JP2010147575A - Receiving device and symbol timing detecting method - Google Patents

Abstract

Symbol timing is accurately detected even in the presence of an interference wave.
A window function generator extracts a left half section T11 of a window function to generate a left section window function, and generates a right section window function using a 1-left section window function. The processing signal generation unit 22 starts from a predetermined period of time TR from a predetermined reference time TR, a time period TO that dates back to the past, a finite interval of the left interval window function, and a time TB of the received signal that dates back to the past. with the first processing signal S1 by multiplying the time series w _k to the received signal, the reference period from the reference time TR, starting from the time tO of the received signal retroactive, right section window function 1-w _k the received signal Is multiplied by the time series to generate the second processed signal S2. The output unit 23 receives a signal obtained by adding the first processing signal S1 and the second processing signal S2 in time series in a time period TI from the reference time point TR to the right section window function 1-w _k before the finite section. Is output to the timing detector 3.
[Selection] Figure 4

Description

  The present invention relates to a receiving apparatus and a symbol timing detection method for detecting the symbol timing of a multicarrier signal including a preamble portion in which known symbols are repeated.

  In an orthogonal division multiplexing (OFDM) communication system typified by 802.11a, correlation processing using a time-axis waveform pattern of a known symbol is employed to detect symbol timing. FIG. 12 is a frame configuration diagram of an OFDM signal of one packet in 802.11a. In the following description, noise added in the transmission path is ignored for convenience of description.

  As shown in FIG. 12, this OFDM signal includes a preamble part D1, a header part D2, and a data part D3. The preamble part D1 includes a short preamble part D11 and a long preamble part D12. The short preamble section D11 repeatedly stores known symbols called short training symbols (STS). The long preamble part D12 repeatedly stores known symbols called long training symbols (LTS). In 802.11a, STS is repeated 10 times and LTS is repeated 2 times.

  In the conventional receiver, the STS time axis waveform pattern is stored in advance, and the symbol timing is detected by obtaining the correlation between the time axis waveform pattern and the received signal input in time series. .

  FIG. 13 shows a block diagram of a synchronization detection circuit for detecting symbol timing in a conventional receiving apparatus. As shown in FIG. 13, the synchronization detection circuit includes a correlator 101, a known symbol storage unit 102, a peak detection unit 103, a threshold storage unit 104, a timing determination unit 105, a counter 106, and an OFDM demodulation circuit 107.

The processing of the correlator 101 will be described below. Assuming that the STS time-axis waveform pattern is s _k and the received signal is s ′ _k , the correlation value Rxx (l) is expressed by equation (1).

However, “*” indicates a complex conjugate, k indicates the sample number of the sample value in the time domain, N indicates the number of samples of the FFT size of the OFDM signal, and l indicates the number of slides of the STS time axis waveform pattern. . In the case of 802.11a, since the fast Fourier transform with 64 FFT-size samples is assumed, the STS time-axis waveform pattern _sk is expressed as shown in Equation (2).

Here, n indicates the sample number of the sample values in the frequency domain, a _n represents Fourier coefficients.

Furthermore, in 802.11a, since the STS cycle is 1/4 of the FFT size, the STS time-axis waveform pattern _sk is expressed by Equation (3).

Since the received signal s ′ _k is subjected to distortion of the transmission path, the received signal s ′ _k is expressed by Expression (4), where H _4n e ^jθ4n represents the distortion in the frequency domain.

  When Expressions (3) and (4) are substituted into Expression (1), Expression (5) is obtained due to the orthogonality of Fourier transform.

  However, l ′ is a positive integer of 0 or more. In the formula (5), if m ≠ n, the added value for k is 0 regardless of n because of orthogonality. In the equation (5), the correlation value for n is similarly 0 when l ≠ 16l ′.

  Eventually, Rxx (l) has a correlation value in 16 sample periods. The peak detection unit 103 shown in FIG. 13 compares a predetermined threshold value with a correlation value so as not to detect noise, and determines a correlation value exceeding the threshold value as a correlation peak.

  The timing determination unit 105 causes the counter 106 to count the peak period of the correlation peak detected by the peak detection unit 103, and detects the symbol timing according to the count value.

  The OFDM demodulation circuit 107 demodulates the received signal according to the symbol timing detected by the timing determination unit 105.

  Note that Patent Documents 1 and 2 are known as related publicly known documents. In Patent Document 1, when transmitting an information signal by OFDM, in order to accurately demodulate the transmission signal by the receiver, the oscillation frequency of the transmitter and the frequency conversion oscillator of the receiver, and the sampling timing of the transmitter and the receiver There has been disclosed a receiver for the purpose of accurately matching the period of.

Further, Patent Document 2 discloses a digital communication apparatus that can establish frequency synchronization in a short time and reliably even when a pilot symbol without a null subcarrier is used.
JP-A-8-223132 JP 2000-22660 A

By the way, when the received signal s ′ _k includes the interference wave am _k , synchronization detection cannot be performed correctly. Expression (6) is obtained by expanding the interference wave am _k by Fourier series.

