the coefficient update circuit 256 shown in FIG. The coefficient update circuit 259 includes an error signal in the range of data corresponding to the number of taps of the variable coefficient FIR filter 241 determined with reference to the start position of the main wave GI in the delayed pre-echo equalization signal, and a signal extraction circuit A sample correlation value of the reference signal supplied from 260 is calculated, and a pre coefficient is generated. When the range of the error signal is cut out from the signal after pre-echo equalization, the position detected by the main wave position detection circuit 251 may be referred to as necessary.

  Specifically, the coefficient update circuit 259 generates a coefficient that cancels the multipath component for a tap having a multipath and a tap corresponding to a delay of a constant multiple thereof. In addition, the coefficient update circuit 259 has a correlation when the coefficient that should be 0 is actually non-zero for the other taps. Therefore, the coefficient update circuit 259 updates the coefficient in a direction to cancel it, and converges the coefficient to 0. .

  The coefficient update circuit 259 outputs the generated pre coefficient to the variable coefficient FIR filter 241.

  FIG. 16 is a diagram illustrating an example of a signal used for generating the pre coefficient.

The signal shown at the top of FIG. 16 is an OFDM time domain signal. In the signal extraction circuit 260, a signal in a section including the copy source of the main wave GI is extracted from the OFDM time domain signal. In the example of FIG. 16, the range indicated by the section W ₃ is extracted, and the signal of the section W ₃ is input to the coefficient update circuit 259 as a reference signal. A signal S ₈ shown at the bottom of FIG. 16 is a reference signal.

The signal shown in the second stage from the top in FIG. 16 is the pre-echo equalized signal S ₁ . In the delay circuit 258, to delay the pre-echo equalized signal S ₁ is performed, among the pre-echo equalized signals S ₁ delayed tap number of the data of the variable coefficient FIR filter 241 is an error signal In the coefficient update circuit 259.

A signal S ₇ shown in the third row from the top in FIG. 16 is an error signal. In this example, the position of the data corresponding to the number of taps of the variable coefficient FIR filter 241 before the reference position is determined with reference to the position that is one sample earlier than the end position of the main wave GI. It is a signal.

  The pre-coefficient is generated as described above.

  FIG. 17 is a diagram illustrating a configuration example of the main wave position detection circuit 251. The detection of the main wave GI start position will be described with reference to FIG. 18 as appropriate.

  As shown in FIG. 17, the main wave position detection circuit 251 includes an effective symbol length delay circuit 291, a complex conjugate calculation circuit 292, a multiplication circuit 293, a GI length moving average calculation circuit 294, an absolute value calculation circuit 295, and a maximum position. The search circuit 296 is configured. The pre-echo equalized signal output from the variable coefficient FIR filter 241 is input to the effective symbol length delay circuit 291 and the multiplier circuit 293.

  The effective symbol length delay circuit 291 delays the pre-echo equalization signal supplied from the variable coefficient FIR filter 241 by the effective symbol length, and outputs the delayed pre-echo equalization signal to the complex conjugate arithmetic circuit 292.

A signal S ₂₁ shown at the top of FIG. 18 is a pre-echo equalized signal including a main wave component and a post-echo component, which is input to the effective symbol length delay circuit 291. FIG. 18 shows an example of a signal in a two-wave environment.

A signal S ₂₂ shown in the second row from the top in FIG. 18 represents a pre-echo equalized signal delayed by an effective symbol length.

  The complex conjugate arithmetic circuit 292 calculates a conjugate complex number using the pre-echo equalized signal supplied from the effective symbol length delay circuit 291, and outputs the calculated conjugate complex number to the multiplier circuit 293.

  The multiplication circuit 293 obtains an autocorrelation value for each sample of the input signal of the main wave position detection circuit 251 by multiplying the post-echo equalization signal by the conjugate complex number calculated by the complex conjugate arithmetic circuit 292. The multiplication circuit 293 outputs the autocorrelation value for each sample to the GI length moving average calculation circuit 294.

  The GI length moving average calculation circuit 294 calculates a moving average of GI lengths of autocorrelation values for each sample. In the GI length moving average calculation circuit 294, a GI length window is set.

FIG signal S ₂₃ shown in the third row from the top of the 18 shows calculated by the GI length moving average calculation circuit 294, the moving average of the GI length of the autocorrelation values for each sample. Thus, the moving average of the GI length of the autocorrelation value for each sample is obtained as a signal obtained by synthesizing the triangular wave.

  The GI length moving average calculation circuit 294 outputs the calculated moving average to the absolute value calculation circuit 295.

  The absolute value calculation circuit 295 calculates the absolute value of the moving average supplied from the GI length moving average calculation circuit 294. The absolute value calculation circuit 295 outputs the calculated absolute value to the maximum position search circuit 296.

  The maximum position search circuit 296 detects a point at which the autocorrelation value is maximum based on the absolute value of the moving average calculated by the absolute value calculation circuit 295. The detected position represents the start position of the main wave GI (the boundary position of the OFDM symbol). The maximum position search circuit 296 outputs a signal representing the detected position to each circuit.

  The position represented by the signal output from the maximum position search circuit 296 is used to generate a post coefficient and a pre coefficient.

  Here, processing of the OFDM receiving apparatus 201 having the above configuration will be described.

  First, the OFDM demodulation processing of the OFDM receiver 201 will be described with reference to the flowchart of FIG. The processing of each step is appropriately performed in parallel with other processing or before and after other processing. The processing of each step in FIGS. 20 to 22 is the same.

  In step S <b> 1, the tuner 212 converts the RF signal received by the receiving antenna 211 into an IF signal and outputs the IF signal to the BPF 213.

  In step S <b> 2, the BPF 213 performs filtering on the IF signal and outputs a signal obtained by performing the filtering to the A / D conversion circuit 214.

  In step S <b> 3, the A / D conversion circuit 214 performs A / D conversion on the signal supplied from the BPF 213 and outputs a digital IF signal to the quadrature demodulation circuit 215.

  In step S <b> 4, the orthogonal demodulation circuit 215 generates an OFDM time domain signal by performing orthogonal demodulation, and outputs the generated OFDM time domain signal to the adaptive equalization filter 231.

  In step S5, the adaptive equalization filter 231 performs time domain equalization processing. The OFDM time domain signal obtained by the time domain equalization process is output from the adaptive equalization filter 231 to the FFT circuit 216. The time domain equalization process will be described later with reference to the flowchart of FIG.

  In step S 6, the FFT circuit 216 performs an FFT operation on the OFDM time domain signal supplied from the adaptive equalization filter 231, and outputs the OFDM frequency domain signal to the SP utilization equalization circuit 217.

  In step S7, the SP use equalization circuit 217 calculates transmission path characteristics of all subcarriers using the SP signal, and compensates for distortion due to the transmission path included in the OFDM frequency band signal. The SP utilization equalization circuit 217 outputs an equalization signal obtained by compensating for distortion caused by the transmission path to the error correction circuit 218.

  In step S8, the error correction circuit 218 performs various processes such as a deinterleave process on the equalized signal supplied from the SP use equalization circuit 217, and outputs the decoded data to a subsequent circuit. The above processing is repeated while the OFDM receiver 201 is receiving an OFDM signal.

  Next, the time domain equalization process performed in step S5 of FIG. 19 will be described with reference to the flowchart of FIG.

  In step S11, the main wave position detection circuit 251 detects a predetermined position of the main wave from the pre-echo equalization signal supplied from the variable coefficient FIR filter 241, and a signal representing the detected position is output to the signal extraction circuit 252, the signal The data is output to the extraction circuit 257 and the signal extraction circuit 260.

  In step S12, pre-echo removal processing is performed. The pre-echo removal process will be described later with reference to the flowchart of FIG.

