JP2009100357A - Waveform equalizer and its control method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a waveform equalizer for actualizing a reduced circuit scale, compared with conventional one. <P>SOLUTION: A filter part 11 uses a delay device group 111 consisting of (k-1) pieces of delay devices 111<SB>1</SB>-111<SB>k-1</SB>connected in series to one another for delaying input signals on a sample to sample basis in sequence. A multiplier group 112 consisting of total k pieces of multipliers 112<SB>0</SB>-112<SB>k</SB>multiplies the input signals delayed on a m-sample basis (m is a natural number satisfying 0≤m≤k-1) by the delay device group 111, by tap coefficients, respectively. An adder 113 adds the input signals which the multiplier group 112 consisting of the multipliers 112<SB>0</SB>-112<SB>k</SB>multiplies by the tap coefficients, respectively. A coefficient control part 14 calculates values for the tap coefficients by which the delay device group 111 multiplies the input signals delayed on the m-sample basis, in accordance with an error signal output from an equalizing error estimating part 13 with a delay on a N-sample basis and an input signal output from a delay part 12 with a delay on a (m+N)-sample basis. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a waveform equalizer that performs waveform equalization of an input signal, and a method for controlling the waveform equalizer.

  In a signal transmission system, a received signal may be distorted due to a reflected wave generated in a transmission path. For example, in terrestrial television broadcasting, as shown in FIG. 9, a reflected wave S2 detoured by reflection of a radio wave (hereinafter referred to as a main wave S1) transmitted from a radio tower 301 by a mountain 302 or the like is received on the receiving side. By reaching 303, interference between the main wave S1 and the reflected wave S2 occurs. Such interference between the main wave and the reflected wave has been known as a so-called ghost since the days of analog broadcasting, but it also has a big problem in digital broadcasting because it greatly affects the reception characteristics. Yes.

  For example, the frequency spectrum of the power density of the received signal when there is no reflected wave as shown in FIG. 10A and the frequency spectrum of the power density of the received signal when there is a reflected wave as shown in FIG. Is compared with the following. That is, the power density of the received signal, which is constant according to the frequency when there is no reflected wave, decreases in a predetermined frequency band due to distortion of the signal due to the reflected wave when there is a reflected wave.

  Therefore, the broadcast signal receiver 403 uses a waveform equalizer 404 as shown in FIG. 11 to remove the distortion contained in the broadcast signal received from the transmission station 401 via the transmission path 402 and to generate a waveform. By equalizing, it is possible to accurately decode the broadcast signal without being affected by the reflected wave in the demodulation / decoding unit 405 at the subsequent stage.

  As shown in FIG. 12A, the waveform equalizer 404 applies a received signal consisting of a main wave reaching time T1 and a reflected wave reaching time T2 after time T1 to FIG. By performing a convolution operation on the impulse response as shown in B), the signal component of the reflected wave is removed to perform waveform equalization. In particular, in the waveform equalizer, since it is necessary to remove the signal component of the reflected wave that arrives at the time T2 that is later than the time T1 in order to cope with various transmission paths, the impulse response of the filter is lengthened. As a result, the number of taps for generating the impulse response increases.

  In such a waveform equalizer, as an applied waveform equalizer used for a transmission line whose tap coefficient cannot be fixed in advance, for example, as shown in FIG. The equalization error estimation unit 520 estimates the equalization error of the signal output from the filter unit 510, and the coefficient control unit 530 controls the tap coefficient of the filter unit 510 based on the equalization error. There is.

  Such an adaptive waveform equalizer is, for example, the circuit configuration of the waveform equalizer described in Patent Document 1 previously filed by the present applicant, specifically as shown in FIGS. This is realized by a circuit configuration.

That is, since the filter unit 510 performs a convolution operation on k (k is a natural number) input signals continuously arranged at predetermined time intervals, the input signal is delayed by one sample (k−1). Delay units 511 _{1 to} 511 _k−1 connected in series, and m (m is a natural number satisfying 0 ≦ m ≦ k−1) output from the delay unit group 511 is delayed. respectively and k multipliers ₅₁₂ 0 to 512 _k-1 multiplier group 512 consisting of multiplying the tap coefficients _{C m} to the input signals, the multipliers ₅₁₂ 0 to 512 _k-1 of the multipliers 512 And an adder 513 for adding the multiplication results.

In the filter unit 510, each delay unit 511 _{1 to} 511 _k−1 delays the input signal d (n) by one sample, and each multiplier 512 _{0 to} 512 _k−1 outputs m samples before the time T _n . The input signal d (n−m) is multiplied by each tap coefficient C _m , and the adder 513 adds the multiplication results of the multipliers 512 _{0 to} 512 _k−1 to obtain a waveform-equalized signal. Output as z (n).

  The equalization error estimation unit 520 estimates an equalization error for z (n) output from the filter unit 510 and outputs an error signal e (n) indicating the equalization error.