Where AM _n represents the complex amplitude of the jamming wave, e ^jΘn represents the complex phase of the jamming wave, and L represents the block size number of the Fourier series expansion. Since disturbance am _k is the sample value in the frequency region other than the frequency index n _0, which corresponds to its frequency f _AM has no formula (6) can be simplified as shown in Equation (7).

Eventually, when such an interference wave am _k exists, the received signal s ′ _k needs to be changed as shown in Expression (8).

The cross-correlation between the received signal s ′ _k + am _k and the STS time axis waveform pattern s _k will be examined. The correlation value Rxx (l) between the received signal s ′ _k + am _k including the interference wave am _k and the time axis waveform pattern s _k is expressed by the equation (9) because of the orthogonality of the Fourier transform.

In Expression (9), the first term shown in the second line from the bottom matches the correlation value in the absence of the interference wave am _k shown in Expression (5), but the second term shown in the first line from the bottom is This results in an error and degrades the symbol timing detection performance. Here, the error shown in the second term, the amplitude component AM _n0 jammer am _k is a major component, disturbance am _k varies depending on the modulation frequency if it is modulated. The error shown in the second term also changes depending on the STS cycle.

As described above, in the presence of the interference wave am _k, a situation occurs in which the correlation value due to the error shown in the second term becomes large, making it difficult to detect the correlation peak and degrading the symbol timing detection performance. . Moreover, in the methods of Patent Documents 1 and 2, there is no description about suppressing the interference wave.

  An object of the present invention is to provide a receiving apparatus and a symbol timing detection method capable of accurately detecting a symbol timing even in the presence of an interference wave.

  (1) A receiving apparatus according to an aspect of the present invention is a receiving apparatus that detects a symbol timing of a multicarrier signal including a preamble portion in which known symbols are repeated, and is based on a reference period that is an integral multiple of the period of the known symbols. A window function processing unit that performs window function processing on a received signal using a window function having a short finite interval, and a correlation between the received signal subjected to the window function processing by the window function processing unit and the time axis waveform pattern of the known symbol A timing detection unit that obtains a peak and detects a symbol timing of the multicarrier signal based on the correlation peak, and the window function processing unit extracts a left half section or a right half section of the window function Based on the extracted window function, a left interval window function that is a window function of the left half interval and a right interval window function that is a window function of the right half interval are generated. A window function generating unit that performs the reference period from a predetermined reference time, a finite section of the left section window function from a time traced back in the past, and the received signal time traced back in the past as a starting point. To the received signal in time series to generate a first processed signal, and starting from the time of the received signal retroactive to the right half of the reference time, the right interval window function is used as the received signal. A processing signal generation unit that multiplies a time series to generate a second processing signal, and a signal obtained by adding the first processing signal and the second processing signal in time series to the past from the reference time point. And an output unit that outputs the received signal up to a finite interval before the right interval window function to the timing detection unit.

  A symbol timing detection method according to another aspect of the present invention is a symbol timing detection method for detecting a symbol timing of a multicarrier signal including a preamble portion in which a known symbol is repeated, and is an integer of a cycle of the known symbol. A window function processing step for performing window function processing on the received signal using a window function having a finite interval shorter than a double reference period, and a time axis waveform of the received signal and the known symbol processed by the window function processing step A timing detection step of obtaining a correlation peak with a pattern and detecting a symbol timing of the multicarrier signal based on the correlation peak, wherein the window function processing step comprises a left half section or a right half of the window function And the left and right interval window functions are generated based on the extracted window function. A window function generation step, a reference period from a predetermined reference time, a finite interval of the left interval window function from a time retroactive to the past, and a reception signal time retroactive to the past as a starting point. The received signal is time-sequentially generated to generate a first processing signal, and the right section window is started from a finite section of the right section window function from the reference time, the time of the received signal going back in the past. A processing signal generating step of generating a second processing signal by multiplying the received signal by a time series with a function; and a signal obtained by adding the first processing signal and the second processing signal in time series to the reference time point And an output step of outputting the received signal as a reception signal before the finite interval of the right interval window function from the past to the past.

  According to these configurations, the left half or right half section of the window function is extracted, and a left section window function and a right section window function are generated based on the extracted window function. Then, from the predetermined reference time point, the reference period, the finite interval of the left interval window function from the time point that went back in the past, the time point of the received signal that goes back in the past, and the left interval window function is time-series to the received signal Is multiplied to generate a first processed signal. Also, the finite interval of the right interval window function from the reference time point, the reception signal time point that goes back in the past, is set as the starting point, and the right interval window function is multiplied by the reception signal in time series to generate the second processing signal.

  Then, a signal obtained by adding the first processing signal and the second processing signal in time series is output as a reception signal from the reference time point to the past and before the finite section of the right section window function.

  As a result, the orthogonality of the received signal is maintained, and at the same time, the interference wave included in the received signal can be suppressed, so that the symbol timing can be detected with high accuracy.