  In step S13, post-echo removal processing is performed. The post-echo removal process will be described later with reference to the flowchart of FIG. After the post-echo removal process is completed, the process returns to step S5 in FIG. 19, and the subsequent processes are performed.

  Next, the pre-echo removal process performed in step S12 of FIG. 20 will be described with reference to the flowchart of FIG.

  In step S 21, the delay circuit 258 of the filter coefficient generation circuit 244 delays the pre-echo equalization signal supplied from the variable coefficient FIR filter 241, and outputs the delayed pre-echo equalization signal to the coefficient update circuit 259. Among the pre-echo equalized signals output from the delay circuit 258, a signal within a predetermined range determined with the start position of the main wave GI as a reference is handled as an error signal in the coefficient update circuit 259.

  In step S22, the signal extraction circuit 260 extracts a signal in a section including the main GI copy source from the OFDM time domain signal supplied from the quadrature demodulation circuit 215, and determines the extracted main GI copy source. The signal in the included interval is output to the coefficient update circuit 259 as a reference signal.

  In step S23, the coefficient update circuit 259 calculates sample correlation values of the error signal and the reference signal, and generates a pre coefficient so as to update the sample correlation value in a direction to cancel. The coefficient update circuit 259 outputs the pre coefficient to the variable coefficient FIR filter 241.

  In step S24, the variable coefficient FIR filter 241 performs filtering on the OFDM time domain signal using the pre coefficient generated by the coefficient update circuit 259, and removes or suppresses the pre-echo component included in the OFDM time domain signal. I do. The variable coefficient FIR filter 241 outputs an OFDM time domain signal with the pre-echo component removed or suppressed. Thereafter, the process returns to step S12 in FIG. 20, and the subsequent processing is performed.

  Next, the post-echo removal process performed in step S13 of FIG. 20 will be described with reference to the flowchart of FIG.

  In step S31, the signal extraction circuit 252 of the filter coefficient generation circuit 244 extracts a signal in a section including the main wave GI from the pre-echo equalized signal according to the position detected by the main wave position detection circuit 251. The signal extraction circuit 252 outputs the extracted signal of the section including the main wave GI to the variable coefficient FIR filter 253.

  In step S32, the variable coefficient FIR filter 253 performs filtering on the signal supplied from the signal extraction circuit 252 to generate a multipath dummy signal. The variable coefficient FIR filter 253 outputs the multipath dummy signal to the subtraction circuit 254.

  In step S33, the subtracting circuit 254 subtracts the multipath dummy signal from the pre-echo equalized signal to generate an error signal. The subtraction circuit 254 outputs the generated error signal to the delay circuit 255.

  In step S <b> 34, the delay circuit 255 delays the error signal and outputs the delayed error signal to the coefficient update circuit 256.

  In step S <b> 35, the signal extraction circuit 257 extracts a signal in a section including the copy source of the main wave GI from the pre-echo equalized signal according to the position detected by the main wave position detection circuit 251. The signal extraction circuit 257 outputs the signal in the section including the copy source of the main wave GI to the coefficient update circuit 256 as a reference signal.

  In step S36, the coefficient update circuit 256 generates a post coefficient based on the error signal and the reference signal, and outputs the generated post coefficient to the variable coefficient FIR filter 242 and the variable coefficient FIR filter 253.

  In step S <b> 37, the variable coefficient FIR filter 242 filters the OFDM time domain signal supplied from the subtraction circuit 243 using the post coefficient, and outputs the signal obtained by performing the filtering to the subtraction circuit 243. Output.

  In step S38, the subtraction circuit 243 removes the post-echo component by subtracting the signal supplied from the variable coefficient FIR filter 242 from the pre-echo equalized signal. The subtraction circuit 243 outputs an OFDM time domain signal from which the post-echo component has been removed. Thereafter, the process returns to step S13 in FIG. 20, and the subsequent processing is performed.

  With the above processing, it is possible to easily generate highly accurate adaptive equalization filter coefficients without estimating a delay profile and without using a known signal that is required in the LMS algorithm.

  Also, by applying this coefficient to the adaptive equalization filter, it is possible to stably and accurately achieve multipath component removal for time-domain OFDM signals, and to improve reception performance. Is possible.

  Note that the configuration is not limited to the above-described configuration, and the configuration of the circuit can be changed as appropriate. Here, some examples of configuration variations will be described.

  Variations of detection of the main wave GI start position will be described.

  In the example described above, the start position of the main wave GI is detected using the pre-echo equalized signal. However, the signal used to detect the position is not limited to the pre-echo equalized signal.

  That is, as long as the time adjustment considering the delay time in each circuit is performed, the main wave position detection circuit 251 having the configuration of FIG. 17 detects the start position of the main wave GI from any OFDM time domain signal. It is possible.

  It is also possible to detect the position using the OFDM frequency domain signal instead of the OFDM time domain signal.

  In this case, for example, as described with reference to FIG. 6 and FIG. 7, the channel characteristic of the frequency band is estimated by removing the modulation component of the SP signal extracted from the OFDM frequency band signal. In addition, the IFFT operation is performed on the estimated transmission line characteristics to calculate the impulse response of the transmission line in the time domain, and the position where the impulse response is maximized is searched to determine the start position of the main wave GI. Detection is performed.

  FIG. 23 shows the configuration of an OFDM receiver 202 that detects the start position of the GI of the main wave using the OFDM frequency band signal and generates the pre-coefficient and the post-coefficient. In the example of FIG. 23, the OFDM frequency band signal obtained by performing the FFT operation by the FFT circuit 216 is input to the filter coefficient generation circuit 244 of the adaptive equalization filter 231.

  A variation of signal extraction in a section including a GI copy source used as a reference signal for updating a pre coefficient and a post coefficient will be described.

  In the above, when updating the pre-coefficient, the signal in the section including the GI copy source is extracted from the OFDM time domain signal before the pre-echo removal is performed, and pre-echo equalization is performed when the post coefficient is updated. It is assumed that the signal of the section including the GI copy source is extracted from the post signal, but since the purpose is to calculate the sample correlation value with the GI component included in the error signal, the GI copy source is included. The signal of the section including the GI copy source may be extracted from the signal at any time.

  Variations in generating a multipath dummy signal used for updating the post coefficient will be described.

  In the above, a signal obtained by adding 0 to the signal of the section including the main wave GI extracted by the signal extraction circuit 252 is used as the input signal of the variable coefficient FIR filter 253. However, it is not always necessary to add 0. It is also possible to use the pre-echo equalized signal as it is. At this time, since the noise term included in the sample correlation value obtained by the coefficient update circuit 256 is increased, care for noise such as reducing the step size μ multiplied by the sample correlation value is required.

  The generation of coefficients as described above can also be applied to an adaptive equalization filter having another configuration, for example, in which a coefficient update unit is incorporated in a loop of an IIR type equalization filter.

  FIG. 24 is a diagram illustrating another configuration example of the adaptive equalization filter 231.

  The configuration of the filter coefficient generation circuit 244 provided in the adaptive equalization filter 231 in FIG. 24 differs from the configuration in FIG. 10 in the position of the delay circuit. That is, in the example of FIG. 10, the delay circuit 255 and the delay circuit 258 are provided at the position where the error signal is delayed, whereas in the example of FIG. 24, the delay circuit 307 is positioned at the position where the reference signal is delayed. A delay circuit 309 is provided.

  In the example of FIG. 10, the error signal is generated using the signal in the section including the main wave GI, and the sample correlation is obtained using the signal in the section including the copy source of the main wave GI as a reference signal. Although the generation is performed, in the example of FIG. 24, the same processing is realized by switching the role of the signal used for obtaining the sample correlation value.

