JP2012151763A - Receiver and propagation path compensation method thereof - Google Patents

Abstract

PROBLEM TO BE SOLVED: To correct a simple point deviation of a DFT window with a simple circuit configuration.SOLUTION: A receiver that receives a signal transmitted using a multicarrier includes a DFT unit which applies a discrete Fourier transformation (hereinafter referred to as DFT) calculation to a reception signal in a time domain for converting it to a plurality of subcarrier signals in a frequency domain, and a propagation path estimation circuit which corrects a propagation path characteristic from the data signal of the plurality of subcarrier signals based on the reference vector of a pilot signal contained in the plurality of subcarrier signals. The propagation path estimation circuit includes a minimum angle information storage part which stores the minimum angle information of rotation factor in DFT calculation, a phase rotation amount calculation part which calculates a first phase rotation amount for the plurality of subcarrier signals according to the minimum angle information and the sample point deviation number of a DFT window, and a phase correction part which corrects the phase of the plurality of subcarrier signals based on the first phase rotation amount.

Description

  The present invention relates to a receiving apparatus and its propagation path compensation method.

  In a wireless communication system using OFDM (Orthogonal Frequency Division Multiplexing) or OFDMA (Orthogonal Frequency Division Multiple Access), which is a communication method for transmitting and receiving signals using multicarriers, The modulated component is modulated by a plurality of subcarriers having frequencies whose phases are orthogonal to each other, IFFT is performed on the plurality of subcarriers, up-converted to a carrier frequency, and transmitted. On the other hand, the receiving side downconverts the received signal, performs FFT, and demodulates the modulated components of a plurality of subcarriers. Since the frequencies of the subcarriers are different, the phases of the received subcarriers are different from each other, and the amplitude varies according to the movement of the propagation path and the receiving apparatus.

  Therefore, a pilot signal, which is a known symbol, is inserted into a plurality of subcarriers at a predetermined ratio on the transmitting side, and the phase and amplitude information of a known modulated component of the pilot signal is extracted on the receiving side. Demodulating the modulated component of the subcarrier to be transmitted. The phase and amplitude information (reference vector) of this pilot signal is an estimated value of propagation path distortion (estimated propagation path value).

  In order to estimate the propagation path characteristics with high accuracy, it is desirable to insert pilot signals into all subcarriers. However, since this reduces the data transmission efficiency, some of the multiple symbols on the time axis have some symbols. A pilot signal is inserted. Then, in each subcarrier, by taking a moving average of the reference vectors obtained from a plurality of pilot signals mapped to some symbols of the same subcarrier, the fluctuation of the reference vector due to propagation path fluctuation is smoothed, The accuracy of the reference vector that is the propagation path estimation value is increased. Further, in accordance with this moving averaged reference vector, the phase and amplitude of the subcarrier in the symbol between the pilot signals are corrected to perform propagation path compensation.

  The pilot signal in the OFDM system is described in the following patent documents.

JP 2004-312372 A JP 2006-60433 A JP-A-9-116521

  In the receiving apparatus, DFT (Discrete Fourier Transform) processing for multiplying a plurality of sample points of the time domain received signal before Fourier transform by multiplying a predetermined coefficient to obtain the frequency domain received signal after Fourier transform Is done. The calculation of the DFT process is performed on a plurality of sample points in the DFT window synchronized with the symbol. Here, the sample point means a sample point when an analog signal is converted into a digital signal by the AD converter. However, the start position of the DFT window is synchronized with the original symbol due to the time variation of the received signal due to the movement of the terminal and the time variation of the received signal due to the disturbance such as the arrival time of the received signal due to the difference in the multipath path. The position may deviate. Due to the shift of the DFT window, the DFT calculation is performed on the sample point shifted from the original sample point of the DFT calculation target.

  For this reason, if the above sample point deviation occurs between pilot signals interpolated at different positions on the time axis, the phases of the pilot signals on both sides differ, and the accuracy of their moving average values decreases, resulting in high accuracy. Propagation path estimation value cannot be obtained.

  SUMMARY OF THE INVENTION An object of the present invention is to provide a receiving apparatus and a propagation path compensation method for correcting the phase of a frequency domain signal that has undergone DFT processing due to the occurrence of a DFT window sample point shift.

A first aspect of the receiving device is a receiving device that receives a signal transmitted using a multicarrier.
A DFT unit that converts a time domain received signal into a plurality of frequency domain subcarrier signals by performing a discrete Fourier transform (DFT) operation;
A propagation path estimation circuit that corrects distortion due to a propagation path from a data signal of the plurality of subcarrier signals based on a reference vector of a pilot signal interpolated in the plurality of subcarrier signals;
The propagation path estimation circuit includes a minimum angle information storage unit that stores minimum angle information of a twiddle factor of a DFT operation, and a first angle for the plurality of subcarrier signals corresponding to the minimum angle information according to the number of sample point deviations of the DFT window. And a phase correction unit that corrects phases of the plurality of subcarrier signals based on the first phase rotation amount.

  According to the first aspect, it is possible to correct the phase fluctuation due to the sample point shift of the DFT window with a simple configuration.

It is a block diagram of the radio | wireless apparatus which concerns on this Embodiment. It is a block diagram of the receiver which concerns on this Embodiment. It is a block diagram of a propagation path estimation circuit. It is a figure which shows the relationship between several subcarriers of OFDM, and several symbols. It is a figure which shows the relationship between several subcarriers of OFDM, and several symbols. It is a figure which shows the determinant of DFT calculation. DFT is a diagram for explaining a rotation factor ^{W N} arithmetic. FT after the frequency domain signals _X 0 to X ₇ of the Q component (sin), is a diagram showing a signal waveform of the eight sub-carrier of the I component (cos). FT after the frequency domain signals _X 0 to X ₇ of the Q component (sin), is a diagram showing a signal waveform of the eight sub-carrier of the I component (cos). It is a figure which shows the DFT determinant when a DFT window has shifted | deviated 1 sample point, since the phase of an incoming wave has advanced. It is a figure which shows a DFT determinant when a DFT window has shifted | deviated 1 sample point because the arrival wave was delayed. It is a block diagram of the propagation path estimation circuit in 1st Embodiment. It is a block diagram of a pilot signal correction | amendment part. It is a block diagram of the pilot signal correction | amendment part 40 which SC1 rotation amount calculation part 411 carries out phase rotation by the complex conjugate of the rotation factor W1. 4 is a configuration diagram of an SC1 rotation amount calculation unit 411 and each SC rotation amount calculation unit 414 in a phase rotation amount calculation unit 41. FIG. 3 is a configuration diagram of a phase correction unit 42. FIG. It is a block diagram of the propagation path estimation circuit in 2nd Embodiment. It is a block diagram of the sample point shift correction unit 40A. It is another block diagram of sample point shift amendment part 40A.

  FIG. 1 is a configuration diagram of a radio apparatus according to the present embodiment. The transmission side of the wireless device includes an encoding circuit 10 that encodes transmission data Txd, a mapping circuit 11 that maps the encoded data on the I and Q coordinate axes, an I component of the mapped coordinate points, Q A modulation circuit 12 that modulates components with a plurality of subcarriers and performs fast inverse Fourier transform (IFFT), a digital-analog conversion circuit 13, and an orthogonal modulation circuit 14 that orthogonally modulates time domain signals of I and Q components with local frequency signals And an IF / RF circuit 15 for up-converting to a high frequency of the carrier wave. The carrier signal is transmitted from the antenna AT via the diplexer 16.

  On the other hand, on the reception side of the wireless device, the received signal received by the antenna AT is input to the RF / IF circuit 20 through the diplexer 16 and down-converted to an intermediate frequency. Then, the quadrature detection circuit 21 performs quadrature detection with the local frequency signal to generate a time domain signal of I component and Q component. Thereafter, the analog-to-digital conversion circuit 22 converts it into a digital signal, and the demodulation circuit 23 performs fast Fourier transform (FFT) to convert it into a frequency domain signal. The propagation path characteristics of each subcarrier are estimated from the pilot signal, and the data signal The channel characteristics are removed from the I and Q components and demodulated, and the demapping circuit 24 demaps from the coordinate points on the I and Q coordinates to reproduce the encoded data, and the decoder circuit 25 decodes it. Received data Rxd is extracted (decoded).