The coefficient control unit 530 updates the tap coefficients C _m by the controllers 530 _{0 to} 530 _k−1 based on the error signal e (n) output from the equalization error estimation unit 520. Here, an example of a coefficient updating formula for updating the tap coefficient is shown in the following formula (1). Expression (1) is an expression for updating the tap coefficient of the equalization filter based on an LMS (Least Mean Square) algorithm.

C _m (n + 1) = C _m (n) −μ × d (n−m) × e (n) (1)
In equation (1), C _m (n) is the coefficient of the m th tap at time T _n , C _m (n−1) is the coefficient of the m th tap at time T _n−1 , e (n) Represents an error signal at time T _n , d (n−m) represents an input signal m samples before time T _n , and μ represents a coefficient value.

  In the above equation (1), it is assumed that no delay occurs until the error signal e (n) output from the equalization error estimation unit 520 is generated with respect to the input signal d (nm). It can be adapted according to conditions.

  However, in the case where an adaptive equalizer is mounted on a circuit, the coefficient update equation represented by Equation (1) cannot often be applied. The reason for this is particularly problematic because the equalization error estimation unit 520 takes time to calculate an error signal.

  Therefore, in the waveform equalizer, the tap coefficient is updated by using a coefficient updating formula represented by the following formula (2) obtained by modifying the formula (1).

C _m (n + 1) = C _m (n) −μ × d (n−M−N) × e (n−N) (2)
In Expression (2), N represents the number of samples due to circuit delay by the equalization error estimation unit 520.

Therefore, in order to implement Expression (2) in a circuit, for example, in the controller 530 _m of the coefficient control unit 530, as shown in FIG. 15, μ × d (n−m−N) × e (n−N) A multiplier 531, a subtracter 532 for subtracting the multiplication result of the multiplier 531 from the current tap coefficient C _m (n) to calculate a tap coefficient C _m (n + 1), and the tap coefficient C _m (n + 1). Is delayed by one sample. The tap coefficient C _m (n + 1) output from the delay unit 533 is fed back to the subtractor 532 in order to calculate the tap coefficient C _m (n + 2).

Further, since the controller 530 _m inputs the input signal d (n−m−N) N samples before the input signal d (n−m) to the multiplier 531, the input signal d (n−m). A delay group 534 in which a total of N delay units that delay one sample are connected in series is required.

In the coefficient controller 530 having a total of k controllers 530 _m having such a configuration, when a delay device that delays the input signal by one sample is mounted, the circuit delay amount by the equalization error estimator is not considered. Compared to the above, a total of k × N delay devices are required.

  Therefore, in the adaptive equalizer, when the bit width of the input signal to be equalized or the number of taps of the filter increases, the number of delay units provided in the coefficient control unit is required correspondingly, and as a result, the circuit This will increase the scale.

Japanese Patent Application No. 2007-203852

  The present invention has been proposed in view of such circumstances, and by reducing the number of delay devices that delay the input signal, the circuit scale of the entire waveform equalizer that adaptively controls the tap coefficient is reduced. An object of the present invention is to provide a waveform equalizer that realizes reduction, and a method for controlling the waveform equalizer.

  As a means for solving the above-described problem, a waveform equalizer according to the present invention is a waveform equalizer that performs waveform equalization of an input signal. K (k is a natural number) continuously arranged at predetermined time intervals. .) A filter that performs a convolution operation on each input signal and outputs a waveform-equalized signal, an error signal that estimates an equalization error for the signal output from the filter, and indicates the equalization error Are output after being delayed by N (N is a natural number) samples with respect to the signal output from the filter, and the tap coefficient of the filter is controlled in accordance with the error signal output from the error estimation means. Coefficient control means and delay means for sequentially delaying the input signal for each sample and supplying it to the coefficient control means using (k + N-1) delay devices connected in series, and the filter is connected in series (K-1) pieces A delay group that sequentially delays the input signal for each sample using a delay unit, and each input signal delayed by m (m is a natural number satisfying 0 ≦ m ≦ k−1) samples by the delay group. And a multiplier group consisting of a total of k multipliers each multiplied by a tap coefficient, and an adder for adding each input signal multiplied by the tap coefficient in each multiplier of the multiplier group. The means outputs the error signal outputted from the error estimation means after delaying the input coefficient delayed by m samples by the delay group by N samples from the error estimation means and the delay means (output from the delay means). It is calculated based on an input signal delayed by m + N) samples.

  In addition, the waveform equalizer control method according to the present invention performs waveform equalization by performing a convolution operation on k (k is a natural number) input signals continuously arranged at a predetermined time interval. An equalization error is estimated for a filter that outputs a signal and a signal output from the filter, and an error signal indicating the equalization error is N samples (N is a natural number) samples with respect to the signal output from the filter. An error estimation unit that outputs after delay, a coefficient control unit that controls a tap coefficient of the filter according to an error signal output from the error estimation unit, and (k + N−1) delay devices connected in series are used. Then, in the control method of the waveform equalizer comprising the delay means for sequentially delaying the input signal for each sample and supplying the input signal to the coefficient control means, the filter comprises (k−1) delay devices connected in series. Input using the delay group Are sequentially delayed for each sample, and for each input signal delayed by m (m is a natural number satisfying 0 ≦ m ≦ k−1) by the delay group, a total of k multipliers are used. The multiplier group is multiplied by the tap coefficient, and each input signal multiplied by the tap coefficient in each multiplier of the multiplier group is added by the adder. The coefficient control means is delayed by m samples by the delay group. Based on the error signal output by delaying N samples from the error estimation means and the input signal delayed by (m + N) samples output from the delay means. It is characterized by calculating.