  (2) The window function generation unit extracts the window function of the left half section, generates the extracted window function as the left section window function, and subtracts the left section window function from 1 to reduce the right section window. It is preferable to generate a function.

  According to this configuration, the window function of the left half section is extracted, the extracted window function is generated as the left section window function, and the right section window function is generated by the 1-left section window function. Therefore, it is possible to generate a left section window function and a right section window function having the same finite section by a simple process.

  (3) The window function generation unit extracts the window function of the right half section, generates the extracted window function as the right section window function, and subtracts the right section window function from 1 to reduce the left section window. It is preferable to generate a function.

  According to this configuration, the window function in the right half section is extracted, the extracted window function is generated as the right section window function, and the left section window function is generated by the 1-right section window function. Therefore, it is possible to generate a left section window function and a right section window function having the same finite section by a simple process.

  (4) There are a plurality of window function processing units, each window function processing unit has a different reference period, and the timing detection unit calculates a correlation peak of each received signal output from each window function processing unit. It is preferable to obtain a peak period and detect the symbol timing of the multicarrier signal based on the correlation peak of the received signal whose peak period is closest to the period of the known symbol.

  According to this configuration, a plurality of window function processing units having different reference periods are provided. Therefore, it is possible to detect the symbol timing using the received signal output from the window function processing unit that can suppress the interference wave intensity most, and to detect the symbol timing more accurately.

  (5) It further includes a known symbol storage unit that stores a plurality of types of time axis waveform patterns having different frequency bands generated in advance by combining a plurality of subcarriers used for generating the known symbol, and the timing detection unit includes: The peak period of the correlation peak between the reception signal output from the window function processing unit and each time axis waveform pattern is obtained, and the correlation of the reception signal by the time axis waveform pattern whose peak period is closest to the period of the known symbol It is preferable to detect the symbol timing of the multicarrier signal based on the peak.

  According to this configuration, a correlation peak is obtained using a plurality of time-axis waveform patterns of known symbols having different frequency bands. Therefore, it is possible to obtain a correlation peak using a time axis waveform pattern in a frequency band in which no disturbing wave is placed, and by detecting a symbol timing using the correlation peak obtained using this time axis waveform pattern. The symbol timing can be detected with higher accuracy.

  (6) It is preferable that the timing detection unit detects a correlation value having a level within a predetermined range as a correlation peak, and obtains the peak period using the correlation peak.

  According to this configuration, since a correlation value having a level within a predetermined range is detected as a correlation peak and the peak period is obtained using the correlation peak, the symbol timing can be detected with high accuracy.

  According to the present invention, since it is possible to suppress the interference wave included in the received signal while maintaining the orthogonality of the received signal, it is possible to detect the symbol timing with high accuracy.

(Embodiment 1)
Hereinafter, the receiving apparatus according to Embodiment 1 of the present invention will be described. FIG. 1 shows a block diagram of a receiving apparatus according to the first embodiment. The receiving apparatus illustrated in FIG. 1 is, for example, an 802.11a standard receiving apparatus that receives a multicarrier signal that is obtained by digitally modulating and multiplexing a plurality of narrowband subcarriers having an orthogonal relationship on the frequency axis. It is.

  This multicarrier signal is an OFDM (Orthogonal Frequency Division Multiplexing) signal of the IEEE 802.11a standard, and includes a preamble part D1, a header part D2, and a data part D3 as shown in FIG. . The preamble part D1 is a PLCP (Physical Layer Convergence Protocol) preamble, for example, and stores a symbol for establishing reception synchronization. The header part D2 stores header information. The data part D3 stores data symbols to be transmitted.

  The preamble part D1 includes a short preamble part D11 and a long preamble part D12. The short preamble part D11 stores a short training symbol (STS) that is a known symbol. The long preamble part D12 stores a long training symbol (LTS). STS and LTS are symbols known on the receiving side. The STS is mainly used for symbol synchronization for detecting symbol timing and coarse adjustment of AFC (Automatic Frequency Control). LTS is mainly used for AFC fine adjustment and channel estimation.

  In the 802.11a standard, the short preamble part D11 stores 10 STSs, and the period of one STS is 0.8 μS. The long preamble part D12 stores two LTSs, and the period of one LTS is 3.2 μS. In this embodiment, the symbol timing is detected using 10 STSs.

  The header information includes, for example, the transmission rate and data length of the data symbol stored in the data part D3.

  The data part D3 includes, for example, a predetermined number of OFDM symbols that accommodate the data symbols D32 and a predetermined number of OFDM symbols that accommodate the pilot signal (PS) D33. Each OFDM symbol includes a guard interval (GI) D31 provided at the head. GID31 is a redundant signal provided to avoid intersymbol interference. GID31 is a signal obtained by copying (copying) a predetermined period at the rear end of the data symbol D32 (in the case of PSD33, a predetermined period at the rear end of PSD33).

  For example, in the 802.11a standard, the period of the data symbol D32 is 3.2 μS, and the period of GID31 is 0.8 μS.