  The propagation path estimation circuit according to the present embodiment is provided in the demodulation circuit 23.

  FIG. 2 is a configuration diagram of the receiving apparatus according to the present embodiment. This receiving apparatus shows the receiving unit of FIG. 1 in more detail, and is applied to an OFDM or OFDMA communication system. The received signal received by the antenna passes through the band pass filter 26 and is amplified by the low noise amplifier LNA and is amplified to a constant amplitude by the automatic gain control amplifier 27. Then, the received signal is multiplied by the local signals FL having phases of 0 ° and π / 2 generated by the reference signal oscillator VCO_RF by the mixers MIXi and MIXq, and down-converted to the baseband via the low-pass filter LPF. That is, the RF / IF circuit 20 and the quadrature detection circuit 21 in FIG. 1 are configured by the mixers MIXi and MIXq in the analog circuit 30A. Then, digital reception signals of I component and Q component are input to the digital baseband unit 30B by the analog-digital converter A / D.

  In the digital baseband unit 30B, the automatic gain control circuit AGC controls the gain of the automatic gain control amplifier 27 so that the amplitudes of the I signal and the Q signal are constant. Further, the DFT timing detection circuit 31 detects the DFT window from the I signal and the Q signal, and supplies the timing signal to the cyclic prefix (CP) removal circuit 32 and the DFT 33. The cyclic prefix removal circuit 32 removes the cyclic prefix added to the head of the effective symbol of the received signal from the time domain I signal and Q signal and outputs the effective symbol component. The DFT 33 outputs the time domain I signal. The signal and the Q signal are DFT transformed to extract a plurality of subcarriers in the frequency domain.

  Here, the DFT 33 is generally referred to as FFT (Fast Fourier Transform), which is a technique for discretizing the Fourier transform and calculating at high speed, and is referred to as a discrete Fourier transform (DFT) in this specification. That is, the analog-to-digital converter A / D converts an analog value at a sampling point at a constant time interval into a digital value, and the DFT 33 performs a DFT operation on the discrete value at that sampling point. That is, the DFT 33 performs a DFT operation on discrete sample values in the DFT window detected by the DFT timing detection circuit 31. This DFT operation will be described in detail later.

  Further, a channel estimation circuit (Channel Estimation) 34 detects the phase and amplitude of a pilot signal, which is a known signal inserted in some symbols and some subcarriers, and places it in a different position on the time axis. An average (time average) of phases and amplitudes (reference vectors) of a plurality of inserted pilot signals is calculated. Thus, by taking the average of the phase and amplitude (reference vector) of a plurality of pilot signals, the accuracy of the phase and amplitude of the reference vector is improved, and the reliability of the propagation path estimation value is improved. Then, the propagation path estimation circuit 34 performs phase rotation or the like on the I and Q signals in the data area of each subcarrier based on the reference vector of each subcarrier, thereby obtaining the absolute values of the I and Q signals in the data area. Get phase etc. If the modulation method is QPSK, the phase information of the data area is acquired, and if it is QAM, the phase information and amplitude information are acquired.

  Then, the demapping circuit 24 demaps the original transmission code from the coordinate points of the I signal and Q signal in the data area, and the decoder circuit 25 performs decoding such as error correction (FEC) and reception. Data Rxd is output.

  The DFT timing detection circuit 31, the cyclic prefix removal circuit 32, the DFT 33, the propagation path estimation circuit 34, and the like correspond to the demodulation circuit 23 in FIG.

  FIG. 3 is a configuration diagram of a propagation path estimation circuit. The propagation path estimation circuit 34 includes a reference vector detection unit 35 that detects the phase and amplitude (reference vector) of the DFT-transformed pilot signal, and an averaging unit that calculates an average value of a plurality of reference vectors detected for each subcarrier. 36, and a phase rotation unit 37 that rotates the phase of the signal in the data region of each subcarrier and corrects it to an absolute phase based on the phase and amplitude of the averaged reference vector, that is, a propagation path compensation circuit.

FIG. 4 is a diagram illustrating a relationship between a plurality of OFDM subcarriers and a plurality of symbols. In this figure, a plurality of subcarriers SC _{0 to} SC ₇ and a plurality of symbols S _{m + 1 to} S _{m + 9} are arranged, among which symbols S _{m + 2} and S _{m + 8} have pilots inserted in all subcarriers, and others Data is modulated on the subcarriers of symbols S _{m + 1} , S _{m + 3 to} S _{m + 7} , and S _{m + 9} .

In FIG. 4, it is assumed that the DFT window sample point shift does not occur in all symbols. In this case, the reference phase of each subcarrier is the same for all symbols. Therefore, the averaging unit 36 in FIG. 3 calculates the average values of the phases and amplitudes (reference vectors) of the pilot signals of the symbols S _{m + 2} and S _{m + 8} for each subcarrier, and the phase rotation unit 37 averages them. Each subcarrier of symbols S _{m + 3 to} S _{m + 7} between symbols S _{m + 2} and S _{m + 8} is corrected based on the phase and amplitude (reference vector) of the pilot signal.

  Here, the reference phase is a phase based on the timing of the DFT window. If a sample point shift of the DFT window occurs, the reference phase shifts. Therefore, the reference phase here is different from the reference vector of the pilot signal due to the propagation path characteristics.

FIG. 5 is a diagram illustrating a relationship between a plurality of OFDM subcarriers and a plurality of symbols. The plurality of subcarriers, the plurality of symbols, and the pilot are the same as those in FIG. In FIG. 5, the DFT window sample point shift occurs between the symbols S _{m + 4} and S _{m + 5} . Therefore, the reference phase of each symbol differs between the symbol before the occurrence of the sample point shift and the symbol after the occurrence.

Therefore, the phases and amplitudes (reference vectors) of the pilot signals of the symbols S _{m + 2} and S _{m + 8} are also different, and their average value becomes a reference vector with low accuracy. As a result, if each subcarrier of the data symbols S _{m + 3 to} S _{m + 7} is corrected using these average values, incorrect channel compensation is performed.

Therefore, in the present embodiment, as a first method, the pilot reference phase of the sample Sm + 8 is corrected so as to be the same as the pilot reference phase of the sample Sm _{+ 2} according to the number of sample point deviations. Is applied to the data symbols S _{m + 3} and S _{m + 4} before the sample point shift occurs. Further, the complex conjugate of the average value (I and Q signals with inverted Qch) is applied to perform propagation path compensation of the data symbols S _{m + 5} , S _{m + 6} , and S _{m + 7} after the sample point deviation occurs. Complex conjugate of the mean value, the sample S _{m + 2} of the pilot reference phase and the phase correction so as to match the sample S _{m + 8} pilot reference phase according to the deviation number of sample points, the pilot taking the average of It is the same as the average value of the reference vector.

  As a second method, all the data symbols and pilot signals after the occurrence of the sample point deviation are corrected to the reference phase before the occurrence of the sample point deviation. Then, the average value of the phase and amplitude (reference vector) of the pilot signal before the occurrence of the sample point deviation and the corrected pilot signal after the occurrence is applied to compensate the propagation path of the data symbol between the pilot signals.

  The first method only needs to perform phase correction corresponding to the deviation of the sample points only for the pilot signal, so that power consumption can be saved.

  Next, the DFT calculation will be described, and the phase correction when a sample point shift occurs in the DFT window will be described.

When the number of points, which is the number of discrete points in the DFT operation, is N, the DFT (Discrete Fourier Transform) operation is performed by performing the following operation on _N complex number sequences x _{0 to} x _N−1 to perform N operations. it is known that the a calculation for obtaining a complex sequence _X 0 _{~X N-1.}
X _j = Σ (k = 0 to N−1) x _k e ^{{(−2π / N) jk}}
In the OFDM or OFDMA system, this calculation is also performed in a DFT calculation for converting a time domain signal into a frequency domain signal. This arithmetic expression can be expressed by the following determinant.