  The waveform equalizer according to the present invention is a waveform equalizer that performs waveform equalization of an input signal. K (k is a natural number) input signals that are continuously arranged at predetermined time intervals. A filter that performs a convolution operation and outputs a waveform-equalized signal, an estimation error is estimated for the signal output from the filter, and an error signal indicating the equalization error is converted into a signal output from the filter. And an error estimation means for outputting after delaying by N (N is a natural number) samples, and the filter uses (k + N−1) delay devices connected in series, and the input signal is sampled for each sample. Are sequentially delayed, and each input signal delayed by m (m is a natural number satisfying 0 ≦ m ≦ k−1) by the delay group is multiplied by a tap coefficient and output. a multiplier group consisting of k multipliers, and a multiplier group An adder for adding the input signals multiplied by the tap coefficients in each multiplier, and the coefficient control means, from the error estimation means, tap coefficients for multiplying signals delayed by m samples by the delay group. The calculation is based on the error signal output after being delayed by N samples and the input signal delayed by (m + N) samples by the delay group.

  In addition, the waveform equalizer control method according to the present invention performs waveform equalization by performing a convolution operation on k (k is a natural number) input signals continuously arranged at a predetermined time interval. An equalization error is estimated for a filter that outputs a signal and a signal output from the filter, and an error signal indicating the equalization error is N (N is a natural number) samples of the signal output from the filter. In a method for controlling a waveform equalizer including an error estimation unit that outputs after delaying, a filter uses a delay group of (k + N−1) delay units connected in series, and inputs an input signal to 1 A multiplier comprising a total of k multipliers for each input signal delayed by m for each sample and delayed by m (m is a natural number satisfying 0 ≦ m ≦ k−1) samples by a delay group. Multiply by tap coefficient for each group and output The adder adds the input signals multiplied by the tap coefficients in each multiplier of the generator group, and the coefficient control means adds error coefficients to the tap coefficients for multiplying the signals delayed by m samples by the delay group. Is calculated based on the error signal output after being delayed by N samples from the input signal and the input signal delayed by (m + N) samples by the delay group.

  According to the present invention, the coefficient control means multiplies the input signal delayed by m samples by the delay group to the value of the tap coefficient delayed by N samples from the error estimation means and the error signal output. Since the calculation is made based on the input signal delayed by (m + N) samples output from the delay means, the number of delay devices for delaying the input signal can be reduced as compared with the prior art, thereby adaptively. The circuit scale of the entire waveform equalizer that controls the tap coefficient can be reduced.

  According to the present invention, the coefficient control means multiplies the signal delayed by m samples by the delay group by the error signal output by the error estimation means and (m + N) by the delay group. ) Since the calculation is performed based on the input signal delayed by the number of samples, the number of delay devices for delaying the input signal can be reduced as compared with the prior art, and thereby a waveform equalizer that adaptively controls the tap coefficient. The entire circuit scale can be reduced.

  Hereinafter, the best mode for carrying out the present invention will be described in detail with reference to the drawings.

  A waveform equalizer to which the present invention is applied performs waveform equalization on an input signal. For example, a single carrier BS (Broadcasting Satellite) digital broadcast wave RF (Radio Frequency) as shown in FIG. ) It is incorporated in the receiving apparatus 1 that receives a signal.

  That is, the receiving apparatus 1 outputs the transport stream signal TS by performing demodulation processing on the baseband signal, the antenna 2 that receives the RF signal from the transmission path, the RF converter 3 that converts the RF signal into a baseband signal, and the baseband signal. And a decoding processing unit 5 that performs decoding processing on the transport stream signal TS.

  The antenna 2 receives the RF signal of the BS digital broadcast wave from the transmission path, and supplies the received RF signal to the RF conversion unit 3.

  The RF conversion unit 3 converts the RF signal supplied from the antenna 2 to a baseband signal by applying a carrier wave having a center frequency of a desired channel, and supplies the baseband signal to the demodulation processing unit 4.

  The demodulation processing unit 4 performs demodulation processing on the baseband signal supplied from the RF conversion unit 3 and supplies the demodulated transport stream signal TS to the decoding processing unit 5.

  Specifically, the demodulation processing unit 4 includes an A / D conversion unit 41, a frequency correction unit 42, a sampling synchronization unit 43, a roll-off filter 44, a frequency synchronization unit 45, a waveform equalizer 46, a phase synchronization unit 47, a phase correction. A unit 48 and an error correction unit 49 are provided.