  Returning to FIG. 1, the reception apparatus includes a reception unit 1, a window function processing unit 2, a timing detection unit 3, a known symbol storage unit 4, and an OFDM demodulation circuit 5.

  The reception unit 1 receives a reception signal via a predetermined transmission path, performs predetermined analog processing, and then performs analog-digital conversion to generate a digital reception signal (reception vector). In the present embodiment, receiving section 1 samples the received signal so that the number of samples in one cycle of STS is 16, for example. Note that received signals for a certain period in the past are stored in a buffer (not shown).

  As the transmission path, for example, either wired or wireless may be adopted. As the cable, various communication lines, power lines used for power line carrier communication, and the like can be adopted.

  The window function processing unit 2 performs window function processing on the received signal using a predetermined window function.

Here, the window function w _k has a finite section T1 shorter than a reference period that is a positive integer multiple of the period of STS, which is a known symbol, has a symmetrical time-axis waveform, and moves toward the center of the finite section. As the value increases. Note that the maximum value of the window function w _k is 1 or less.

  In this embodiment, the reference period is four times the STS period, and the number of samples is 16 × 4 = 64. This reference period is equal to the FFT (Fast Fourier Transform) size, that is, the OFDM symbol period excluding GI.

In the present embodiment, for example, a Gaussian window is adopted as the window function w _k , but the present invention is not limited to this, but a rectangular window, a Gaussian window, a Hann window, a Hamming window, a Blackman window, a Kaiser window, a Bartlett window, and an exponent window. Various things such as these may be adopted.

FIGS. 2A and 2B are waveform diagrams showing an example of the window function w _k employed in the present embodiment. FIG. 2 (a), the (b), the vertical axis represents the value of the window function w _k, the horizontal axis represents time. In the window functions w _k of FIGS. 2A and 2B, the number of samples in the finite interval T1 is both 14, which is an even number.

  As shown in FIGS. 2A and 2B, the finite section T1 indicates a period from the leftmost sample point PL to the rightmost sample point PR. The left half section T11 is a section defined by the left half sample points in the finite section T1, and the right half section T12 is a section defined by the right half sample points in the finite section T1.

  In FIG. 2A, there are sample points on the vertical axis, and the number of sample points on the left side excluding the sample points on the vertical axis is seven, but the sample on the right side excluding the sample points on the vertical axis. The number of points is 6, and the number of sample points is not symmetrical. Therefore, in the case of FIG. 2 (a), the period consisting of seven sample points from the sample point PL to the nearest sample point on the left side of the vertical axis becomes the left half section T11, and the sample from the sample point on the vertical axis A period composed of seven sample points up to the point PR is a right half section T12. In the case of FIG. 2A, the value of the sample point on the vertical axis is 1.

  On the other hand, in FIG. 2B, there are no sample points on the vertical axis, and the number of sample points on the left side of the vertical axis and the number of sample points on the right side are both 7. The number is equal on the left and right. For this reason, in the case of FIG. 2B, the period consisting of seven sample points from the sample point PL to the leftmost sample point on the left side of the vertical axis is the left half section T11, and A period composed of seven sample points from the sample point to the sample point PR is a right half section T12. In the case of FIG. 2B, since there is no sample point on the vertical axis, the value of the nearest sample point on the left side of the vertical axis and the value of the nearest sample point on the right side of the vertical axis are both 1. Is less than.

  2 (a) and 2 (b), only 14 sample points are shown for convenience, but in the following description, p + 1 sample points exist in the left half section T11, and in the right half section T12. It is assumed that there are p + 1 sample points and a total of 2p + 2 sample points.

  FIG. 3 is a block diagram showing a detailed configuration of the window function processing unit 2. 4A and 4B are explanatory diagrams of window function processing by the window function processing unit 2. FIG. As shown in FIG. 3, the window function processing unit 2 includes a window function generation unit 21, a processing signal generation unit 22, and an output unit 23.

The window function generator 21, generates the left section window function w _k retrieves a section T11 of the left half of the window function w _k, based on the left section window function w _k, right section window function 1-w _k Is generated. Here, the window function generator 21, FIG. 2 (a), the (b), the taken out p + 1 samples points in the left half of the section T11, and generates a left section window function w _k. By _{1-w k} when the value of each sample point and _{w k} of the left section window function _{w k,} to produce a right section window function _{1-w k.}

Accordingly, in this case, the finite sections of the left section window function w _k and the right section window function 1-w _k are both the left half section T11.

Further, as shown in FIG. 4A, the processing signal generation unit 22 receives a retroactive reception of a left interval window function from a predetermined reference time TR, a reference period, and a retroactive time TO from the past. The first processing signal S1 is generated by multiplying the received signal by time series with the left interval window function w _k starting from the time point TB of the signal.

Further, as illustrated in FIG. 4A, the processing signal generation unit 22 uses the right interval window function 1-w _k as the reception signal starting from the time TO of the reception signal traced back to the reference period and the past from the reference time TR. Is multiplied by the time series to generate the second processed signal S2.