  The main wave position detection circuit 301 detects a predetermined position of the main wave from the pre-echo equalized signal supplied from the variable coefficient FIR filter 241, and a signal representing the detected position is extracted as a signal extraction circuit 302, a signal extraction circuit 306, And output to the signal extraction circuit 308.

  The signal extraction circuit 302 extracts a signal in the section including the copy source of the main wave GI from the pre-echo equalized signal supplied from the variable coefficient FIR filter 241 according to the position detected by the main wave position detection circuit 301. . The signal extraction circuit 302 outputs the signal of the section including the extracted main wave GI copy source to the variable coefficient FIR filter 303.

  The variable coefficient FIR filter 303 filters the signal supplied from the signal extraction circuit 302 using the post coefficient set by the coefficient update circuit 305. The variable coefficient FIR filter 303 outputs the multipath dummy signal obtained by performing the filtering to the subtraction circuit 304.

  The subtracting circuit 304 subtracts the multipath dummy signal from the pre-echo equalized signal to generate an error signal. The subtraction circuit 254 outputs the generated error signal to the coefficient update circuit 305.

  The coefficient update circuit 305 generates a post coefficient based on the error signal supplied from the subtraction circuit 304 and the delayed reference signal supplied from the delay circuit 307, and the generated post coefficient is converted into a variable coefficient FIR filter. 242 and the variable coefficient FIR filter 303. The coefficient update circuit 305 has a configuration similar to that of the coefficient update circuit 256 shown in FIG.

  The signal extraction circuit 306 extracts a signal in the section including the main wave GI from the pre-echo equalized signal according to the position detected by the main wave position detection circuit 301. The signal extraction circuit 306 outputs the extracted signal in the section including the main wave GI to the delay circuit 307 as a reference signal.

  The delay circuit 307 uses the reference signal supplied from the signal extraction circuit 306 for a time corresponding to an interval one sample less than the effective symbol length with reference to the position one sample after the start position of the main wave OFDM symbol. The delayed reference signal is output to the coefficient update circuit 305.

  The signal extraction circuit 308 extracts a signal in the section including the GI of the main wave from the OFDM time domain signal supplied from the quadrature demodulation circuit 215 according to the position detected by the main wave position detection circuit 301. The signal extraction circuit 308 outputs the extracted signal in the section including the main wave GI to the delay circuit 309 as a reference signal.

  The delay circuit 309 delays the reference signal supplied from the signal extraction circuit 308 and outputs the delayed reference signal to the coefficient update circuit 310.

  The coefficient update circuit 310 regards a signal in a predetermined section including the GI copy source among the pre-echo equalized signals supplied from the variable coefficient FIR filter 241 as an error signal, and supplies the error signal and the delay circuit 309. Based on the reference signal thus generated, a pre coefficient is generated. The coefficient update circuit 310 outputs the generated pre coefficient to the variable coefficient FIR filter 241.

  The adaptive equalization filter 231 having the above configuration can also be realized in a form having the above-described variations.

  FIG. 25 is a diagram illustrating still another configuration example of the adaptive equalization filter 231.

  Normally, the number of taps becomes very large when an OFDM signal is demodulated using an adaptive equalization filter. The number of taps becomes even larger when trying to cope with multipath having a delay time exceeding GI.

  In the variable coefficient FIR filter of the adaptive equalization filter 231 shown in FIG. 10 and the like, since the coefficient of taps without multipath is 0, in some cases, most of the outputs of a large number of taps are 0. May be. In the adaptive equalization filter 231 of FIG. 25, such a waste of taps can be eliminated.

  The configuration shown in FIG. 25 is a configuration assuming that one post-echo wave is removed by one variable coefficient FIR filter and one pre-echo wave is removed or suppressed by two variable coefficient FIR filters. The number of necessary circuits increases according to the number of multipaths assumed.

  The adaptive equalization filter 231 in FIG. 25 has three configurations: a configuration for removing pre-echo, a configuration for removing post-echo, and a configuration for generating coefficients.

  The configuration for removing the pre-echo includes variable length delay circuits 321 and 322, variable coefficient FIR filters 323 and 324, and addition circuits 325 and 326.

  The configuration for removing post-echo includes a subtraction circuit 327, a variable length delay circuit 328, and a variable coefficient FIR filter 329.

  The configuration for generating the coefficients includes a delay profile estimation circuit 330, a pre-echo position detection circuit 331, a post-echo position detection circuit 332, a main wave position detection circuit 333, signal extraction circuits 334, 335, and 336, and variable length delay circuits 337 and 338. , Coefficient update circuits 339 and 340, signal extraction circuits 341, 342, and 343, variable coefficient FIR filter 344, variable length delay circuit 345, subtraction circuit 346, variable length delay circuit 347, and coefficient update circuit 348.

  In the configuration for generating the coefficients, the delay profile estimation circuit 330 estimates the delay profile using the OFDM time domain signal supplied from the orthogonal demodulation circuit 215. As a profile estimation method using an OFDM time domain signal, for example, a matched filter using GI is known. The delay profile estimation circuit 330 outputs the estimated delay profile to the pre-echo position detection circuit 331 and the post-echo position detection circuit 332.

  The pre-echo position detection circuit 331 detects a pre-echo from the delay profile estimated by the delay profile estimation circuit 330, and calculates a predetermined position serving as a pre-echo reference such as a GI start position and a pre-echo delay time.

  The pre-echo position detection circuit 331 outputs the pre-echo position to the signal extraction circuit 335 and the signal extraction circuit 336. The pre-echo position detection circuit 331 outputs the pre-echo delay time to the variable-length delay circuit 337 and the variable-length delay circuit 338, and also outputs the pre-echo delay time to the variable-length delay circuit 321 and the variable-length delay circuit 322.

  The post-echo position detection circuit 332 detects a post-echo from the delay profile estimated by the delay profile estimation circuit 330, and calculates a predetermined position serving as a post-echo reference such as a GI start position and a post-echo delay time. To do.

  The post-echo position detection circuit 332 outputs the post-echo position to the signal extraction circuit 342. The post-echo position detection circuit 332 outputs the post-echo delay time to the variable-length delay circuit 345, the variable-length delay circuit 347, and the variable-length delay circuit 328.

  As described above, in the adaptive equalization filter 231 in FIG. 25, the delay profile is estimated, and the position and delay time of the pre-echo and the post-echo are calculated.

  The main wave position detection circuit 333 detects a predetermined position of the main wave from the pre-echo equalized signal supplied from the addition circuit 326. The main wave position detection circuit 333 has, for example, a configuration similar to the configuration of FIG. The main wave position detection circuit 333 outputs a signal representing the detected position to the signal extraction circuit 334, the signal extraction circuit 341, and the signal extraction circuit 343.

  With reference to FIG. 26 as appropriate, a configuration relating to generation of a post coefficient will be described.

  The signal extraction circuit 341 extracts a signal in a section including the copy source of the main wave GI from the pre-echo equalized signal supplied from the adder circuit 326 according to the position detected by the main wave position detection circuit 333. The signal extraction circuit 341 outputs the extracted signal in the section including the main wave GI copy source to the coefficient update circuit 348 as a reference signal.

A signal S ₃₁ shown at the bottom of FIG. 26 is a reference signal output from the signal extraction circuit 341.

  The signal extraction circuit 342 extracts a signal in a section including the post echo GI from the pre-echo equalized signal supplied from the addition circuit 326 according to the post echo position detected by the post echo position detection circuit 332. The signal extraction circuit 342 outputs the signal in the section including the extracted post-echo GI to the subtraction circuit 346.

A signal S ₃₂ shown in the fifth row from the top in FIG. 26 is a signal in a section including the post-echo GI output from the signal extraction circuit 342.