  The A / D conversion unit 41 converts the baseband signal, which is an analog signal supplied from the RF conversion unit 3, into a digital baseband signal and supplies the converted signal to the frequency correction unit 42.

  The frequency correction unit 42 corrects the center frequency of the baseband signal supplied from the A / D conversion unit 41 according to the frequency error signal supplied from the frequency synchronization unit 45 described later, and the corrected baseband signal is obtained. The sampling synchronization unit 43 is supplied. The reason why such correction processing is performed is that the center frequency of the baseband signal converted by the RF conversion unit 3 includes an error with respect to the center frequency of the channel.

  The sampling synchronization unit 43 performs processing for establishing symbol point synchronization with respect to the baseband signal supplied from the frequency correction unit 42. Specifically, the sampling synchronization unit 43 samples the baseband signal at a sampling interval that is an integral multiple of the period between symbol points, for example, and synchronizes with each symbol point, and the base synchronized with the symbol point is obtained. The band signal is supplied to the roll-off filter 44.

  The roll-off filter 44 performs, for example, a Nyquist square-root roll-off characteristic filter process on the baseband signal supplied from the sampling synchronization unit 43 in order to perform band limitation to suppress intersymbol interference. The roll-off filter 44 supplies the band-limited baseband signal to the frequency synchronization unit 45 and the waveform equalizer 46, respectively. Here, the band limitation processing is performed by transmitting a broadcast signal that has been subjected to waveform shaping so as to have a Nyquist square root characteristic on the transmission side in standards such as BS digital broadcasting, and on the reception side. This is because the intersymbol interference is suppressed in the entire transmission / reception system by passing through the Nyquist square root characteristic filter.

  Based on the baseband signal supplied from the roll-off filter 44, the frequency synchronization unit 45 determines whether the center frequency of the baseband signal is synchronized with the desired channel center frequency. For example, the frequency synchronization unit 45 calculates a difference between the power density of the baseband signal and a predetermined reference value, and supplies the calculation result to the frequency correction unit 42 as a frequency synchronization signal. This processing is performed because the baseband signal supplied from the roll-off filter 44 is band-limited as described above, so if the center frequency is shifted from the center frequency of the desired channel, This is because the power density is greatly reduced.

  Note that the frequency correction unit 42 performs frequency conversion using the frequency synchronization signal as an index to correct the center frequency of the baseband signal.

  The waveform equalizer 46 receives the baseband signal supplied from the roll-off filter 44, and, as will be described later, is converted into k input signals (k is a natural number) continuously arranged at predetermined time intervals. A convolution operation is performed on the waveform to equalize the waveform, and the waveform-equalized baseband signal is supplied to the phase synchronization unit 47 and the phase correction unit 48, respectively.

  The phase synchronization unit 47 determines whether the baseband signal supplied from the waveform equalizer 46 is synchronized with respect to the phase, and supplies the determination result to the phase correction unit 48 as a phase synchronization signal. . Specifically, the phase synchronization unit 47 calculates a phase error of the baseband signal using a pilot signal indicating fixed amplitude and phase information transmitted on a transmission frame basis on the transmission side, for example, and uses it as a phase synchronization signal. This is supplied to the phase correction unit 48.

  The phase correction unit 48 corrects the phase of the baseband signal supplied from the waveform equalizer 46 based on the phase synchronization signal supplied from the phase synchronization unit 47 and supplies the corrected signal to the error correction unit 49. In addition, the phase correction unit 48 supplies information regarding the corrected phase difference to the waveform equalizer 46.

  The error correction unit 49 generates a transport stream signal TS by performing error correction on the baseband signal supplied from the phase correction unit 48 using an error correction code added to the baseband signal on the transmission side. To the decoding processing unit 5.

  The decoding processing unit 5 performs decoding processing on the transport stream signal TS supplied from the demodulation processing unit 4 and outputs video data, audio data, control data, and the like.

  In the receiving apparatus 1 having the above-described configuration, the following description will be focused on the configuration and operation of the waveform equalizer 46 that receives a baseband signal and performs waveform equalization.

  First, a first embodiment of the waveform equalizer 46 will be described using a waveform equalizer 46 as shown in FIGS.

In other words, the waveform equalizer 46, as shown in FIG. 2, the baseband signal supplied from the roll-off filter 44 at time T _n (hereinafter, the input signal d (n) referred to.) Performing waveform equalization filter An equalization error with respect to the signal z (n) output from the filter unit 11, the delay unit 12 that sequentially delays the input signal for each sample, and outputs each delayed input signal. An equalization error estimation unit 13 for estimation and a coefficient control unit 14 for controlling tap coefficients of the filter unit 11 are provided.

As shown in FIG. 3, the filter unit 11 performs a convolution operation on k (k is a natural number) input signals arranged continuously at a sampling interval and outputs a waveform equalized signal. , A delay group 111 that sequentially delays the input signal d (n) input at time T _n for each sample, and m (m satisfies 0 ≦ m ≦ k−1) output from the delay group 111. Natural number.) Multiplier group 112 for multiplying each input signal d (n−m) at time T _nm delayed by samples by tap coefficients C _{0 to} C _k−1 , and each of multiplier group 112 And an adder 113 for adding the multiplication results.