As shown in FIG. 4 (b), the output unit 23 traces a signal obtained by adding the first processed signal S1 and the second processed signal S2 in time series from the reference time point TR to the past. output to timing detection section 3 as a reception signal period TI to a finite interval preceding w _k.

  Note that the window function processing unit 2 executes the above processing, for example, every sampling period.

  Returning to FIG. 1, the timing detection unit 3 obtains a correlation peak between the received signal subjected to the window function processing by the window function processing unit 2 and the STS time axis waveform pattern, and based on the correlation peak, the symbol of the multicarrier signal Detect timing.

  Specifically, the timing detection unit 3 includes a correlator 31, a peak detection unit 32, a threshold storage unit 33, a timing determination unit 34, and a counter 35.

  The correlator 31 obtains a correlation value based on the cross-correlation between the received signal subjected to the window function processing and the STS time axis waveform pattern stored in the known symbol storage unit 4, and outputs the correlation value to the peak detection unit 32.

  The peak detector 32 detects the correlation peak of the correlation value output from the correlator 31. Here, when the correlation value output from the correlator 31 is within the range of the predetermined threshold value a and the predetermined threshold value b (b <a) stored in the threshold value storage unit 33, the peak detection unit 32 The correlation value is determined as a correlation peak. This can prevent noise from being detected as a correlation peak.

  Here, the threshold value a is, for example, a received signal transmitted from the nearest transmitter, and a value obtained by adding a certain margin to a correlation value obtained from a received signal on which no interference wave is superimposed is adopted. . Thereby, the correlation value obtained from the received signal to which the energy of the interference wave is added can be invalidated. For the threshold value b, for example, a correlation value obtained from a reception signal that is a reception signal transmitted from the farthest transmitter and has a maximum value that can be separated from the noise signal can be employed. Thereby, the sensitivity of synchronization can be maximized while excluding the influence of noise.

  The timing determination unit 34 causes the counter 35 to count the number of samples of the peak interval of the correlation peak detected by the peak detection unit 32. When the timing determination unit 34 repeatedly detects the peak interval of 16 samples corresponding to one cycle of the STS a predetermined number of times, the timing determination unit 34 detects the detection time of the first correlation peak of this repetition as the symbol timing.

  The known symbol storage unit 4 stores in advance an STS time axis waveform pattern which is a known symbol. This time axis waveform pattern is digital data represented by, for example, 16 sample values. Although the known symbol stores the STS time-axis waveform pattern in advance, the present invention is not limited to this, and a frequency-domain symbol value may be stored in advance. In this case, a known symbol generation unit is provided between the known symbol storage unit 4 and the correlator 31, and the known symbol generation unit converts the symbol value of the frequency domain stored in the known symbol storage unit into an IFFT (inverse Fast Fourier transform). Transform: Inverse fast Fourier transform), a time axis waveform pattern may be calculated and output to the correlator 31.

  The OFDM demodulating circuit 5 performs OFDM demodulation processing on the received signal output from the receiving unit 1 in accordance with the synchronization timing detected by the timing determining unit 34. Here, the OFDM demodulating circuit 5 performs processing such as preamble part removal, FFT, phase correction, demodulation of complex symbols primarily modulated by BPSK, QPSK, and error correction on the received signal. Thereby, the data symbol D32 is restored.

  Next, details of the processing of the receiving apparatus shown in FIG. 1 will be described. First, a case where the window function processing unit 2 performs normal window function processing using a window function having a 64-sample period different from the above description will be described.

In the window function processing, generally, a time window function w _k the same number of samples and the FFT size of the multicarrier signal of interest may be held in advance calculated and, the received signal with the window function w _k for each sample An inner product process for multiplying the series is performed. Assuming that the received signal subjected to the inner product processing is s ′ _k and the transfer function of the transmission path is 1 for convenience of explanation, s ′ _k is expressed by Expression (10).

In this case, the time-axis waveform pattern _{s k} of STS, the correlation value between the received signal window function processing by the equation (10) Rxx (l) is expressed by equation (11).

Here, since the received signal is assumed to be an OFDM signal having 64 FFT-size samples, the window function w _k has 64 sample points. Further, since the number of sample points is 64 and an even number, the window function w _k is a function symmetric with respect to the time axis between the 31st sample and the 32nd sample. Therefore, w _k = w _63−k holds.

  FIG. 5A is a conceptual diagram when adding received signals not subjected to window function processing for the FFT size. FIG. 5B is a conceptual diagram when the received signals subjected to the window function processing are added for the FFT size. Each point plotted in FIGS. 5A and 5B indicates the sample value of the received signal, the vertical axis indicates the imaginary axis, and the horizontal axis indicates the real axis.