  The signal extraction circuit 343 extracts the signal in the section including the main wave GI from the pre-echo equalized signal supplied from the adder circuit 326, and the variable coefficient FIR filter 344 extracts the signal in the section including the main wave GI. Output to.

The signal S ₃₃ shown in the second stage from the top in FIG. 26 is a signal in a section including the main wave GI output from the signal extraction circuit 343.

  The variable coefficient FIR filter 344 filters the signal supplied from the signal extraction circuit 343 using the post coefficient set by the coefficient update circuit 348.

  The variable coefficient FIR filter 253 in FIG. 10 has a number of taps corresponding to the assumed maximum delay time, whereas the variable coefficient FIR filter 344 has a sufficiently smaller number of taps. Not done. For example, the variable coefficient FIR filter 253 in FIG. 10 has several thousand taps, whereas the number of taps in the variable coefficient FIR filter 344 is several to several tens.

  The variable coefficient FIR filter 344 outputs a multipath dummy signal obtained by filtering to the variable length delay circuit 345.

A signal S ₃₄ shown in the third stage from the top in FIG. 26 is a multipath dummy signal output from the variable coefficient FIR filter 344. In the example of FIG. 26, the number of taps of the variable coefficient FIR filter 344 is 1.

  The variable-length delay circuit 345 detects the post-echo position of the multipath dummy signal supplied from the variable coefficient FIR filter 344 until the signal in the section including the post-echo GI is output from the signal extraction circuit 342 to the subtraction circuit 346. The delay is performed according to the delay time calculated by the circuit 332. The variable length delay circuit 345 outputs the delayed multipath dummy signal to the subtraction circuit 346.

The signal S ₃₅ shown in the fourth stage from the top in FIG. 26 is a delayed multipath dummy signal output from the variable length delay circuit 345.

  The subtraction circuit 346 subtracts the multipath dummy signal supplied from the variable length delay circuit 345 from the signal supplied from the signal extraction circuit 342 to generate an error signal. The subtraction circuit 346 outputs the generated error signal to the variable length delay circuit 347.

A signal S ₃₆ shown in the sixth stage from the top in FIG. 26 is an error signal output from the subtraction circuit 346.

  The variable length delay circuit 347 delays the error signal supplied from the subtraction circuit 346 according to the delay time calculated by the post-echo position detection circuit 332 so that the start point of the reference signal and the start point of the error signal are the same. The variable length delay circuit 347 outputs the delayed error signal to the coefficient update circuit 348.

A signal S ₃₇ shown in the seventh stage from the top in FIG. 26 is an error signal output from the variable length delay circuit 347.

  The coefficient update circuit 348 has the same configuration as that of the coefficient update circuit 256 shown in FIG. 12 except that the selector 271 is not provided. That is, in the configuration of FIG. 25, the delay time in the variable coefficient FIR filter 344 is sufficiently shorter than the GI length, so the selector 271 of FIG. 12 is not necessary.

  The coefficient update circuit 348 generates a post coefficient using the reference signal supplied from the signal extraction circuit 341 and the error signal supplied from the variable length delay circuit 347, and uses the generated post coefficient as a variable coefficient FIR filter 329. And output to the variable coefficient FIR filter 344.

  A configuration relating to post-echo removal will be described.

  The subtraction circuit 327 subtracts the post-echo removal signal supplied from the variable coefficient FIR filter 329 from the pre-echo equalization signal supplied from the addition circuit 326 to thereby add a post-echo signal included in the post-echo equalization signal. Remove ingredients. The subtraction circuit 327 outputs a signal obtained by removing the post-echo component to the outside of the adaptive equalization filter 231 and also outputs it to the variable length delay circuit 328.

  The variable length delay circuit 328 delays the signal supplied from the subtraction circuit 327 by the delay time calculated by the post-echo position detection circuit 332, and outputs the delayed signal to the variable coefficient FIR filter 329.

  The variable coefficient FIR filter 329 performs filtering on the signal supplied from the variable length delay circuit 328 using the post coefficient generated by the coefficient update circuit 348. The variable coefficient FIR filter 329 outputs a post-echo removal signal obtained by filtering to the subtraction circuit 327.

  In this way, by estimating the delay profile, thereby detecting the post-echo position, and focusing on only the detected post-echo position, the number of taps of the variable coefficient FIR filter can be reduced, At the same time, the scale of the coefficient update circuit can be reduced.

  Next, a configuration related to pre-echo will be described with reference to FIG.

  Similarly to the signal extraction circuit 260 of FIG. 10, the signal extraction circuit 334 copies the main wave GI from the OFDM time domain signal before removing the pre-echo component according to the position detected by the main wave position detection circuit 333. The signal of the section including is extracted. The signal extraction circuit 334 outputs a signal in a section including the copy source of the main wave GI to the coefficient update circuit 339 and the coefficient update circuit 340 as a reference signal.

A signal S ₄₁ shown at the bottom of FIG. 27 is a reference signal output from the signal extraction circuit 334.

  The signal extraction circuit 335 extracts a signal in a section including the pre-echo GI from the pre-echo equalized signal supplied from the addition circuit 326 according to the position detected by the pre-echo position detection circuit 331. The signal extraction circuit 335 outputs the extracted signal including the pre-echo GI to the variable length delay circuit 337 as an error signal.

A signal S ₄₂ shown in the third row from the top in FIG. 27 is an error signal output from the signal extraction circuit 335.

  Based on the position detected by the pre-echo position detection circuit 331, the signal extraction circuit 336 uses a pre-echo component GI having a delay time that is twice the actual delay time from the pre-echo equalized signal supplied from the addition circuit 326. The signal of the section including is extracted. The signal extraction circuit 336 outputs the extracted signal in the section including the pre-echo GI to the variable length delay circuit 338 as an error signal.

A signal S ₄₃ shown in the fourth stage from the top in FIG. 27 is an error signal output from the signal extraction circuit 336.

  The variable length delay circuit 337 generates the error signal supplied from the signal extraction circuit 335 according to the delay time calculated by the pre-echo position detection circuit 331, and the start point of the error signal is the same as the start point of the reference signal output from the signal extraction circuit 334. Delay to be. The variable length delay circuit 337 outputs the delayed error signal to the coefficient update circuit 339.

A signal S ₄₄ shown in the fifth stage from the top in FIG. 27 is an error signal output from the variable length delay circuit 337. The error signal S ₄₄ is input to the coefficient updating circuit 339 with a reference signal S _41.

  The variable length delay circuit 338 generates the error signal supplied from the signal extraction circuit 336 according to the delay time calculated by the pre-echo position detection circuit 331, and the start point of the error signal is the same as the start point of the reference signal output from the signal extraction circuit 334. Delay to be. The variable length delay circuit 338 outputs the delayed error signal to the coefficient update circuit 340.

A signal S ₄₅ shown in the sixth stage from the top in FIG. 27 is an error signal output from the variable length delay circuit 338. This error signal S ₄₅ is input to the coefficient update circuit 340 together with the reference signal S ₄₁ .

  The coefficient update circuit 339 has a configuration similar to that of the coefficient update circuit 256 shown in FIG. The coefficient update circuit 339 generates a pre coefficient based on the reference signal supplied from the signal extraction circuit 334 and the error signal supplied from the variable length delay circuit 337. The coefficient update circuit 339 outputs the generated pre coefficient to the variable coefficient FIR filter 323.

  The coefficient update circuit 340 has the same configuration as that of the coefficient update circuit 256 shown in FIG. The coefficient update circuit 340 generates a pre coefficient based on the reference signal supplied from the signal extraction circuit 334 and the error signal supplied from the variable length delay circuit 338. The coefficient update circuit 340 outputs the generated pre coefficient to the variable coefficient FIR filter 324.