FIG. 6 is a diagram showing a determinant of the DFT operation. In FIG. 6A, x _{0 to} x _N−1 are values of sample points _{0 to N−1} of the time domain signal to be DFT-calculated. W ^{0 to} W ^{(N−1) ^ 2} are twiddle factors multiplied by the time domain signals x _{0 to} x _N−1 . Further, X _{0 to} X _N−1 are frequency domain signals obtained by DFT calculation. The frequency domain signals _X 0 _{~X N-1} have different frequencies 0, f, 2f, 3f~ and (N-1) f, the respective phase and amplitude. Note that (N-1) ^ 2 means (N-1) ² .

  That is, in this DFT determinant, the number of points in the DFT window is N points (0 to N-1), and the number of subcarriers (frequency) is N (0 to N-1). The number of points is the number of sample points subject to DFT operation. When all sample points are points, all sample points in the DFT window are subject to DFT operation.

FIG. 6B shows an example in which the number of DFT points is N = 8. That is, the sample point _x 0 ~x ₇ time domain signal within the DFT window, the frequency domain signals _X 0 to X ₇ is calculated by multiplying the twiddle factor ^W 0 ^{to W-49.}

FIG. 7 is a diagram for explaining the twiddle factor W ^N of the DFT operation. FIG. 7A shows twiddle factors W ^{0 to} W ^N-1 when the number of DFT points is N. Twiddle factors W _N has a phase component which rotates clockwise from W ⁰ angle 0 ° by the angle 2 [pi / N. FIG. 7B shows twiddle factors W ^{0 to} W ⁷ when the number of DFT points is eight. In this case, it has a phase component that rotates clockwise from the angle 0 ° to 45 °.

It can be understood that the twiddle factors W ^{0 to} W ^{N−1 in} FIGS. 6 and 7 correspond to e ^{{(−2π / N) jk}} in the above DFT arithmetic expression.

8, 9, Q components of the frequency domain signals _X 0 to X ₇ after the Fourier transform (sin), is a diagram showing a signal waveform of the eight sub-carrier of the I component (cos). From this figure, the meaning of the phase of the twiddle factors W ^{0 to} W ^N-1 is understood. Unlike the DFT operation, the FT operation is an operation for continuous points. Therefore, as shown in FIGS. 8 and 9, the frequency domain signals X _{0 to} X ₇ consist of continuous points and can recognize the waveform. On the other hand, in the case of the DFT calculation, the calculation is performed on discrete points called sample points, and the obtained frequency domain signal is also discrete point data.

As shown in this figure, the frequency of each subcarrier _X 0 to X ₇ are, _{X 0} is 0 _(DC), X 1 so to X ₇ is a F～7f, each with the same eight sampling points SP0~SP7 The subcarriers have different phases. For example, as shown in FIG.
X ₁ : 0, π / 4, 2π / 4, 3π / 4, 4π / 4, 5π / 4, 6π / 4, 7π / 4
X ₂ : 0, 2π / 4, 4π / 4, 6π / 4, 8π / 4, 10π / 4, 12π / 4, 14π / 4
X ₃ : 0,3π / 4, 6π / 4, 9π / 4, 12π / 4, 15π / 4, 18π / 4, 21π / 4
X ₄ : 0,4π / 4, 8π / 4, 12π / 4, 16π / 4, 20π / 4, 24π / 4, 28π / 4
X ₅ : 0,5π / 4, 10π / 4, 15π / 4, 20π / 4, 25π / 4, 30π / 4, 35π / 4
X ₆ : 0,6π / 4, 12π / 4, 18π / 4, 24π / 4, 30π / 4, 36π / 4, 42π / 4
X ₇ : 0,7π / 4, 14π / 4, 21π / 4, 28π / 4, 35π / 4, 42π / 4, 49π / 4
Thus, since the frequency of each subcarrier has a relationship of 0, f to 7f, the phase at each sampling point is also different. The above phase coincides with the phase components of the twiddle factors W ^{0 to} W ^{49 although} the signs are opposite.

FIG. 10 is a diagram showing a DFT determinant when the DFT window is shifted by one sample point because the phase of the incoming wave is advanced. The DFT 33 in FIG. 2 performs an SFT calculation on the sample points of the time domain signal within a predetermined DFT window. Therefore, when the phase of the incoming wave advances and the DFT window shifts by one sample point, the sample points of the time domain signal to be subjected to the DFT calculation are x _{1 to} x ₇ and x ₀ . Therefore, the frequency domain signals X _{0 to} X of each subcarrier obtained by the DFT operation as a result of multiplying the sampling points x _{1 to} x ₇ and x ₀ of the time domain signal by the twiddle factors W ^{0 to} W ⁴⁹ by the DFT operation. ₇ is advanced by the phase of one sample point.

In Figure 10, the sample point x ₁ ~x _7, x ₀ of the original if the time domain signals, where should the rotation factors surrounded by a broken line in FIG multiplied, wrong rotation factor W ⁰ is different from the broken lines - W ⁴⁹ is multiplied. As a result, the frequency domain signals X _{0 to} X ₇ obtained by the DFT calculation have phases π / 4, 2π / 4, 3π / 4, 4π / 4, 5π / 4 corresponding to one sample point of each subcarrier. , 6π / 4, 7π / 4, the phase is advanced.

This phase shift corresponds to the phase of the rotation factors W ^{1 to} W ⁷ of each subcarrier, and corresponds to the phase of each subcarrier between the sample points SP0 and SP1 in FIG. Accordingly, the DFT calculation result X _{0 to} X ₇ is multiplied by the correction amount at the time of one sample shift indicated by the arrow in FIG. 8 to correct the phase shift of the DFT calculation result due to the one sample shift, and the DFT calculation result is corrected. It is possible to obtain a DFT calculation result without a window sample point shift.

This can be explained by actual matrix calculation as follows. In the matrix operation of Fig. 10, for example, the frequency domain signals _{X 1} is
X ₁ = W ⁰ * x ₁ + W ¹ * x ₂ + W ² * x ₃ + W ³ * x ₄ + W ⁴ * x ₅ + W ⁵ * x ₆ + W ⁶ * x ₇ + W ⁷ * x ₀
Therefore, by multiplying them by the rotation factor W ¹ (phase −π / 4) of the minimum angle corresponding to the deviation of one sample point of the subcarrier X ₁ ,
X ₁ = W ¹ * (W ⁰ * x ₁ + W ¹ * x ₂ + W ² * x ₃ + W ³ * x ₄ + W ⁴ * x ₅ + W ⁵ * x ₆ + W ⁶ * x ₇ + W ⁷ * x ₀ )
= W ¹ * x ₁ + W ² * x ₂ + W ³ * x ₃ + W ⁴ * x ₄ + W ⁵ * x ₅ + W ⁶ * x ₆ + W ⁷ * x ₇ + W ⁰ * x ₀
Next, it is possible to advanced phases obtain operation results X ₁ no sample point deviation is corrected.

Similarly, by multiplying the other frequency domain signals X _{2 to} X ₇ by the rotation factors W ^{2 to} W ⁷ of the respective minimum angles, it is possible to obtain a calculation result without sample point deviation.

In the case of two sample point shifts, the frequency domain signals X _{1 to} X ₇ have a twiddle factor corresponding to the two sample points of each subcarrier, that is, a phase twice the minimum angle twiddle factor W ^{1 to} W ^7. Multiplication by twiddle factors W ^{2 to} W ⁸ having can yield an operation result without sample point deviation. It is as follows.
X ₁ = W ² * (W ⁰ * x ₂ + W ¹ * x ₃ + W ² * x ₄ + W ³ * x ₅ + W ⁴ * x ₆ + W ⁵ * x ₇ + W ⁶ * x ₀ + W ⁰ * x ₁ )
= W ² * x ₂ + W ³ * x ₃ + W ⁴ * x ₄ + W ⁵ * x ₅ + W ⁶ * x ₆ + W ⁷ * x ₇ + W ⁰ * x ₀ + W ¹ * x ₁
In the case of a deviation of M sample points, if the frequency domain signals X _{1 to} X ₇ are multiplied by a rotation factor W ^{M to} W ^{M + 7} which is M times the minimum angle, an operation result without deviation of the sample points is obtained. be able to.