The delay group 111 uses a total of (k−1) delay devices 111 _{1 to} 111 _k−1 connected in series to sequentially delay the input signal for each sample, and is delayed by m samples. Each input signal d (nm) of T _nm is output to the multiplier group 112.

The multiplier group 112 multiplies the input signals d (n) to d (n−k + 1) output from the delay group 111 by a total of k multipliers that respectively multiply the tap coefficients C _{0 to} C _k−1 . 112 _{0 to} 112 _k−1 .

The adder 113 adds the multiplication results of the multipliers 112 _{0 to} 112 _k−1 of the multiplier group 112 and outputs the result as a waveform equalized signal z (n).

The delay unit 12 sequentially delays the input signal d (n) for each sample using (k + N−1) delay devices 12 _{1 to} 12 _{k + N−1} connected in series, and each input signal d (n ) To d (n−k + 1−N) are output. Here, N represents the number of samples of circuit delay by the equalization error estimation unit 13 described later.

  The equalization error estimation unit 13 estimates an equalization error of z (n) output from the filter unit 11 based on the phase error information supplied from the phase correction unit 48, and an error signal indicating the equalization error. e (n) is output.

  Specifically, the equalization error estimation unit 13 uses the difference between the output signal z (n) of the filter and the transmission point s (n) that is the source of the signal as the error signal e (n), and the following equation (3) ).

e (n) = z (n) × Δφ (n) −s (n) (3)
In equation (3), z (n) represents the output signal of the waveform equalizer 46, and s (n) represents the estimated value of the transmission symbol. Δμ (n) is phase error information supplied from the phase correction unit 48 and represents a correction amount that the phase correction unit 48 performs phase correction on z (n).

  Here, several methods for estimating the transmission point s (n) that is the source of the output signal z (n) of the waveform equalizer 46 are known. Two methods are mentioned.

  As a first method, there is a method of transmitting a fixed sequence for equalization. In this method, the transmission side transmits a fixed sequence for each transmission frame, for example. When the position of the transmission frame and the fixed sequence is specified on the receiving side, the original transmission signal can be specified only in the fixed sequence section.

  As a second method, there is a method in which a signal point closest to the signal point after equalization is assumed to be a transmission point. In this method, for example, in the case of QPSK (Quadrature Phase Shift Keying) modulation, as shown in FIG. 4, one of the four points A, B, C, and D is a transmission point. Here, as shown in FIG. 4, if a point X on the IQ plane is received, it can be assumed that the point A closest to the point X is transmitted and the point X is received.

  Using these methods, an estimated value of the transmission symbol is obtained.

Coefficient control unit 14, an equalization error based on the estimated unit 13 the error signal e output from the (n), the sum of k tap coefficients _C 0 _{-C k-1} total update respectively the k controller 14 ₀ ~ 14 _k-1 .

Here, an example of a coefficient update equation for updating the tap coefficients C _m the following equation (4). Equation (4) is an equation for updating the tap coefficient of the equalization filter based on an LMS (Least Mean Square) algorithm.

C _m (n + 1) = C _m (n) −μ × d (n−m) × e (n) (4)
In Equation (4), C _m (n) is the coefficient of the m th tap at time T _n , C _m (n−1) is the coefficient of the m th tap at time T _n−1 , e (n) Is an error signal at time T _n , d (n−m) is an input signal at time T _n−m , and μ is a coefficient value.

  In the above equation (4), it is assumed that no delay occurs until the error signal e (n) output from the equalization error estimation unit 13 is generated with respect to the input signal d (n−m). It can be adapted according to conditions.

  However, when an adaptive equalizer is mounted on a circuit, the coefficient update equation represented by equation (4) cannot often be applied. This is particularly problematic because the error signal calculation takes time in the equalization error estimation unit.

Therefore, in the controller 14 _m that updates the m-th tap coefficient C _m , the tap coefficient is updated by using a coefficient updating formula represented by the following formula (5) obtained by modifying the formula (4).

C _m (n + 1) = C _m (n) −μ × d (n−m−N) × e (n−N) (5)
In Equation (5), N represents the number of samples of the circuit delay amount by the equalization error estimation unit.

That is, in the controller 14 _m , since the equation (5) is implemented by a circuit, as shown in FIG. 5, a multiplier 141 _m that calculates μ × d (n−N−N) × e (n−N). A subtractor 142 _m that calculates the tap coefficient C _m (n + 1) at the next time T _{n + 1} by subtracting the multiplication result of the multiplier 141 _m from the current tap coefficient C _m (n), and the tap coefficient C _m A delay unit 143 _m that delays (n + 1) by one sample. Incidentally, the tap coefficients output from the delayer _{143 m C m (n + 1} ) is fed back to the subtractor 142 _m to calculate the tap coefficients _C m (n + 2).