  Thus, when removing the pre-echo with the configuration of FIG. 10, the number of taps corresponding to a constant multiple of the assumed maximum delay time is required, whereas with the configuration of FIG. In the case of removal, the number of taps can be dramatically reduced as in the case of post-echo.

  A configuration related to removal or suppression of pre-echo will be described.

  The variable length delay circuit 321 delays the OFDM time domain signal supplied from the orthogonal demodulation circuit 215 by the delay time calculated by the pre-echo position detection circuit 331, and the delayed OFDM time domain signal is combined with the variable length delay circuit 322. Output to the variable coefficient FIR filter 324.

  The variable length delay circuit 322 further delays the OFDM time domain signal supplied from the variable length delay circuit 321 by the delay time calculated by the pre-echo position detection circuit 331, and sends the delayed OFDM time domain signal to the adder circuit 326. Output.

  The variable coefficient FIR filter 323 performs filtering on the OFDM time domain signal supplied from the orthogonal demodulation circuit 215 using the pre coefficient generated by the coefficient update circuit 339. The variable length delay circuit 322 outputs a signal obtained by performing filtering to the addition circuit 325.

  The variable coefficient FIR filter 324 filters the OFDM time domain signal supplied from the variable length delay circuit 321 using the pre coefficient generated by the coefficient update circuit 340. The variable coefficient FIR filter 324 outputs a signal obtained by performing filtering to the adder circuit 325.

  The adder circuit 325 removes or suppresses part of the pre-echo by adding the signal supplied from the variable coefficient FIR filter 323 and the signal supplied from the variable coefficient FIR filter 324, and sends the obtained signal to the adder circuit 326. Output.

  The adder circuit 326 removes or suppresses the remaining part of the pre-echo by adding the signal supplied from the variable length delay circuit 322 and the signal supplied from the adder circuit 325, and after pre-echo equalizing the obtained signal Output as a signal.

  As described above, the delay profile is estimated, only the position where the multipath component exists and the delay time are detected from the estimated value of the delay profile, and the variable coefficient FIR filter is configured in a form focused on the target position. Therefore, it is possible to reduce the calculation amount and the circuit scale.

  As described above, the configuration of the adaptive equalization filter 231 shown in FIG. 25 is a configuration assuming that the interference of one pre-echo wave and one post-echo wave is removed. Depending on the environment, either pre-echo or post-echo may have two or more waves, or both pre-echo and post-echo may have two or more waves. In this case, it is necessary to limit the multipath to be subjected to interference removal. .

  That is, it is necessary to select which multipath interference is to be removed. In order to efficiently perform interference removal, it is preferable to select a multipath with large interference and remove the component.

  FIG. 28 is a diagram illustrating a configuration example of the adaptive equalization filter 231.

  28, the same reference numerals are given to the same components as those in FIG. The configuration of the adaptive equalization filter 231 shown in FIG. 28 is different from the configuration of FIG. 25 in that a path selection circuit 351 is provided on the output side of the pre-echo position detection circuit 331 and the post-echo position detection circuit 332. The overlapping description will be omitted as appropriate.

  In the adaptive equalization filter 231 shown in FIG. 28, when there are more multipaths than the number of assumed multipaths (one pre-echo wave and one post-echo wave), the multipath to be subjected to interference removal is selected and selected. Coefficients are generated to remove multipath interference.

  A configuration for generating coefficients will be described.

  The delay profile estimation circuit 330 estimates a delay profile using the OFDM time domain signal supplied from the orthogonal demodulation circuit 215 and outputs the estimated delay profile to the pre-echo position detection circuit 331 and the post-echo position detection circuit 332.

  The pre-echo position detection circuit 331 detects a pre-echo based on the delay profile estimated by the delay profile estimation circuit 330, and in addition to a predetermined position serving as a pre-echo reference such as a GI start position and a pre-echo delay time, Calculate pre-echo power. The pre-echo position detection circuit 331 outputs the pre-echo position, delay time, and power to the path selection circuit 351.

  The post-echo position detection circuit 332 detects a post-echo based on the delay profile estimated by the delay profile estimation circuit 330, a predetermined position serving as a post-echo reference, such as a GI start position, and a post-echo delay time. In addition, the post-echo power is calculated. The post-echo position detection circuit 332 outputs the position, delay time, and power to the path selection circuit 351.

  The path selection circuit 351 selects a pre-echo to be subjected to interference removal based on the pre-echo position, delay time, and power supplied from the pre-echo position detection circuit 331. The path selection circuit 351 outputs the pre-echo position selected as the interference removal target to the signal extraction circuits 335 and 336 and outputs the delay time to the variable length delay circuits 321, 322, 337 and 338.

  Further, the path selection circuit 351 selects a post-echo to be subject to interference removal based on the post-echo position, delay time, and power supplied from the post-echo position detection circuit 332. The path selection circuit 351 outputs the position of the post-echo selected as the interference removal target to the signal extraction circuit 342 and outputs the delay time to the variable length delay circuits 328, 345 and 347. Multipath selection by the path selection circuit 351 will be described later.

  Other configurations for generating coefficients shown in FIG. 28 are basically the same as those described above.

  That is, the main wave position detection circuit 333 detects the position of the main wave from the pre-echo equalized signal supplied from the addition circuit 326 and outputs signals representing the detected positions to the signal extraction circuits 334, 341, and 343.

The signal extraction circuit 341 in the configuration that performs processing related to the post-echo extracts a signal in a section including the main wave GI copy source from the pre-echo equalized signal. The signal extraction circuit 341 outputs the signal in the section including the GI copy source of the main wave to the coefficient update circuit 348 as a reference signal (signal S _{31 in} FIG. 26).

The signal extraction circuit 342 extracts a signal in a section including the post-echo GI from the post-echo equalized signal according to the post-echo position supplied from the path selection circuit 351. The signal extraction circuit 342 outputs the signal in the section including the post-echo GI (signal S _{32 in} FIG. 26) to the subtraction circuit 346.

The signal extraction circuit 343 extracts the signal in the section including the main wave GI (signal S _{33 in} FIG. 26) from the pre-echo equalized signal, and outputs the signal in the section including the main wave GI to the variable coefficient FIR filter 344. To do.

Variable coefficient FIR filter 344, by using the post coefficient set by the coefficient updating circuit 348 performs filtering on the signal supplied from the signal extraction circuit 343, multi-path dummy signal (signal S ₃₄ of FIG. 26) Is output to the variable length delay circuit 345.

The variable length delay circuit 345 uses the multipath dummy signal supplied from the variable coefficient FIR filter 344 as the path selection circuit 351 until the signal in the interval including the post-echo GI is output from the signal extraction circuit 342 to the subtraction circuit 346. Delay according to the delay time supplied from The variable length delay circuit 345 outputs the delayed multipath dummy signal (signal S _{35 in} FIG. 26) to the subtraction circuit 346.

The subtraction circuit 346 generates an error signal (signal S _{36 in} FIG. 26) by subtracting the multipath dummy signal supplied from the variable length delay circuit 345 from the signal supplied from the signal extraction circuit 342, and generates a variable length. Output to the delay circuit 347.

The variable length delay circuit 347 delays the error signal supplied from the subtraction circuit 346 so that the start point of the reference signal and the start point of the error signal are simultaneously according to the delay time supplied from the path selection circuit 351. The variable length delay circuit 347 outputs the delayed error signal (signal S _{37 in} FIG. 26) to the coefficient update circuit 348.