FIG. 11 is a diagram showing a DFT determinant when the DFT window is shifted by one sample point because the phase of the incoming wave is delayed. Sample point _x 7 of the time domain _signal, x 0 ~x ₆ in rotation factor ^W 0 ^{results to W-49} is multiplied, the frequency domain signals _X 0 to X ₇ of each subcarrier obtained in the DFT calculation, respectively 1 It is delayed by the phase of the sample point.

11, the sample point x _7, x ₀ ~x ₆ of the original if the time domain signals, where should the rotation factors surrounded by a broken line in FIG multiplied, wrong rotation factor W ⁰ is different from the broken lines - W ⁴⁹ is multiplied. As a result, the frequency domain signals X _{0 to} X ₇ obtained by the DFT calculation have phases π / 4, 2π / 4, 3π / 4, 4π / 4, 5π / 4 corresponding to one sample point of each subcarrier. , 6π / 4, 7π / 4, the phase is delayed.

Therefore, in this case, the frequency domain signals X _{1 to} X ₇ after the DFT operation are divided by the minimum angle twiddle factors W ^{1 to} W ⁷ corresponding to one sample point of each carrier or the twiddle factors W ^{1 to} W If the complex conjugate of ⁷ is multiplied, an operation result with no sample point deviation can be obtained.

According to the above description, when the phase of the incoming wave advances, the twiddle factor W ¹ corresponding to the minimum angle is multiplied by the number of sample point deviations (phase rotation) to correct the subcarrier X ₁ of the frequency f. The phase is obtained, and the correction phase of subcarriers X _{2 to} X ₇ of frequencies 2f to 7f (in the case of N, (N−1) f) is 2 to 7 times the correction phase of subcarrier X ₁ of frequency f, respectively. (In the case of N, (N-1) times). Then, multiply their correction phase in _X 1 to X ₇ which is DFT operation to (phase rotation).

Instead, the correction phase of the subcarrier X ₁ of frequency f is obtained by multiplying (phase rotation) the rotation factor W ¹ corresponding to the minimum angle by the number of times (number of DFT points−number of sample point deviations) as the correction phase. Then, the correction phase of other subcarriers X _{2 to} X ₇ is also obtained by multiplying by 2 to 7 times, and the corrected phase is divided by DFT operation X _{1 to} X ₇ or multiplied by a complex conjugate (phase rotation). You may make it do.

Conversely, in the above description, when the phase of the incoming wave is delayed, the rotation factor W ¹ corresponding to the minimum angle is multiplied by the number of sample point deviations (phase rotation), and the subcarrier X ₁ of the frequency f The correction phase is obtained, and the correction phases of the subcarriers X _{2 to} X ₇ having the frequencies 2f to 7f are obtained by multiplying the correction phase of the subcarrier X _{1 having} the frequency f by 2 to 7 times, respectively. Then, X _{1 to} X ₇ subjected to the DFT operation are divided by their correction phases or multiplied by complex conjugates (phase rotation).

Instead, the correction phase of the subcarrier X ₁ of frequency f is obtained by multiplying (phase rotation) the rotation factor W ¹ corresponding to the minimum angle by the number of times (number of DFT points−number of sample point deviations) as the correction phase. Then, the correction phase of the other subcarriers X _{2 to} X ₇ is also obtained by multiplying by 2 to 7 times, and X _{1 to} X ₇ subjected to the DFT operation are multiplied (phase rotation) by those correction phases. Good. When advanced one sample point worth phase of FIG. 10, the frequency domain signals _X 1 to X ₇ which are FT operation corresponding to the frequency domain signals _X 1 to X ₇ which is DFT operation is as follows.
X ₁ = A ₁ e ^{j (2πft + θ1 + π / 4)}
Here, A ₁ is the amplitude, θ ₁ is the phase component, and f is the frequency component.
X ₂ = A ₂ e ^{j (2π2ft + θ2 + 2π / 4)}
X ₃ = A ₃ e ^{j (2π3ft + θ3 + 3π / 4)}
X ₄ = A ₄ ej ^{(2π4ft + θ4 + 4π / 4)}
X ₅ = A ₅ ej ^{(2π5ft + θ5 + 5π / 4)}
X ₆ = A ₆ e ^{j (2π6ft + θ6 + 6π / 4)}
X ₇ = A ₇ ej ^{(2π7ft + θ7 + 7π / 4)}
As shown in FIG. 8, the phase of X ₁ is advanced by π / 4 corresponding to one sample point due to a deviation of one sample point in the DFT window. Therefore, the minimum angle twiddle factor corresponding to one sample point W ¹ = e ^{−j (π / 4)}
Is multiplied by
^{_{X 1 * W 1 = A 1}} e j (2πft + θ1 + π / 4) * e -j (π / 4) = A 1 e j (2πf1t + θ1)
The advanced phase component can be removed.

Similarly,
X ₂ = A ₂ e ^{j (2π2ft + θ2 + 2π / 4)}
W ² = e ^{−j (2π / 4)}
Is multiplied by
^{_{X 2 * W 2 = A 2}} e j (2π2ft + θ2 + 2π / 4) * e -j (2π / 4) = A 2 e j (2π2ft + θ2)
The same applies to X _{3 to} X ₇ .

On the other hand, in the case of FIG. 11, X ₁ is delayed by π / 4 corresponding to one sample point due to a shift of one sample point in the DFT window.
X ₁ = A ₁ e ^{j (2πft + θ1−π / 4)}
Therefore, the twiddle factor W ¹
W ¹ = e ^{−j (π / 4)}
Dividing
^{_{X 1 / W 1 = A 1}} e j (2πft + θ1-π / 4) / e -j (π / 4) = A 1 e j (2πft + θ1)
Similarly,
X ₂ = A ₂ ej ^{(2π2ft + θ2-2π / 4)}
W ² = e ^{−j (2π / 4)}
Dividing
^{_{X 2 / W 2 = A 2}} e j (2π2ft + θ2-2π / 4) / e -j (2π / 4) = A 2 e j (2π2ft + θ2)
Since the above division is equivalent to multiplying the complex conjugate of the twiddle factor,
^{_{X 1 * W 1 = A 1}} e j (2πft + θ1-π / 4) * e -j (-π / 4) = A 1 e j (2πft + θ1)
^{_{X 2 * W 2 = A 2}} e j (2π2ft + θ2-2π / 4) * e -j (-2π / 4) = A 2 e j (2π2ft + θ2)
It becomes. The same applies to the other frequency domain signals X _{3 to} X ₇ .

As described above, when the DFT window is deviated from the original window, X _{1 to} X ₇ after the DFT calculation are performed according to the number of sample points deviated and the phase of the minimum angle of the rotation factor of each subcarrier. What is necessary is just to correct | amend a phase. Then, the phase of the minimum angle of rotation factor of each sub-carrier is phase as W ¹ to ^W-7 of FIG. Since X ₀ after the DFT operation is a DC wave having no phase component, phase correction is not necessary, and correction may be made with a phase zero rotation factor W ⁰ .

From the above, the correction value of the sample point deviation of the DFT window can be obtained by calculation by rotating the phase of the twiddle factor W ¹ having the minimum angle corresponding to the number of sample point deviations and the subcarriers. Using this concept, in this embodiment, the propagation path estimation circuit CE obtains the number of sample point deviations from the DFT timing signal, and obtains the phase correction amount of each subcarrier according to the number of deviations. Specific examples will be described below.