  As shown in FIG. 5A, when the received signal that has not been subjected to the window function processing is added by the FFT size due to the orthogonality of the OFDM signal, the addition value (complement) becomes zero. Further, according to equation (10), the denominator of the exponent of the exponential function is 16. The number of FFT size samples of the received signal is 64, where k = 0 to 63. Therefore, when the received signal is added by the FFT size, the received signal rotates four times on the complex plane as shown in FIG.

  On the other hand, the received signal on which the window function processing has been performed changes in a spiral shape with the amplitude gradually increasing from the 0th sample to the 31st sample as shown by the black dots in FIG. Then, as shown by the white dots in FIG. 5B, the center 32 sample starts to decrease, and the amplitude gradually decreases and changes spirally.

  That is, when the window function processing is performed, the imaginary number component of the received signal has symmetry as shown in FIG. 5B and cancels when the FFT size is added, and the added value becomes almost zero. Maintains orthogonality. On the other hand, since the real number component does not have symmetry as shown in FIG. 5B, even if the FFT size is added, the added value does not become 0 and the orthogonality is broken.

  Therefore, when the window function process is performed on the received signal, the orthogonality is lost and the subcarriers interfere with each other, and the demodulated OFDM symbol includes many errors and cannot be accurately demodulated.

Therefore, in the present embodiment, the window function processing shown in FIGS. 4A and 4B described above is performed. Hereinafter, the window function processing shown in FIGS. 4A and 4B will be described in detail. In the following description, in FIGS. 4A and 4B, the number of samples in the reference period is 64, and p = 31. Therefore, in the window function w _k of FIGS. 2A and 2B, the number of samples in the finite interval T1 is 2p + 2 = 64. The reference time point TR is the 64th sample of the i-th reference period. Since the window function w _k has 64 samples and is an even number, assuming that the sample point PL is the 0th sample and the sample point PR is the 63rd sample, the window function w _k is almost symmetrical between the 31st and 32nd samples. Function.

First, as shown in FIGS. 2A and 2B, the window function generation unit 21 extracts the sample points of the left half section T11 from the sample points of the window function w _k and outputs the left section window function w. to generate a _k, a 1-w _k When each sample point and w _k of the left section window function to generate a right section window function 1-w _k.

  Next, the processed signal generation unit 22 extracts p + 1 (= 32) sample points from the −p (= −31) sample to the 0th sample of the received signal as shown in Expression (12).

Then, the processed signal generator 22, as shown in equation (13), the p + 1 pieces of sample points left section window function w _k from 0-th sample of the left section window function w _k to p th sample k = set as -P～0, the left section window function w _k set to act on the p + 1 samples points from -p th sample of the received signal to zero-th sample.

  Next, the processing signal generation unit 22 extracts p + 1 (= 32) sample points from the 64-p (= 33) sample to the 64th sample of the received signal as shown in Expression (12).

Next, as shown in Expression (14), the processing signal generation unit 22 determines p + 1 sample points from the 0th sample to the pth sample of the right interval window function 1-w _k as the right interval window function 1-w. _k is set as k = 64-p to 64, and the set right interval window function 1-w _{k is applied} to p + 1 sample points from the 64-pth sample to the 64th sample of the received signal.

Then, the output unit 23, as shown in equation (15), and _s'k of the formula (13) (k = -p~0) , s'k of formula (14) (k = 64- p~ 64) are added and output as sample points from the k = 64-pth sample to the 64th sample of the received signal.

Received signal _s'k obtained by the equation (15), since the term of _{w k} is not present, the received signal _s'k of the FFT size of the respect to the origin as shown in FIG. 5 (a) Arranged point-symmetrically. As a result, the received signal s ′ _k can maintain orthogonality by this window function processing.

Here, as the range of k, k = −p to 64 was considered, but k = −p to 16q (q is an integer value of 1 or more and p + 1 ≦ 16q / 2) (however, the size of w _k is The same holds true for 2p + 2).

Thus, by employing the window function processing described above, it is possible to maintain the orthogonality of the received signal _s'k, it is possible to take out the OFDM symbols accurately.

Next, consider the case where the window function processing of the present embodiment is applied to the interference wave am _k of the received signal s′k. In the following description, only k = −1, 63, which are complementary sample points, will be considered for convenience of description. First, when the number of FFT-sized samples is 64 and the interference wave am _k is expanded in the Fourier series, Equation (16) is obtained for the interference wave.

Here, ε _k represents an error between the interference wave expanded in the Fourier series and the actual interference wave am _k .

When the window function w _k is applied to the disturbance wave am _k at the sample points of k = −1, 63, the equation (17) is obtained.

These w _-1 am _-1 and w ₆₃ am ₆₃ are added as shown in Expression (18).

As shown in Expression (18), w ₋₁ is related to the term of ε ₋₁ −ε ₋₆₃ . Here, w ₋₁ satisfies 0 <w ₋₁ ≦ 1. Therefore, by the above window function processing, the influence of ε ₋₁ -ε ₋₆₃ can be reduced by w ₋₁ times, and the interference wave am _k can be suppressed.