  The coefficient update circuit 348 generates a post coefficient using the reference signal supplied from the signal extraction circuit 341 and the error signal supplied from the variable length delay circuit 347, and uses the generated post coefficient as a variable coefficient FIR filter 329. , 344.

The signal extraction circuit 334 in the configuration for performing the processing related to the pre-echo is based on the main wave position detected by the main wave position detection circuit 333, and the GI copy source of the main wave from the OFDM time domain signal before the pre-echo component is removed. The signal of the section including is extracted. The signal extraction circuit 334 outputs the signal in the section including the copy source of the main wave GI to the coefficient update circuits 339 and 340 as a reference signal (signal S _{41 in} FIG. 27).

The signal extraction circuit 335 extracts a signal in a section including the pre-echo GI from the pre-echo equalized signal according to the pre-echo position supplied from the path selection circuit 351. The signal extraction circuit 335 outputs the signal in the section including the pre-echo GI to the variable length delay circuit 337 as an error signal (signal S _{42 in} FIG. 27).

In accordance with the pre-echo position supplied from the path selection circuit 351, the signal extraction circuit 336 extracts a signal in a section including a pre-echo component GI having a delay time that is twice the actual delay time from the pre-echo equalization signal. Extract. The signal extraction circuit 336 outputs the signal in the section including the pre-echo GI to the variable length delay circuit 338 as an error signal (signal S _{43 in} FIG. 27).

The variable length delay circuit 337 determines the error signal supplied from the signal extraction circuit 335 according to the delay time of the pre-echo supplied from the path selection circuit 351, and the start point of the error signal is the start point of the reference signal output from the signal extraction circuit 334. Delay to be simultaneous. The variable length delay circuit 337 outputs the delayed error signal (signal S _{44 in} FIG. 27) to the coefficient update circuit 339.

The variable length delay circuit 338 determines the error signal supplied from the signal extraction circuit 336 according to the delay time of the pre-echo supplied from the path selection circuit 351, and the start point of the error signal is the start point of the reference signal output from the signal extraction circuit 334. Delay to be simultaneous. The variable length delay circuit 338 outputs the delayed error signal (signal S _{45 in} FIG. 27) to the coefficient update circuit 340.

  The coefficient update circuit 339 generates a pre coefficient based on the reference signal supplied from the signal extraction circuit 334 and the error signal supplied from the variable length delay circuit 337 and outputs the pre coefficient to the variable coefficient FIR filter 323.

  The coefficient update circuit 340 generates a pre coefficient based on the reference signal supplied from the signal extraction circuit 334 and the error signal supplied from the variable length delay circuit 338 and outputs the pre coefficient to the variable coefficient FIR filter 324.

  Here, selection of a multipath to be subjected to interference removal will be described.

  As described above, information on the position, delay time, and power of each pre-echo is supplied from the pre-echo position detection circuit 331 to the path selection circuit 351. Further, the post-echo position detection circuit 332 supplies information on the position, delay time, and power of each post-echo to the path selection circuit 351.

  In the path selection circuit 351, for each pre-echo, based on the position, delay time, and power, the amount of intersymbol interference that will remain after the FFT calculation when interference removal is not performed is estimated, and the interference amount is used as a reference. A pre-echo to be subject to interference cancellation is selected.

  For each post-echo, based on the position, delay time, and power, the amount of intersymbol interference that will remain after the FFT calculation when interference cancellation is not performed is estimated. The target post-echo is selected.

The amount of interference of each multipath (each pre-echo and post-echo) with respect to the main wave is expressed by the following equation.
(Interference amount of each multipath) = (symbol interval before and after entering the FFT interval) x (power)

  If the FFT interval is determined, the symbol interval before and after entering the FFT interval is identified from the multipath position and the delay time.

  FIG. 29 is a diagram illustrating a specific example of the interference amount.

  The horizontal direction in FIG. 29 represents the time direction. In order from the upper band, pre-echo, main wave, and post-echo are represented, and the width of each band represents the power of each path. The main wave power is 1, the pre-echo power is α, and the post-echo power is β.

  The same symbol is attached to the same symbol transmitted in each path. When an FFT section is set in a section as shown in FIG. 29 with a dot-attached symbol as a decoding target, a part of the next symbol transmitted by the pre-echo is included in this FFT section. Enter only the amount corresponding to time t1. In addition, a part of the previous symbol transmitted by the post-echo enters only as much as time t2.

  The amount of pre-echo interference is represented by t1 × α, and the amount of post-echo interference is represented by t2 × β.

  In the case of the configuration shown in FIG. 28 in which the interference of one pre-echo wave and one post-echo wave is removed, the path selection circuit 351 uses the interference amount calculated in this way as a reference to pre-echo and post-echo to be subjected to interference removal. Are selected one by one.

  FIG. 30 is a block diagram illustrating a configuration example of the path selection circuit 351.

  As shown in FIG. 30, the path selection circuit 351 includes an FFT interval estimation unit 351A, an interference amount calculation unit 351B, and a selection unit 351C. Information on the position, delay time, and power of each pre-echo output from the pre-echo position detection circuit 331 and information on the position, delay time, and power of each post-echo output from the post-echo position detection circuit 332 include an FFT interval estimation unit. 351A, the interference amount calculation unit 351B, and the selection unit 351C are input.

  The FFT interval estimation unit 351A estimates the FFT interval set in the subsequent FFT calculation from the path configuration.

  FIG. 31 is a diagram illustrating an example of an FFT interval estimated when the path configuration includes only a pre-echo and a main wave.

  In the example of FIG. 31, it is assumed that there are two pre-echo waves. The power of the pre-echo 1 is α, and the power of the pre-echo 2 is β.

  In this case, as shown in FIG. 31, the most forward section in time is estimated as the FFT section within a range in which other symbols transmitted on the main wave (symbols other than the symbol to be demodulated) do not interfere. Is done. The start position of GI is the same position as the start position of the FFT section.

  FIG. 32 is a diagram illustrating an example of an FFT interval estimated when the path configuration includes only a post-echo and a main wave.

  In the example of FIG. 32, it is assumed that there are two post-echo waves. The power of the post echo 1 is α, and the power of the post echo 2 is β.

  In this case, as shown in FIG. 32, the section that is the most backward in time is estimated as the FFT section within a range in which other symbols transmitted by the main wave do not interfere. The boundary position with the next symbol is the same as the end position of the FFT interval.

  FIG. 33 is a diagram illustrating an example of an FFT interval estimated when the path configuration includes a pre-echo, a main wave, and a post-echo.

  In the example of FIG. 33, there are two pre-echoes and post-echoes. The power of the pre-echo 1 is α, the power of the pre-echo 2 is β, the power of the post-echo 1 is γ, and the power of the post-echo 2 is δ.

  In this case, since it can only be expected that other symbols transmitted in the main wave are set within a range that does not cause interference, the interval in which the worst-case interference amount is calculated for each of the pre-echo and the post-echo. Is estimated as the FFT interval.

  That is, as shown in FIG. 33, with regard to the pre-echo, an interval that is the rearmost in time is estimated as an FFT interval within a range in which other symbols transmitted on the main wave do not interfere.

  As for the post-echo, the most temporal interval in time is estimated as the FFT interval within a range where other symbols transmitted by the main wave do not interfere.

  The FFT interval estimation unit 351A outputs a signal representing the FFT interval estimated as described above to the interference amount calculation unit 351B.

  The interference amount calculation unit 351B calculates the interference amount of each multipath based on the FFT section estimated by the FFT section estimation unit 351A and information supplied from the pre-echo position detection circuit 331 and the post-echo position detection circuit 332. The interference amount calculation unit 351B outputs a signal representing the calculated interference amount to the selection unit 351C.

  The selection unit 351C selects one pre-echo wave and one post-echo wave having a larger interference amount as multipaths to be interference-removed based on the interference amount calculated by the interference amount calculation unit 351B.