[First Embodiment]
FIG. 12 is a configuration diagram of a propagation path estimation circuit according to the first embodiment. In this embodiment, for simplification, the number of points of DFT is 8. The propagation path estimation circuit 34 has a data pilot identification unit 45 that identifies and separately outputs a data signal and a pilot signal from the I and Q channel signals DFT computed by the DFT computation unit 33. Furthermore, the propagation path estimation circuit 34 includes a pilot signal correction unit 40 that corrects the sample point deviation of the DFT window of the pilot signal. The pilot signal correction unit 40 includes a phase rotation amount calculation unit 41 and a phase correction unit 42.

  The phase rotation amount calculation unit 41, in accordance with the deviation (number of sample point deviations) from the original DFT timing of the DFT timing signal S31 generated by the DFT timing generation unit 31, and the minimum angle of the rotation factor of each subcarrier, A phase rotation amount for correcting the sample point deviation is calculated. Then, the phase correction unit 42 performs phase correction by rotating the phase of the pilot signal subjected to the DFT calculation by the rotation amount calculated by the phase rotation amount calculation unit 41. The phase correction unit 42 includes a phase rotator described later.

  Further, in the propagation path estimation circuit 34, the pilot average unit 36 calculates an average value of a plurality of pilot signals whose DFT window sample point deviation is corrected by the pilot signal correction unit 40. The propagation path compensation unit 37 corrects the phase and amplitude of the data signal based on the phase and amplitude (reference vector) of the time-averaged pilot signal. The data signal is delayed for a predetermined time by the data delay unit 46 and matched with the output timing of the pilot averaging unit 36. The propagation path compensation unit 37 is a phase rotator described later, and is a phase rotator equivalent to the phase correction unit 42 in the present embodiment.

When the example of FIG. 5 is applied, the pilot signal of the symbol S _{m + 8} is a symbol after the occurrence of the sample point deviation, so that the pilot signal correction unit 40 performs phase correction corresponding to the sample point deviation. Specifically, it is a phase rotation process. Then, a time average of the pilot signal of the symbol S _{m + 8} phase-corrected by the pilot averaging unit 36 and the pilot signal of the symbol S _{m + 2} without sample point deviation is obtained. Further, in the propagation path compensation unit 37, propagation path compensation is performed on the data signals of the symbols S _{m + 3} and S _{m + 4} before the occurrence of the sample point deviation with a time-averaged pilot signal. Specifically, it is a phase rotation process.

Further, for symbols S _{m + 5} , S _{m + 6} , and S _{m + 7} after the occurrence of the sample point deviation, it is desirable that the propagation path compensation is performed with the time average of the pilot signal phase-corrected after the sample point deviation.

For this purpose, a complex conjugate generator 43 for obtaining a complex conjugate of the time average of the pilot signal of the symbol S _{m + 8 and} the pilot signal of the symbol S _{m + 2 having} no sample point deviation, and a Q channel code inversion enable signal S44. And a selector 44 for selecting the output of the complex conjugate generator 43 according to the above. The complex conjugate is phase rotated pilot signal symbols S _{m + 2} in the opposite direction of the phase rotation amount by the phase rotation amount calculation section 41 calculated, consistent with those obtained by averaging pilot signals and time symbols S _{m + 8.} Then, the propagation path compensation unit 37 performs propagation path compensation on the symbols S _{m + 5} , S _{m + 6} , and S _{m + 7} after the occurrence of the sample point deviation by the complex conjugate of the pilot signal.

  FIG. 13 is a configuration diagram of the pilot signal correction unit. As described with reference to FIG. 12, the pilot signal correction unit 40 includes a phase rotation amount calculation unit 41 and a phase correction unit 42. In FIG. 13, reference numerals 410 to 415 are assigned to the phase rotation amount calculation unit 41. FIG. 13 shows signals S1 and S2 in the sample point deviation correction processing in each symbol period when there is no sample point deviation, when one sample point deviation occurs, and when two sample point deviations occur. , S3, S4 are shown.

The phase rotation amount calculation unit 41 calculates the phase rotation amount in the phase correction of the pilot signal after the DFT calculation associated with the sample point deviation for each subcarrier. The phase rotation amount calculation unit 41 has a W ¹ table 410 in which I and Q channel information S0 of a rotation factor W ¹ having minimum angle information among a plurality of rotation factors is stored. The I and Q channel signals of this twiddle factor W ¹ are I = 1 / √2, Q = −1 / √2.

  Further, the control signal generation unit 413 obtains the number of DFT window shifts (number of sample points) from the DFT timing signal S31, generates the enable signal S413 corresponding to the number of shifts, and generates the reset signal RST in synchronization with the symbol period. , The aforementioned Q channel code inversion enable signal S44 is generated.

Furthermore, the phase rotation amount calculation unit 41 includes a flip-flop 412 and a rotation amount calculation unit 411 of the subcarrier SC1, and obtains the phase rotation amount of the subcarrier SC1 (X1) according to the sample point deviation amount. When the control signal generation unit 413 outputs the reset signal RST at the start timing of the symbol period, the flip-flop 412 is reset, and the signal S1 of the two output terminals Q includes the I channel component “1” and the Q channel component “ 0 "is output. This corresponds to the rotation factor ^{W 0.} When the control signal generation unit 413 outputs the enable signal S413 as many times as the number of sample point deviations, the flip-flop 412 causes the SC1 rotation amount calculation unit 411 to output the initial value (I = 1 / √2, Q = −) of the signal S1. I / Q signals obtained by rotating 1 / √2) by the number of sample point deviations of the minimum angle information W1 of the rotation factor are output.

  The number of rotations is (DFT point number−sample point deviation number) when the phase of the incoming wave is advanced as shown in FIG. 10, and sample point deviation when the phase is delayed as shown in FIG. Is a number.

Therefore, as shown in FIG. 13, the rotation factors W ⁰ is the signal S1 is at no symbol period sample point deviation, the signal S1 at symbol periods after one sample point deviation of the phase delay in the rotation factor W ^2, further phase signal S1 becomes twiddle factor W ³ being a symbol period after 2 sample point deviation of delay. On the contrary, although not shown in FIG. 13, the signal S1 becomes the twiddle factor W ^{2 *} in the symbol period after one sample point of phase advance, and the signal S1 becomes the symbol period after two sample points of phase advance. The twiddle factor W ^{3 *} . * Indicates a complex conjugate.

Furthermore, the phase rotation amount calculation unit 41 includes a flip-flop 415 and each SC rotation amount calculation unit 414, and obtains the phase rotation amount corresponding to the sample point shift corresponding to the subcarrier. That is, the flip-flop 415 outputs the I channel component “1 / √2” and the Q channel component “−1 / √2” corresponding to the twiddle factor W ₀ to the signal S2 of both outputs Q by the reset signal RST. Is done. Then, in synchronization with the clock CLK that controls the phase correction timing of the subcarriers X0 to X7 within the symbol period, the flip-flop 415 causes each SC rotation amount calculation unit 414 to generate an initial value (I = 1 / √2) of the signal S2. , Q = −1 / √2), the phase correction amount S2 whose phase is rotated in order by the phase rotation amount S1 held in the flip-flop 412 is output.

  Then, the phase correction unit 42 multiplies the pilot signal S3 output from the data pilot identification unit 45 by the complex conjugate of the phase correction amount S2 (phase rotation), and the pilot signal S4 in which the phase shift due to the sample point shift is corrected. Is output.

Further, the control signal generation unit 413 generates a Q channel code inversion enable signal S44 to be output to the upstream selector 44 of the downstream channel compensation unit 37. This is because in the propagation path compensation for the data signal of the symbol after the sample point deviation detected based on the DFT timing signal S31, as described in the example of FIG. 5, the pilot signal of the symbol S _{m + 2} and the symbol S _{m + 8} This corresponds to the need to multiply symbols S _{m + 5} , S _{m + 6} , and S _{m + 7} after the sampling point deviation by a time-average complex conjugate with the pilot signal of which the sampling point deviation is corrected. Therefore, the control signal generation unit 413 outputs the Q channel code inversion enable signal S44 at the phase correction timing of the symbols S _{m + 5} , S _{m + 6} , S _{m + 7} by the propagation path compensation unit 42.