Here, in the controller 14 _m , the input signal d (n−m−N) output from the delay unit 12 and the d (n−N) output from the equalization error estimation unit 13 are input to the multiplier 141 _m. By inputting the error signal e (n−N) indicating the equalization error, the coefficient is updated so as to satisfy the condition of Expression (5).

In this way, in the coefficient control unit 14, each controller 14 _m multiplies the tap coefficient C _m by which the signal d (n−m) delayed by m samples at the time T _n by the delay group 111 is multiplied. , and d (n-n) error signal e indicating the equalization error of (n-n) outputted from the equalization error estimation section 13 at time _{T n,} at time _{T n} which is output from the delay unit 12 (m + n ) Calculate based on the signal d (n−m−N) delayed by the number of samples.

  On the other hand, the conventional adaptive waveform equalizer outputs the input signal d (n−m−N) N samples before the input signal d (n−m) in each controller. Therefore, a delay group is provided in which a total of N delay units that delay the input signal by one sample are connected in series.

  That is, in the conventional adaptive waveform equalizer, N × k × B shift registers are required to implement the delay group provided in the coefficient control unit with a circuit. Here, B indicates the bit width of the input signal. That is, one delay unit is composed of B shift registers.

On the other hand, in the waveform equalizer 46, the delay unit 12 is mounted on a circuit with (N + k−1) × B shift registers, and each controller 14 _{m corresponds to} the input signal d (n−m). There is no need to provide a delay device that outputs the input signal d (n−m−N) before N samples. Therefore, the waveform equalizer 46 can reduce the number of shift registers mounted on the circuit by the number represented by the following equation (6), as compared with the conventional adaptive waveform equalizer described above.
N × k × B− (N + k−1) × B (6)
As the values of N, k, and B increase, the waveform equalizer 46 increases the number of shift registers mounted on the circuit as compared to the conventional adaptive waveform equalizer described above. Can be reduced.

  Here, in order to perform waveform equalization with higher accuracy, it is necessary to increase the number of taps k and the bit width B of the input signal. In such a case, the waveform equalizer 46 uses a conventional adaptive waveform or the like. It is possible to reduce the number of shift registers more effectively compared to the generator.

  In the waveform equalizer 46, since the number of shift registers can be reduced, the circuit scale for mounting the waveform equalizer can be reduced, and the power consumption used for the shift register can also be reduced.

  As described above, in the waveform equalizer 46 according to the first embodiment, the value of the tap coefficient that the coefficient control unit 14 multiplies the input signal delayed by m samples output from the delay group 111. Is calculated based on the error signal output after being delayed by N samples from the equalization error estimator 13 and the input signal delayed by (m + N) samples output from the delay unit 12. The number of delay devices to be delayed can be reduced as compared with the prior art, and thereby the circuit scale of the entire waveform equalizer that adaptively controls the tap coefficient can be reduced.

  Next, a second embodiment of the waveform equalizer 46 will be described using a waveform equalizer 20 as shown in FIGS.

  That is, as shown in FIG. 6, the waveform equalizer 20 includes a filter unit 21 that performs waveform equalization of an input signal, and an equalization error estimation that estimates an equalization error with respect to a signal output from the filter unit 21. Unit 22 and a coefficient control unit 23 that controls the tap coefficient of the filter unit 21.

Filter unit 21 for outputting a signal waveform-equalized by a convolution operation with respect to aligned k input signals successively at a predetermined time interval, as shown in FIG. 5, a certain time T _n each input of the input signal d and the delay unit group 211 for sequentially delaying (n) to each sample, the time T time is m samples delayed output from the delay unit group 211 in _n T _n-m inputted to A multiplier group 212 that multiplies each of the signals d (n−m) by tap coefficients C _{0 to} C _k−1 and an adder 213 that adds the multiplication results of the multiplier group 212 are provided.

Delay unit group 211, using the series-connected (k + N-1) pieces of delay devices _{211 1 ~211 k + N-1} , sequentially delaying the input signal for each sample, m sample delayed by the time T Each of the _n−m input signals d (n−m) is output to the multiplier group 212.

The multiplier group 212 multiplies each of the input signals d (n) to d (n−k + 1) output from the delay group 211 by a total of k multipliers that are multiplied by tap coefficients C _{0 to} C _k−1 , respectively. 212 _{0 to} 212 _k−1 .

The adder 213 adds the multiplication results of the multipliers 212 _{0 to} 212 _k−1 of the multiplier group 212 and outputs the result as a waveform equalized signal z (n).

  The equalization error estimation unit 22 estimates an equalization error of z (n) output from the filter unit 21 based on the phase error information supplied from the phase correction unit 48, and an error signal indicating the equalization error. e (n) is output. Since the specific error signal calculation method is the same as the calculation method performed by the equalization error estimation unit 13 according to the first embodiment, the description thereof is omitted.

Coefficient control unit 23, equalization error estimator based on the error signal e (n) output from 22, the sum of k tap coefficients _C 0 _{-C k-1} total update respectively the k controller 23 ₀ ~ 23 _k-1 .