  For example, the path configuration consists only of a pre-echo and a main wave, and as shown in FIG. 31, the forefront section is estimated as the FFT section within a range where other symbols transmitted by the main wave do not interfere with each other. In this case, the interference amount calculation unit 351B calculates the interference amount of the pre-echo 1 as t1 × α. Further, the interference amount of the pre-echo 2 is calculated as t2 × β.

  In the selection unit 351C, the interference amounts t1 × α and t2 × β are compared, and the pre-echo of the pre-echoes 1 and 2 in which the larger interference amount is obtained is selected as the pre-echo to be subjected to interference removal.

  Also, as shown in FIG. 32, the path configuration consists of only the post-echo and the main wave, and as shown in FIG. 32, the rearmost interval is estimated as the FFT interval within a range where other symbols transmitted by the main wave do not interfere. In this case, the interference amount calculation unit 351B calculates the interference amount of the post-echo 1 as t1 × α. Further, the interference amount of the post-echo 2 is calculated as t2 × β.

  In the selection unit 351C, the amount of interference t1 × α and t2 × β are compared, and the post-echo for which the larger amount of interference is obtained among the post-echoes 1 and 2 is selected as the post-echo for interference removal. The

  The path configuration is composed of pre-echo, main wave, and post-echo. As shown in FIG. 33, the most backward section of the pre-echo within the range in which other symbols transmitted on the main wave do not interfere is the FFT section. When estimated, the interference amount of the pre-echo 1 is calculated as t1 × α, and the interference amount of the pre-echo 2 is calculated as t2 × β.

  For the post-echo, when the most forward section is estimated as the FFT section within the range where other symbols transmitted in the main wave do not interfere, the interference amount of the post-echo 1 is calculated as t3 × γ, The interference amount of the post-echo 2 is calculated as t4 × δ.

  In the selection unit 351C, the interference amounts t1 × α and t2 × β are compared, and the pre-echo of the pre-echoes 1 and 2 in which the larger interference amount is obtained is selected as the pre-echo to be subjected to interference removal.

  Also, comparing the amount of interference t3 × γ and t4 × δ, the post-echo for which the larger amount of interference is obtained among the post-echoes 1 and 2 is selected as the post-echo to be subject to interference removal.

  The multipath position and delay time selected as described above are output from the selection unit 351C. The position of the pre-echo output from the selection unit 351C is supplied to the signal extraction circuits 335 and 336, and the delay time is supplied to the variable length delay circuits 321, 322, 337, and 338. Further, the position of the post-echo output from the selection unit 351C is supplied to the signal extraction circuit 342, and the delay time is supplied to the variable length delay circuits 328, 345, and 347.

  As described above, the multipath is selected by appropriately using both the path power and the arrival time difference from the main wave, and is preferentially removed from the selected multipath components. Can be removed.

  If the multipath with the large arrival time difference from the main wave is selected in order, when the power of the multipath with the large arrival time difference is small, it is more efficient to select the multipath with the small arrival time difference but the large power. However, it can be prevented from happening.

  In addition, when selecting multipaths in descending order of power, it is more efficient to select multipaths with low power but large arrival time differences when the arrival time difference from the main wave of the multipath with large power is small. Sometimes this can be prevented.

  In the above description, the case where it is assumed that the interference of one pre-echo wave and one post-echo wave is removed has been described. However, the same applies to the case where it is assumed that the circuit implementation is increased and interference of two or more waves is removed. Thus, it is possible to select a multipath as an interference removal target.

  For example, a circuit implementation is made assuming that the interference of two pre-echo waves is removed. When there are three or more pre-echoes, the two pre-echoes are selected as the pre-echoes for interference removal in order from the largest interference amount. The

  In addition, circuit implementation is made assuming that the interference of two post-echo waves is removed. When there are three or more post-echoes, the post-echoes of two waves are the post-interference-removed posts in descending order of the amount of interference. Selected as an echo.

  Further, in the above description, the path selection circuit 351 estimates the FFT section by itself and calculates the interference amount of each multipath. However, the FFT selection is not estimated by the path selection circuit 351 but is performed in the subsequent stage. You may make it notify from the circuit 216 (FIG. 9).

  In the path selection circuit 351, multipath selection is performed based on the FFT section notified from the FFT circuit 216 and information supplied from the pre-echo position detection circuit 331 and the post-echo position detection circuit 332.

  FIG. 34 is a diagram illustrating another configuration example of the coefficient update circuit. 34, the same reference numerals are given to the same components as those shown in FIG. The overlapping description will be omitted as appropriate.

  A circuit having the same configuration as that shown in FIG. 34 is provided in the adaptive equalization filter 231 of FIG. 10 as coefficient update circuits 256 and 259. The coefficient update circuits 305 and 310 may be provided in the adaptive equalization filter 231 in FIG. 24, and the coefficient update circuits 339, 340, and 348 may be provided in the adaptive equalization filter 231 in FIGS. May be.

  Usually, when the error signal is large, such as when the OFDM signal is pulled in or when the transmission path fluctuates, the step size μ is increased. After the error signal has converged, the step size μ is desired to be reduced for stabilization. There is a request. The VSS-LMS algorithm has been conventionally used as a method for satisfying this requirement. However, adjusting the step size μ in the same manner as this method, as described above, is an error signal input to the coefficient update circuit 256 of FIG. Is difficult because it contains many noise terms.

  Therefore, in the coefficient update circuit 361 in FIG. 34, the step size μ is not adjusted according to the error signal, but according to the sample correlation value. As a result, the step size μ can be adjusted stably.

  As shown in FIG. 34, the coefficient update circuit 361 further includes a selector 371, a variable μ update circuit 372, and a maximum tap determination circuit 373.

  The selector 371 selects the sample correlation that is the basis of the coefficient set for the tap of the tap number determined by the maximum tap determination circuit 373 from the sample correlation values after removal of the noise term output from each integrator of the integration circuit 274. Extract the value. The selector 371 outputs the extracted sample correlation value to the variable μ update circuit 372.

  The variable μ update circuit 372 updates the step size μ by performing calculation according to the above equation (5) or (6). In equation (5) or (6), the sample correlation value extracted by the selector 371 is used instead of the error signal represented by e [k].

  The maximum tap determination circuit 373 determines the tap number that maximizes the tap coefficient based on the output from each integrator of the integration circuit 276 and outputs a signal representing the tap number to the selector 371.

  This makes it possible to appropriately adjust the step size μ in accordance with the sample correlation value that is less varied than the error signal.

  The series of processes described above can be executed by hardware or can be executed by software. When a series of processing is executed by software, a program constituting the software executes various functions by installing a computer incorporated in dedicated hardware or various programs. The program is installed from a program recording medium on a general-purpose personal computer capable of processing.

  FIG. 35 is a block diagram illustrating a hardware configuration example of a computer that executes the above-described series of processing by a program.

  A CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, and a RAM (Random Access Memory) 503 are connected to each other by a bus 504.

  An input / output interface 505 is further connected to the bus 504. The input / output interface 505 includes an input unit 506 made up of a keyboard, mouse, microphone, etc., an output unit 507 made up of a display, a speaker, etc., a storage unit 508 made up of a hard disk or nonvolatile memory, and a communication unit 509 made up of a network interface. A drive 510 for driving a removable medium 511 such as an optical disk or a semiconductor memory is connected.

  In the computer configured as described above, the CPU 501 loads the program stored in the storage unit 508 to the RAM 503 via the input / output interface 505 and the bus 504 and executes the program, for example. Is done.