  The operations of the phase rotation amount calculation unit 41 and the phase correction unit 42 constituting the pilot signal correction unit 40 will be described according to the timing chart of the signals S1, S2, S3, S4 shown in FIG.

First, when there is no sample point shift in the leftmost DFT window of the timing chart of FIG. 13, the control signal generation unit 413 detects that there is no sample point shift based on the DFT timing signal S31 and enables it. The signal S413 is not output. Thereby, in response to the reset signal RST, the flip-flop 412 outputs the I and Q signals (I = 0, Q = 0) of the twiddle factor W ⁰ from the output signal S1. In response to the reset signal RST, the flip-flop 415 also outputs the I and Q signals of the twiddle factor W ⁰ from the output signal S2. Each SC rotation amount calculation unit 414 multiplies the output S2 of the flip-flop 415 by the rotation factor W ⁰ having the angle of 0 °, so that all the signals S2 stored in the flip-flop 415 in synchronization with the clock CLK are W ⁰ .

As a result, the phase correction unit 42 multiplies the pilot signal S3 of each subcarrier by a twiddle factor W0 ^* with a phase correction amount of 0 °, that is, without correcting the phase, the pilot signal S3 is corrected as it is. Output as pilot signal S4.

Next, when a phase lag of one sample point occurs in the middle DFT window of the timing chart of FIG. 13, the control signal generator 413 detects the one sample point lag based on the DFT timing signal S31, The enable signal S413 is output once. Thus, after the flip-flop 412 outputs the I and Q signals of the twiddle factor W ⁰ from the output signal S 1 in response to the reset signal RST, the I signal of the twiddle factor W ¹ is output from the output signal S 1 by one enable signal S 413. , Q signals (I = 1 / √2, Q = −1 / √2) are output.

At the beginning of the symbol period, in response to the reset signal RST, the flip-flop 415 outputs the I and Q signals of the twiddle factor W ⁰ as the output signal S2. Each SC rotation amount calculation section 414, because multiplying the twiddle factor ^{W 1} of this angle - [pi] / 4 to the output S2 of the flip-flop 415, flip-flop 415, ^{W 1} and signal S2 in synchronization with the clock CLK , W ² , W ³ , W ⁴ , W ⁵ , W ⁶ , W ⁷ .

As a result, the phase correction unit 42 performs a phase correction amount of 0 °, −π / 4, −2π / 4, −3π / 4, −4π / 4, −5π / 4, −6π / with respect to the pilot signal S3. 4, −7π / 4 twiddle factors W ⁰ , W ¹ , W ² , W ³ , W ⁴ , W ⁵ , W ⁶ , W ⁷ complex conjugate (phase rotation amount is 0 °, + π / 4, + 2π / 4, + 3π / 4, + 4π / 4, + 5π / 4, + 6π / 4, + 7π / 4), and the pilot signal S3 is output as a phase-corrected pilot signal S4.

When a two sample point shift occurs in the rightmost DFT window in the timing chart of FIG. 13, the control signal generation unit 413 detects the two sample point shift based on the DFT timing signal S31 and outputs the enable signal S413 twice. Output. Thus, after the flip-flop 412 outputs the I and Q signals of the twiddle factor W ⁰ from the output signal S 1 in response to the reset signal RST, the I signal of the twiddle factor W ² is output from the output signal S 1 by the two enable signals S 413. , Q signal (I = 0, Q = −1) is output.

At the beginning of the symbol period, in response to the reset signal RST, the flip-flop 415 outputs the I and Q signals of the twiddle factor W ⁰ as the output signal S2. Each SC rotation amount calculation section 414, because multiplying the twiddle factor ^{W 2} of this angle -2.pi. / 4 to the output S2 of the flip-flop 415, flip-flop 415, the signal S2 ^{W 2} in synchronization with the clock CLK , W ⁴ , W ⁶ , W ⁸ , W ¹⁰ , W ¹² , W ¹⁴ .

As a result, the phase correction unit 42 performs a phase correction amount of 0 °, −2π / 4, −4π / 4, −6π / 4, −8π / 4, −10π / 4, −12π / with respect to the pilot signal S3. 4, −14π / 4 twiddle factors W ⁰ , W ² , W ⁴ , W ⁶ , W ⁸ , W ¹⁰ , W ¹² , W ¹⁴ complex conjugate (phase rotation amount is 0 °, + 2π / 4, + 4π / 4, + 6π / 4, + 8π / 4, + 10π / 4, + 12π / 4, + 14π / 4), and the pilot signal S3 is output as a phase-corrected pilot signal S4.

  As described above, the phase-corrected pilot signal is averaged by the pilot averaging unit 36 and the data conjugate is multiplied by the complex conjugate of the average pilot signal by the propagation path compensation unit 37 as described with reference to FIG. As a result, the I and Q signals of the data signal from which distortion due to the propagation path has been removed are output to the demapping unit 24.

As can be understood from the twiddle factor shown in FIG. 7, when the point point is 8, the twiddle factors W ⁵ , W ⁶ , and W ⁷ are complex conjugates of the twiddle factors W ¹ , W ² , and W ³ . Therefore, the twiddle factors W ⁵ , W ⁶ , and W ⁷ do not necessarily have to rotate −π / 4, which is the phase of the twiddle factor W 1 having the minimum angle from the twiddle factor W ⁰ , by 5, 6 and 7 times, respectively. It is possible to obtain + π / 4, which is the phase of the complex conjugate of the twiddle factor W ¹ , by rotating the phase by 3, 2, and 1 times. Moreover, the number of rotations is reduced by using the complex conjugate.

FIG. 14 is a configuration diagram of the pilot signal correction unit 40 in which the SC1 rotation amount calculation unit 411 rotates in phase with the complex conjugate of the twiddle factor W1. The configuration different from FIG. 13 is that a complex conjugate generation unit 416 that performs sign inversion of the Q channel output of the W1 table 410 that stores the minimum angle information W ¹ , and its output or a Q channel output that is not sign inverted. A selector 417 that selects according to the selection signal S417.

Accordingly, when the signal S1 is set to W ¹ , W ² , W ³ according to the number of sample point deviations detected based on the DFT timing signal S31, the control signal generation unit 413 does not code the selection signal S417. On the inversion side, the enable signal S413 is output once, twice and three times, respectively. This operation is the same as in FIG. On the other hand, when the signal S1 is set to W ⁵ , W ⁶ , W ⁷ , the control signal generation unit 413 sets the selection signal S417 to the sign inversion side and outputs the enable signal S413 three times, two times, and once, respectively. To do. Thus, the signal S1 is rotated from ^{W 0} and ^{W 7, W 6, W 5} . Other configurations and operations are the same as those in FIG.

  When the incoming wave has a phase advance as shown in FIG. 10, the selection signal S417 generated by the control signal generation unit 413 is the complex conjugate generation unit 416 if (number of points−number of sample point deviations) exceeds N / 2. The enable signal S417 is output (number of points− (number of points−number of sample point deviations)) times. Conversely, in the case of phase delay, the selection signal S417 generated by the control signal generation unit 413 selects the complex conjugate generation unit 416 side if the sample point deviation number exceeds N / 2, and the enable signal S417 is ( The number of points minus the number of sample point deviations)) is output only times. As a result, the number of rotation calculations of the SC1 rotation amount calculation unit 411 can be suppressed to N / 2 at the maximum.

  FIG. 15 is a configuration diagram of the SC1 rotation amount calculation unit 411 and each SC rotation amount calculation unit 414 in the phase rotation amount calculation unit 41. The signal in the figure indicates the signal of the SC1 rotation amount calculation unit 411, and the parenthesis indicates the signal of each SC rotation amount calculation unit 414. M1 to M4 denote multipliers (mixers), D1 denotes a subtractor, and D2 denotes an adder.