As in the first embodiment, the controller 23 _m that updates the m-th tap coefficient C _m updates the tap coefficient by using the coefficient update formula shown by the above-described formula (5).

That is, in the controller 23 _m , since the equation (5) is implemented by a circuit, as shown in FIG. 8, a multiplier 231 _m that calculates μ × d (n−N−N) × e (n−N). A subtractor 232 _m that calculates the tap coefficient C _m (n + 1) at the next time T _{n + 1} by subtracting the multiplication result of the multiplier 231 _m from the current tap coefficient C _m (n), and this tap coefficient C _m A delay unit 233 _m that delays (n + 1) by one sample. Incidentally, the tap coefficients output from the delay unit _{233 m C m (n + 1} ) is fed back to the subtractor 232 _m to calculate the tap coefficients _C m (n + 2).

Here, in the controller 23 _m , the multiplier 231 _m receives the input signal d (n−m−N) output from the delay group 211 and d (n−N) output from the equalization error estimation unit 13. ) To input the error signal e (n−N) indicating the equalization error, the coefficient is updated so as to satisfy the condition of the equation (5).

In this way, in the coefficient control unit 23, each controller 23 _m multiplies the signal d (n−m) delayed by m samples at the time T _n by the delay group 111 with the tap coefficient C _m . , and d (n-n) error signal e indicating the equalization error of (n-n) outputted from the equalization error estimation section 13 at time _{T n,} at time _{T n} which is output from the delay unit group 211 ( m + N) Based on the signal d (n−m−N) delayed by the number of samples.

The waveform equalizer 20, for supplying a signal d (n-m-N) to each controller 23 _m, compared to the number of delay units of delay unit group 211 to the delay unit group of a conventional filter unit N Although there are many, it is not necessary to provide a delay device that outputs the input signal d (n−m−N) N samples before the input signal d (n−m) in each controller 23 _m . . Therefore, the waveform equalizer 20 can reduce the number of shift registers mounted on the circuit by the number represented by the following equation (7), compared to the conventional adaptive waveform equalizer described above.
N × k × B−N × B (7)

  In this way, the waveform equalizer 20 has a shift mounted on the circuit as compared with the conventional adaptive waveform equalizer described above as the values of N, k, and B increase. The number of registers can be further reduced.

  In particular, in order to perform waveform equalization with higher accuracy, it is necessary to increase the number of taps k and the bit width B of the input signal. In such a case, the waveform equalizer 46 uses a conventional adaptive waveform or the like. It is possible to reduce the number of shift registers more effectively compared to the generator.

  In addition, since the number of shift registers can be reduced in the waveform equalizer 20, in addition to reducing the circuit scale for mounting the waveform equalizer, the power consumption used for the shift register can also be reduced.

  As described above, in the waveform equalizer 20 according to the second embodiment, the coefficient control unit 23 multiplies the signal delayed by m samples output from the delay unit group 211 of the filter unit 21. Since the coefficient is calculated based on the error signal output from the equalization error estimation unit 22 and the input signal delayed by (m + N) samples output from the delay group 211, a delay device that delays the input signal As described above, the number of the above can be reduced as compared with the prior art, and thereby the circuit scale of the entire waveform equalizer that adaptively controls the tap coefficient can be reduced.

  In addition, this invention is not limited only to the above embodiment, Of course, a various change is possible in the range which does not deviate from the summary of this invention.

It is a figure which shows the whole structure of a receiver. It is a figure which shows the whole structure of the waveform equalizer which concerns on a 1st Example. It is a figure which shows the circuit structure of the waveform equalizer which concerns on a 1st Example. It is a figure explaining the method of estimating the transmission point used as the origin of the output signal of a waveform equalizer. It is a figure which shows the structure of the coefficient control part which concerns on a 1st Example. It is a figure which shows the whole structure of the waveform equalizer which concerns on a 2nd Example. It is a figure which shows the circuit structure of the waveform equalizer which concerns on a 2nd Example. It is a figure which shows the structure of the coefficient control part which concerns on a 2nd Example. It is a figure explaining the interference of an electromagnetic wave. It is a figure explaining the change of the frequency spectrum by the presence or absence of a reflected wave. It is a figure explaining a waveform equalizer. It is a figure explaining the relationship between an impulse response and a tap coefficient. It is a figure explaining the whole structure of the conventional waveform equalizer. It is a figure explaining the circuit structure of the conventional waveform equalizer. It is a figure explaining the structure of the coefficient control part of the conventional waveform equalizer.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Receiving device, 2 Antenna, 3 RF conversion part, 4 Demodulation process part, 41 A / D conversion part, 42 Frequency correction part, 43 Sampling synchronization part, 44 Roll-off filter, 45 Frequency synchronization part, 46, 20 Waveform equalization Unit, 47 phase synchronization unit, 48 phase correction unit, 49 error correction unit, 5 decoding processing unit, 11, 21 filter unit, 111, 211 delay group, 111 _{1 to} 111 _k−1 , 12 _{1 to} 12 _{k + N−1} , 143 _m , 211 _{1 to} 211 _k−1 , 233 _m delay unit, 112, 212 multiplier group, 112 _{0 to} 112 _k , 141 _m , 212 _{0 to} 212 _k , 231 _m multiplier, 113, 213 adder, 12 delay units, 13, 22 equalization error estimation units, 14, 23 coefficient control units, 14 _{0 to} 14 _k , 23 ₀ to 23 _k controllers, 142 _m and 232 _m subtractors