  The program executed by the CPU 501 is recorded in the removable medium 511 or provided via a wired or wireless transmission medium such as a local area network, the Internet, or digital broadcasting, and is installed in the storage unit 508.

  The program executed by the computer may be a program that is processed in time series in the order described in this specification, or in parallel or at a necessary timing such as when a call is made. It may be a program for processing.

  The embodiments of the present invention are not limited to the above-described embodiments, and various modifications can be made without departing from the scope of the present invention.

It is a figure which shows an OFDM symbol. It is a figure which shows the arrangement pattern in the OFDM symbol of SP signal. It is a figure which shows the structural example of the conventional OFDM receiver. It is a figure which shows the other structural example of the conventional OFDM receiver. It is a figure which shows the further another structural example of the conventional OFDM receiver. It is a figure which shows the structural example of an adaptive equalization filter. It is a figure which shows the structural example of the variable coefficient filter of FIG. It is a figure which shows the example of the circuit structure to which the LMS algorithm is applied. It is a figure which shows the structural example of the OFDM receiver which concerns on one Embodiment of this invention. It is a figure which shows the structural example of the adaptive equalization filter of FIG. It is a figure which shows the example of the signal used for the production | generation of the coefficient for posting. It is a figure which shows the structural example of a coefficient update circuit. It is a figure which shows the example of the signal used for the production | generation of the coefficient of the tap without multipath. It is a figure which shows the example of the signal used for the production | generation of the coefficient of a tap with a multipath. It is a figure which shows the example of the signal after pre-echo equalization. It is a figure which shows the example of the signal used for the production | generation of the pre coefficient. It is a figure which shows the structural example of a main wave position detection circuit. It is a figure which shows the example of the signal used for the detection of a main wave position. 5 is a flowchart for explaining OFDM demodulation processing of the OFDM receiver. It is a flowchart explaining the time-domain equalization process performed in step S5 of FIG. FIG. 21 is a flowchart for describing pre-echo removal processing performed in step S12 of FIG. 21 is a flowchart for describing post-echo removal processing performed in step S13 of FIG. It is a figure which shows the other structural example of an OFDM receiver. It is a figure which shows the other structural example of an adaptive equalization filter. It is a figure which shows the further another structural example of an adaptive equalization filter. It is a figure which shows the example of the signal used for the production | generation of the coefficient for posting. It is a figure which shows the example of the signal used for the production | generation of the coefficient for posting. It is a figure which shows the structural example of an adaptive equalization filter. It is a figure which shows the specific example of interference amount. It is a block diagram which shows the structural example of a path selection circuit. It is a figure which shows the example of a FFT area. It is a figure which shows the other example of an FFT area. It is a figure which shows the further another example of a FFT area. It is a figure which shows the structural example of a coefficient update circuit. It is a figure which shows the hardware structural example of a computer.

Explanation of symbols

  201 OFDM receiver, 231 adaptive equalization filter, 241 variable coefficient FIR filter, 242 variable coefficient FIR filter, 243 subtraction circuit, 244 filter coefficient generation circuit, 251 main wave position detection circuit, 252 signal extraction circuit, 253 variable coefficient FIR filter , 254 subtraction circuit, 255 delay circuit, 256 coefficient update circuit, 257 signal extraction circuit, 258 delay circuit, 259 coefficient update circuit, 260 signal extraction circuit, 351 path selection circuit, 351A FFT interval estimation unit, 351B interference amount calculation unit, 351C selector

Claims (4)

OFDM signal receiving means for receiving an OFDM signal;
A pre-echo component is removed by applying an adaptive filter to a time-domain OFDM signal received by the OFDM signal receiving means, having a predetermined number of taps in which a variable first coefficient is set, and after pre-echo equalization First filter means for generating a signal;
Subtracting means for subtracting a post-echo component signal from the pre-echo equalized signal generated by the first filter means to generate an equalized signal;
A second number of taps having a predetermined number of taps set with a variable second coefficient and generating a signal of the post-echo component by applying an adaptive filter to the equalized signal generated by the subtracting means; Filter means,
Coefficient generating means for generating the first and second coefficients based on the time domain OFDM signal received by the OFDM signal receiving means and the pre-echo equalized signal generated by the first filter means When,
FFT calculation means for performing the FFT calculation on the equalized signal generated by the subtraction means, and generating an OFDM signal in the frequency domain;
With
The coefficient generating means includes
First extraction means for extracting a signal in a section including the GI of the main wave from the pre-echo equalized signal generated by the first filter means;
Third filter means for generating a dummy signal of a post-echo component by applying an adaptive filter to the signal extracted by the first extraction means using the same coefficient as the second coefficient;
Error signal generating means for subtracting the dummy signal generated by the third filter means from the pre-echo equalized signal and generating an error signal representing an error of the second coefficient;
A second extraction means for extracting a signal of a section including a copy source of the main wave GI from the pre-echo equalized signal and outputting it as a reference signal;
Post-echo coefficient updating means for updating the second coefficient based on the correlation value between the error signal generated by the error signal generating means and the reference signal extracted by the second extracting means;
With
Receiver device. The coefficient generating means includes
A third extracting means for extracting a signal in a section including a copy source of the main wave GI from the time-domain OFDM signal and outputting the extracted signal as a reference signal;
The first coefficient is updated based on a correlation value between the reference signal extracted by the third extraction unit and a signal in a section including the main wave GI in the pre-echo equalized signal. The receiving apparatus according to claim 1 , further comprising: a pre-echo coefficient updating unit. Receive the OFDM signal,
A predetermined number of taps in which a variable first coefficient is set, and a pre-echo component is removed by applying an adaptive filter to the received time-domain OFDM signal to generate a pre-echo equalized signal;
Subtract the post-echo component signal from the generated pre-echo equalized signal to generate an equalized signal,
Generating a signal of the post-echo component by applying an adaptive filter to the generated equalized signal having a predetermined number of taps in which a variable second coefficient is set;
Based on the received time-domain OFDM signal and the pre-echo equalized signal, the first and second coefficients are generated,
Perform an FFT operation on the generated equalized signal to generate an OFDM signal in the frequency domain ,
Extract the signal of the section including the main wave GI from the pre-echo equalized signal,
A dummy signal of a post-echo component is generated by applying an adaptive filter to the extracted signal using the same coefficient as the second coefficient,
Subtracting the dummy signal from the pre-echo equalized signal to generate an error signal representing the error of the second coefficient;
The signal of the section including the copy source of the main wave GI is extracted from the pre-echo equalized signal and output as a reference signal.
A receiving method comprising: updating the second coefficient based on a correlation value between the error signal and the reference signal . Receive the OFDM signal,
A predetermined number of taps in which a variable first coefficient is set, and a pre-echo component is removed by applying an adaptive filter to the received time-domain OFDM signal to generate a pre-echo equalized signal;
Subtract the post-echo component signal from the generated pre-echo equalized signal to generate an equalized signal,
Generating a signal of the post-echo component by applying an adaptive filter to the generated equalized signal having a predetermined number of taps in which a variable second coefficient is set;
Based on the received time-domain OFDM signal and the pre-echo equalized signal, the first and second coefficients are generated,
Perform an FFT operation on the generated equalized signal to generate an OFDM signal in the frequency domain ,
Extract the signal of the section including the main wave GI from the pre-echo equalized signal,
A dummy signal of a post-echo component is generated by applying an adaptive filter to the extracted signal using the same coefficient as the second coefficient,
Subtracting the dummy signal from the pre-echo equalized signal to generate an error signal representing the error of the second coefficient;
The signal of the section including the copy source of the main wave GI is extracted from the pre-echo equalized signal and output as a reference signal.
A program that causes a computer to execute processing including a step of updating the second coefficient based on a correlation value between the error signal and the reference signal .

Images