  In the case of the SC1 rotation amount calculation unit 411, the output S1 of the flip-flop 412 is multiplied by the signal S0 from the W1 table 410 and output to the input of the flip-flop 412. On the other hand, in the case of each SC rotation amount calculation unit 414, the output S2 of the flip-flop 415 is multiplied by the output S1 of the flip-flop 412 and output to the input of the flip-flop 415. In either case, since the input has an amplitude component of 1, both operate as a phase rotator that performs phase rotation on the signals S1 and S2 by the phases of the signals S0 and S1.

  FIG. 16 is a configuration diagram of the phase correction unit 42. M1 to M4 are multipliers (mixers), D1 is an adder, and D2 is a subtractor. In FIG. 15, the subtraction and addition of D1 and D2 are reversed. Therefore, the phase correction unit 42 in FIG. 16 is a phase rotator, which multiplies the pilot signal S3 before phase correction by the complex conjugate of the phase rotation amount S2, and outputs a phase-corrected pilot signal S4.

  The propagation path compensation unit 37 in the propagation path estimation circuit 34 also has the same configuration as the phase rotator in FIG. In this case, a data signal that is a compensated signal is input to signal S3 in FIG. 16, and an average value of pilot signals having a reference phase is input to signal S2.

If the average value of the pilot signal is the reference phase before the sampling point shift, the complex conjugate of the average value of the pilot signal is equal to the data signal of the same reference phase, S _{m + 3} and S _{m + 4} in FIG. Multiplying to remove propagation path characteristics.

On the other hand, when the average value of the pilot signal is the phase after the sample point deviation, the data signal after the sample point deviation is compared with the data signal after the same sample point deviation, S _{m + 5} , S _{m + 6} , and S _{m + 7} in FIG. The complex conjugate of the average value of the phase pilot signal is multiplied to simultaneously remove the propagation path characteristic and the sample point shift.

Channel estimation circuit 34 described above, the first, if only store minimum angle information S0 twiddle factor W ¹ in table 410, and all relative signals X ₁ to X ₇ of all the sub-carrier samples The correction phase amount with respect to the number of point deviations can be obtained by calculation. Therefore, the storage capacity of the table 410 can be reduced. Usually, since the number of points of DFT is a large number such as 2 ¹⁰ = 1024, it is significant to reduce the storage capacity of the table 410.

  Second, the propagation path estimation circuit 34 performs phase correction only on the pilot signal affected by the sample point deviation of the DFT window to correct the reference phase without the influence of the sample point deviation. Therefore, unlike the second embodiment to be described later, phase correction is not performed on the data signal, so that power can be saved accordingly.

  Third, a complex conjugate generation unit 43 and a selector 44 are provided on the output side of the pilot signal averaging unit 36 so that the pilot signal average value of the reference phase before the sample point deviation or the reference phase pilot after the sample point deviation is obtained. If any one of the signal average values is obtained, it is only necessary to select the Q channel signal generated by the complex conjugate generator 43 as the other average value. Therefore, the configuration of the pilot averaging unit 36 can be simplified.

[Second Embodiment]
FIG. 17 is a configuration diagram of a propagation path estimation circuit according to the second embodiment. Also in this embodiment, the number of DFT points is 8. The propagation path estimation circuit 34 includes a sample point deviation correction unit 40A that corrects the sample point deviation of the DFT window for the pilot signal and the data signal I and Q channel signals DFT computed by the DFT computation unit 33. The sample point deviation correction unit 40A includes a phase rotation amount calculation unit 41A and a phase correction unit 42A.

  The phase rotation amount calculation unit 41A performs sampling according to the deviation from the original timing (number of sample point deviations) of the DFT timing signal S31 generated by the DFT timing generation unit 31 and the minimum angle of the rotation factor of each subcarrier. A phase rotation amount for correcting the point shift is calculated. Then, the phase correction unit 42A performs phase correction by rotating the phases of the pilot signal and the data signal that have been subjected to the DFT calculation by the phase rotation amount calculated by the phase rotation amount calculation unit 41A.

  Furthermore, the propagation path estimation circuit 34 includes a data pilot identification unit 45 that identifies a data signal and a pilot signal from the signal corrected for sample point deviation and outputs them separately. In the propagation path estimation circuit 34, the pilot average unit 36 calculates an average value of a plurality of pilot signals in which the sample point deviation of the DFT window is corrected by the sample point deviation correction unit 40A. The propagation path compensation unit 37 corrects the phase and amplitude of the data signal based on the phase and amplitude (reference vector) of the pilot signal averaged over time. The data signal is delayed for a predetermined time by the data delay unit 46 and matched with the output timing of the pilot averaging unit 36.

That is, the propagation path estimation circuit 34 of the second embodiment will be described by applying the example of FIG. 5. Signals after the symbol S _{m + 5} in which the sample point deviation has occurred are represented by the symbols S _{m + 5} , S _{m + 6} , S _{m + 7.} Both the data signal and the pilot signal of the symbol S _{m + 8} are subjected to phase correction corresponding to the sample point shift by the sample point shift correction unit 40A. As a result, all the symbols before and after the occurrence of the sample point deviation are aligned with the same reference phase, and the pilot averaging unit 36 performs the phase correction of the pilot signal of the symbol S _{m + 8 and} the pilot signal of the symbol S _{m + 2} without the sample point deviation. In the propagation path compensation unit 37, the data signals of the symbols S _{m + 3} and S _{m + 4} before the occurrence of the sample point deviation and the data signals of the symbols S _{m + 5} , S _{m + 6} and S _{m + 7} whose sample point deviation has been corrected are obtained. On the other hand, the propagation path is compensated with a time-averaged pilot signal.

  FIG. 18 is a configuration diagram of the sample point deviation correction unit 40A. The configuration is almost the same as the sample point deviation correction unit 40 of FIG. The difference is that the control signal generation unit 413 does not output the Q channel code inversion enable signal, and the phase correction unit 42 performs phase correction on the data signal in addition to the pilot signal.

The SC1 rotation amount calculation unit 411 and the flip-flop 412 calculate the phase correction amount for the subcarrier SC1 (X ₁ ) corresponding to the number of sample point deviations to the signal S1, and each SC rotation amount calculation unit 414 and the flip-flop. In 415, the phase correction amount for each of the subcarriers SC2 to SC7 (X _{2 to} X ₇ ) is calculated as the signal S2, as in FIG.

  FIG. 19 is another configuration diagram of the sample point deviation correction unit 40A. The configuration is almost the same as that of the sample point deviation correction unit 40 of FIG. The difference is that the control signal generation unit 413 does not output the Q channel code inversion enable signal, and the phase correction unit 42 performs phase correction on the data signal in addition to the pilot signal.

The SC1 rotation amount calculation unit 411 and the flip-flop 412 calculate the phase correction amount for the subcarrier SC1 (X ₁ ) corresponding to the sample point deviation number to the signal S1, and the number of sample point deviations in the case of phase delay, When the phase advance (number of points−number of sample point deviations) exceeds 1/2 (N / 2) of the number of points, the SC1 phase amount calculation unit 411 calculates the complex conjugate of W ¹ in the table 410. The multiplication and the calculation of the phase correction amount for each of the subcarriers SC2 to SC7 (X _{2 to} X ₇ ) in each SC rotation amount calculation unit 414 and flip-flop 415 are the same as in FIG. .

  4 and 5, pilot signals are interpolated in all subcarriers, but even if pilot signals are interpolated in some subcarriers of a plurality of subcarriers, pilot signals are interpolated based on them. A reference vector of subcarriers not inserted can be obtained. Therefore, in this case, of course, the present embodiment can be applied.

  The above embodiment is summarized as follows.

(Appendix 1)
In a receiving apparatus that receives a signal transmitted using a multicarrier,
A DFT unit that converts a time domain received signal into a plurality of frequency domain subcarrier signals by performing a discrete Fourier transform (DFT) operation;
A propagation path estimation circuit that corrects distortion due to a propagation path from a data signal of the plurality of subcarrier signals based on a reference vector of a pilot signal interpolated in the plurality of subcarrier signals;
The propagation path estimation circuit includes a minimum angle information storage unit that stores minimum angle information of a twiddle factor of a DFT operation, and a first angle for the plurality of subcarrier signals corresponding to the minimum angle information according to the number of sample point deviations of the DFT window. A phase rotation amount calculation unit for calculating the phase rotation amount of each of the subcarrier signals, and a phase correction unit for correcting the phases of the plurality of subcarrier signals based on the first phase rotation amount.