It should be noted that the other complementary sample points k = −2, 62, k = −3, 61, k = −4, 60,... _The interference wave am _k can be suppressed by _k .

  As described above, according to the receiving apparatus according to the first embodiment, it is possible to suppress the interference wave and to accurately detect the symbol timing without breaking the orthogonality of the reception signal s′k.

In the above description, the window function generator 21 retrieves the segment T11 of the left half, but generates a left section window function w _k, the present invention is not limited thereto. That is, the window function generator 21, window function w _k retrieves the right half of the section T12 of generating the right section window function w _k, based on the right section window function w _k, the left section window function 1-w _k may be generated.

Then, the window function generator 21, as shown in FIG. 2 (a), (b) , taken out p + 1 samples points in the right half of the interval T12, and generates a right section window function _{w k.} When the value of each sample point in the right section window function and w _k by 1-w _k, may be generated left section window function 1-w _k. In this case, the finite sections of the right section window function w _k and the left section window function 1-w _k are both the right half section T12.

  In the above description, the first processing signal S1 is generated first, and then the second processing signal is generated. However, the present invention is not limited to this. First, the second processing signal S2 is generated. Thereafter, the first processing signal may be generated. These aspects are also applicable to the following embodiments.

(Embodiment 2)
The receiving apparatus according to the second embodiment is characterized in that a plurality of window function processing units 2 having different reference periods are provided. In the present embodiment, the same elements as those in the first embodiment are not described.

  FIG. 6 shows a block diagram of a receiving apparatus according to the second embodiment. As illustrated in FIG. 6, the reception device includes three window function processing units 201 to 203.

In the window function processing unit 201, the number of samples in the reference period and the number of samples in the finite section T1 of the window function w _k are 64, which is three times the STS. Window function processing unit 202, the number of samples in a finite interval T1 of sample number and window function w _k of the reference period is 32 twice the STS. Window function processing unit 203, the number of samples in a finite interval T1 of sample number and window function w _k of the reference period is 16 1 times the STS. Note that details of the processing of the window function processing units 201 to 203 are the same as those of the window function processing unit 2 of the first embodiment, and thus the description thereof is omitted.

  The timing detection unit 3 obtains the correlation peak period of each reception signal output from the window function processing units 201 to 203, and based on the correlation peak of the reception signal whose correlation peak period is closest to the STS period, the multicarrier Detect the symbol timing of the signal.

  Specifically, the timing detection unit 3 includes three correlators 311 to 313 corresponding to the window function processing units 201 to 203, peak detection units 321 to 323, and peak counters 351 to 353. The timing detection unit 3 includes a threshold value storage unit 33 and a timing determination unit 34.

  Correlators 311 to 313 calculate correlation values between the reception signals output from window function processing sections 201 to 203 and the STS time-axis waveform pattern, and output the correlation values to peak detection sections 321 to 323.

  The peak detectors 321 to 323 detect correlation peaks of the correlation values output from the correlators 311 to 313, respectively. Note that the correlation peak detection method by the peak detection units 321 to 323 is the same as that of the peak detection unit 32 of the first embodiment, and thus detailed description thereof is omitted.

  The peak counters 351 to 353 count the number of samples at the peak interval of the correlation peak detected by the peak detectors 321 to 323, respectively.

  When any of the peak counters 351 to 353 detects a peak interval of 16 samples corresponding to one cycle of STS, the timing determination unit 34 sets the peak counter as a target peak counter, and a correlator corresponding to the target peak counter. For example, an interval of 32 samples from the detection time of the first correlation peak is set as an effective interval.

  Then, when the peak counter of interest can repeatedly count the peak interval of 16 samples within the effective interval, the timing determination unit 34 detects the first correlation peak of the correlator corresponding to the peak counter of interest. Are detected as symbol timing. Note that the timing determination unit 34 may determine that the peak interval is valid even if the count value of the peak interval by the target peak counter is not 16, but is within a range of 16 ± 1, for example.

  FIGS. 7A to 7D are graphs showing calculation results of correlation values by the correlator, and FIG. 7A is a graph showing STS time-axis waveform patterns for received signals that have not been subjected to window function processing. (B) to (d) show the correlation values calculated by the correlators 311 to 313 shown in FIG. 6, respectively. 7A to 7D, the vertical axis indicates the correlation value, and the horizontal axis indicates the time as the number of samples. 7A to 7D, the threshold value indicates the threshold value b.

  In FIG. 7A, since the peak interval from when the correlation peak is first detected until the next correlation peak is detected is not 16 samples, the symbol timing cannot be detected.

  In FIG. 7B, since the number of samples from when the peak detection unit 321 first detects a correlation peak to when the next correlation peak is detected is 16, 32 samples from the time of detection of the first correlation peak. An effective interval of minutes is set. In this effective section, the peak interval of 16 samples is repeated twice. Therefore, the timing determination unit 34 detects the time point when the correlator 311 first detects the correlation peak as the symbol timing. That is, the symbol timing has been successfully detected.