Claims (4)

In the waveform equalizer that performs waveform equalization of the input signal,
A filter that performs a convolution operation on k (where k is a natural number) input signals continuously arranged at predetermined time intervals and outputs a waveform-equalized signal;
An equalization error is estimated for the signal output from the filter, and the error signal indicating the equalization error is delayed by N (N is a natural number) samples from the signal output from the filter. Error estimation means to output;
Coefficient control means for controlling the tap coefficient of the filter in accordance with the error signal output from the error estimation means;
Using (k + N−1) delay devices connected in series, delay means for sequentially delaying the input signal for each sample and supplying the delay to the coefficient control means,
The above filter
A delay group that sequentially delays the input signal for each sample using (k−1) delay units connected in series;
Multipliers comprising a total of k multipliers for multiplying each input signal delayed by m (m is a natural number satisfying 0 ≦ m ≦ k−1) samples by the delay group. Group,
An adder for adding each input signal multiplied by a tap coefficient in each multiplier of the multiplier group,
The coefficient control means delays the input coefficient delayed by m samples by the delay group for the value of the tap coefficient by N samples from the error estimation means, and outputs the error signal. A waveform equalizer which calculates based on an input signal output from the delay means and delayed by (m + N) samples. A filter that performs a convolution operation on k (k is a natural number) input signals continuously arranged at a predetermined time interval and outputs a waveform-equalized signal;
An equalization error is estimated for the signal output from the filter, and the error signal indicating the equalization error is delayed by N (N is a natural number) samples from the signal output from the filter. Error estimation means to output;
Coefficient control means for controlling the tap coefficient of the filter in accordance with the error signal output from the error estimation means;
In a method of controlling a waveform equalizer, comprising (k + N−1) delay devices connected in series, delay means for sequentially delaying the input signal for each sample and supplying the input signal to the coefficient control means,
The above filter
Using the delay group consisting of (k−1) delay units connected in series, the input signal is sequentially delayed every sample,
For each input signal delayed by m (m is a natural number satisfying 0 ≦ m ≦ k−1) by the delay group, each of the taps is multiplied by a multiplier group consisting of a total of k multipliers. Multiply by a factor
Each input signal multiplied by the tap coefficient in each multiplier of the above multiplier group is added by an adder,
The coefficient control means delays the input coefficient delayed by m samples by the delay group for the value of the tap coefficient by N samples from the error estimation means, and outputs the error signal. A method for controlling a waveform equalizer, comprising: calculating based on an input signal delayed by (m + N) samples output from a delay means. In the waveform equalizer that performs waveform equalization of the input signal,
A filter that performs a convolution operation on k (where k is a natural number) input signals continuously arranged at predetermined time intervals and outputs a waveform-equalized signal;
An equalization error is estimated for the signal output from the filter, and an error signal indicating the equalization error is delayed by N (N is a natural number) samples from the signal output from the filter. An error estimation means for outputting,
The above filter
A group of delay units that sequentially delay the input signal for each sample using (k + N−1) delay units connected in series;
Each of the input signals delayed by m samples (m is a natural number satisfying 0 ≦ m ≦ k−1) by the delay group is composed of k multipliers that respectively multiply and output the tap coefficients. A multiplier group;
An adder for adding the input signals multiplied by the tap coefficient in each multiplier of the multiplier group,
The coefficient control means delays the tap coefficient for multiplying the signal delayed by m samples by the delay group by the N samples from the error estimation means, and outputs the error signal. The waveform equalizer is calculated based on the input signal delayed by (m + N) samples. A filter that performs a convolution operation on k (k is a natural number) input signals continuously arranged at a predetermined time interval and outputs a waveform-equalized signal;
An equalization error is estimated for the signal output from the filter, and an error signal indicating the equalization error is delayed by N (N is a natural number) samples from the signal output from the filter. In a control method of a waveform equalizer comprising an error estimation means for outputting,
The above filter
Using the delay group consisting of (k + N−1) delay units connected in series, the input signal is sequentially delayed every sample,
For each input signal delayed by m (m is a natural number satisfying 0 ≦ m ≦ k−1) by the delay group, the tap coefficients are respectively set by a multiplier group consisting of a total of k multipliers. Multiply
The input signal multiplied by the tap coefficient in each multiplier of the above multiplier group is added by an adder,
The coefficient control means delays the tap coefficient for multiplying the signal delayed by m samples by the delay group by the N samples from the error estimation means, and outputs the error signal. And (m + N) based on the input signal delayed by the number of samples.

Images