(Appendix 2)
In Appendix 1,
The phase rotation amount calculation unit
A second phase rotation amount calculation unit for calculating a second phase rotation amount by rotating the minimum angle information according to the number of sample point deviations of the DFT window; and the second phase rotation amount for each subcarrier. A receiving apparatus comprising: a first rotation amount calculation unit configured to calculate a first phase rotation amount for the plurality of subcarrier signals by performing phase rotation according to a signal.

(Appendix 3)
In Appendix 2,
The second rotation amount calculation unit subtracts the sample point deviation number from the number of points in the complex conjugate of the minimum angle information when the sample point deviation number exceeds 1/2 of the number of points in the DFT calculation. A receiver that rotates the phase by a number.

(Appendix 4)
In any one of appendices 1 to 3,
The phase rotation amount calculation unit obtains the first phase rotation amount with respect to the subcarrier signal of the pilot signal among the plurality of subcarrier signals, and the phase correction unit is based on the first phase rotation amount. A receiving device for correcting a phase of a subcarrier signal of the pilot signal.

(Appendix 5)
In Appendix 4,
The propagation path estimation circuit further includes:
A pilot averaging unit that generates an average of the pilot signal of one symbol before and after the occurrence of the sample point deviation of the DFT window and the pilot signal of the other symbol before and after the occurrence and the phase-corrected pilot signal;
And a propagation path compensator that compensates for the data signal of one symbol before and after the DFT window sample point shift occurs according to the average value of the pilot signal.

(Appendix 6)
In Appendix 5,
The propagation path estimation circuit further includes:
A complex conjugate part for generating a complex conjugate of an average value of pilot signals generated by the pilot average part;
The propagation path compensation unit compensates for the data signal of the other symbol before and after the DFT window sample point shift occurs according to the complex conjugate of the average value of the pilot signals generated by the complex conjugate unit. .

(Appendix 7)
In Appendix 1 or 2,
The phase rotation amount calculation unit obtains the first phase rotation amount for the plurality of subcarrier signals,
The phase correction unit corrects the phase of the pilot signal and the subcarrier signal of the data signal based on the first phase rotation amount.

(Appendix 8)
Based on a DFT unit that performs discrete Fourier transform (hereinafter DFT) operation on a received signal in the time domain and converts it into a plurality of subcarrier signals in the frequency domain, and a reference vector of a pilot signal interpolated in the plurality of subcarrier signals In a propagation path compensation method for a receiving apparatus that receives a signal transmitted using a multicarrier having a propagation path estimation circuit that corrects distortion due to a propagation path from the data signals of the plurality of subcarrier signals,
A phase rotation amount calculation step of calculating first angle rotation amounts for the plurality of subcarrier signals according to the number of sample point deviations of the DFT window, as minimum angle information of a rotation factor of DFT operation;
And a phase correction step of correcting the phases of the plurality of subcarrier signals based on the first phase rotation amount.

(Appendix 9)
In Appendix 8,
The phase rotation amount calculating step includes:
A second phase rotation amount calculating step of calculating a second phase rotation amount by rotating the minimum angle information according to the number of sample point deviations of the DFT window;
And a first rotation amount calculation step of calculating a first phase rotation amount for the plurality of subcarrier signals by rotating the second phase rotation amount according to each subcarrier signal.

(Appendix 10)
In Appendix 9,
In the second rotation amount calculation step, when the sample point deviation number exceeds 1/2 of the DFT calculation point number, the complex conjugate of the minimum angle information is subtracted from the sample point deviation number. A propagation path compensation method that rotates the phase by a number.

(Appendix 11)
In any of appendices 8 to 10,
The phase rotation amount calculating step obtains the first phase rotation amount for the subcarrier signal of the pilot signal among the plurality of subcarrier signals,
The phase correction step is a propagation path compensation method for correcting a phase of a subcarrier signal of the pilot signal based on the first phase rotation amount.

34: propagation path compensation circuit 40: phase rotation amount calculation unit 41: phase rotation amount calculation unit 42: phase correction unit 410: minimum angle information table 411, 412: second phase rotation amount calculation units 414, 415: first Phase rotation amount calculation unit

Claims (8)

In a receiving apparatus that receives a signal transmitted using a multicarrier,
A DFT unit that converts a time domain received signal into a plurality of frequency domain subcarrier signals by performing a discrete Fourier transform (DFT) operation;
A propagation path estimation circuit that corrects distortion due to a propagation path from a data signal of the plurality of subcarrier signals based on a reference vector of a pilot signal interpolated in the plurality of subcarrier signals;
The propagation path estimation circuit includes a minimum angle information storage unit that stores minimum angle information of a twiddle factor of a DFT operation, and a first angle for the plurality of subcarrier signals corresponding to the minimum angle information according to the number of sample point deviations of the DFT window. A phase rotation amount calculation unit for calculating the phase rotation amount of each of the subcarrier signals, and a phase correction unit for correcting the phases of the plurality of subcarrier signals based on the first phase rotation amount. In claim 1,
The phase rotation amount calculation unit
A second phase rotation amount calculation unit for calculating a second phase rotation amount by rotating the minimum angle information according to the number of sample point deviations of the DFT window; and the second phase rotation amount for each subcarrier. A receiving apparatus comprising: a first rotation amount calculation unit configured to calculate a first phase rotation amount for the plurality of subcarrier signals by performing phase rotation according to a signal. In claim 2,
The second rotation amount calculation unit subtracts the sample point deviation number from the number of points in the complex conjugate of the minimum angle information when the sample point deviation number exceeds 1/2 of the number of points in the DFT calculation. A receiver that rotates the phase by a number. In any one of Claims 1 thru | or 3,
The phase rotation amount calculation unit obtains the first phase rotation amount with respect to the subcarrier signal of the pilot signal among the plurality of subcarrier signals, and the phase correction unit is based on the first phase rotation amount. A receiving device for correcting a phase of a subcarrier signal of the pilot signal. In claim 4,
The propagation path estimation circuit further includes:
A pilot averaging unit that generates an average of the pilot signal of one symbol before and after the occurrence of the sample point deviation of the DFT window and the pilot signal of the other symbol before and after the occurrence and the phase-corrected pilot signal;
And a propagation path compensator that compensates for the data signal of one symbol before and after the DFT window sample point shift occurs according to the average value of the pilot signal. In claim 5,
The propagation path estimation circuit further includes:
A complex conjugate part for generating a complex conjugate of an average value of pilot signals generated by the pilot average part;
The propagation path compensation unit compensates for the data signal of the other symbol before and after the DFT window sample point shift occurs according to the complex conjugate of the average value of the pilot signals generated by the complex conjugate unit. . In claim 1 or 2,
The phase rotation amount calculation unit obtains the first phase rotation amount for the plurality of subcarrier signals,
The phase correction unit corrects the phase of the pilot signal and the subcarrier signal of the data signal based on the first phase rotation amount. Based on a DFT unit that performs discrete Fourier transform (hereinafter DFT) operation on a received signal in the time domain and converts it into a plurality of subcarrier signals in the frequency domain, and a reference vector of a pilot signal interpolated in the plurality of subcarrier signals In a propagation path compensation method for a receiving apparatus that receives a signal transmitted using a multicarrier having a propagation path estimation circuit that corrects distortion due to a propagation path from the data signals of the plurality of subcarrier signals,
A phase rotation amount calculation step of calculating first angle rotation amounts for the plurality of subcarrier signals according to the number of sample point deviations of the DFT window, as minimum angle information of a rotation factor of DFT operation;
And a phase correction step of correcting the phases of the plurality of subcarrier signals based on the first phase rotation amount.

Images