  In FIG. 7C, the symbol timing cannot be detected as in FIG. 7A. In FIG. 7D, as in FIG. 7B, the symbol timing has been successfully detected.

In the above equation (18), if ε ₋₆₃ + w ₋₁ (ε ₋₁ −ε ₋₆₃ ) = 0, the effect of the window function processing is maximized, and the interference wave am _k can be most suppressed. In order for this to hold, ε ₋₁ and ε ₋₆₃ need to satisfy equation (19).

However, since ε _k depends on the phase and frequency of the jamming wave am _k , the relationship of Expression (19) does not always hold. Therefore, in the second embodiment, three window function processing units 201 to 203 having different reference periods are provided. As a result, any one of the peak detectors 321 to 323 can detect the correlation peak with high accuracy, more reliably detect the symbol timing, and enhance the effect of the window function processing. be able to.

(Embodiment 3)
The receiving apparatus according to the third embodiment is characterized in that two known symbol storage units are provided in the receiving apparatus according to the second embodiment. FIG. 8 shows a block diagram of a receiving apparatus according to the third embodiment.

  Each of the known symbol storage units 41 and 42 stores a plurality of types of time-axis waveform patterns having different frequency bands generated in advance by combining a plurality of subcarriers used for generating an STS. In 802.11a, an STS time-axis waveform pattern is generated using 12 subcarriers out of 64 subcarriers.

  FIGS. 9A to 9C show the power spectrum of the subcarriers used for generating the STS time-axis waveform pattern. FIG. 9A shows subcarriers used for generating a conventional STS time axis waveform pattern, and FIG. 9B shows an STS time axis waveform pattern stored in the known symbol storage unit 41. FIG. 9C shows the subcarriers used for generating the STS time axis waveform pattern stored in the known symbol storage unit 42.

  As shown in FIG. 9A, the conventional STS time-axis waveform pattern is generated using 12 subcarriers. Therefore, when an interfering wave is present in the frequency band of any one of the 12 subcarriers, the symbol timing cannot be accurately detected.

  Therefore, as shown in FIG. 9 (b), the known symbol storage unit 41 has subbands in a total of 6 both-end frequency bands, 3 from the left and 3 from the right of the 12 subcarriers in FIG. 9 (a). An STS time axis waveform pattern generated using a carrier is stored in advance. Further, as shown in FIG. 9 (c), the known symbol storage unit 42 uses six central frequency band subcarriers from the fourth to the ninth from the left of the 12 subcarriers in FIG. 9 (a). The STS time axis waveform pattern generated in advance is stored in advance.

  Returning to FIG. 8, the correlators 311 to 313 obtain correlation values between the reception signals output from the window function processing units 201 to 203 and the STS time-axis waveform pattern stored in the known symbol storage unit 41, respectively. .

  Correlators 314 to 316 obtain correlation values between the reception signals output from window function processing sections 201 to 203 and the STS time axis waveform pattern stored in known symbol storage section 42, respectively.

  The peak detectors 321 to 323 detect correlation peaks of the correlation values output from the correlators 311 to 313, respectively. The peak detectors 324 to 326 detect correlation peaks of the correlation values output from the correlators 314 to 316, respectively. Since the correlation peak detection method by the peak detection units 321 to 326 is the same as that of the peak detection unit 32 of the first embodiment, detailed description thereof is omitted.

  The peak counters 351 to 356 count the number of samples at the peak interval of the correlation peak detected by the peak detectors 321 to 323, respectively.

  The timing determination unit 34 detects the symbol timing using the same method as in the second embodiment, according to the peak intervals counted by the peak counters 351 to 356, respectively. That is, the timing determination unit 34 determines the symbol timing of the multicarrier signal based on the correlation peak of the received signal according to the time axis waveform pattern in which the peak period of the correlation peak counted by the peak counters 351 to 356 is closest to the STS period. To detect.

  The threshold storage unit 33 outputs the threshold value a and the threshold value b to each of the peak detection units 321 to 326.

Next, processing of correlators 311 to 316 according to the present embodiment will be described using mathematical expressions. First, it is assumed that the interference wave am _k is made so small that the error ε _k can be ignored by the window function processing by the window function processing units 201 to 203 and can be approximated as shown in Expression (20).

Here, it is assumed that the disturbance am _k is present in the low-frequency side of the both ends frequency band of n = 0 to 15, disturbance am _k of n = 16-63 is negligible. In this case, the interference wave am _k can be approximated by the equation (21).

  Also, the STS time-axis waveform pattern stored in the known symbol storage unit 41 is expressed by Expression (22). However, the sampling interval of m is four times the sampling interval of n.

  On the other hand, the STS time-axis waveform pattern stored in the known symbol storage unit 42 is expressed by Expression (23).

The correlation value Rxx (l) between the STS time axis waveform pattern of the equation (22) in which the interference wave am _k is not placed and the received signal s ′ _k + am _k including the interference wave am _k is expressed by the equation (24). The Since it is assumed that the interference wave is on the lower side of the frequency bands at both ends of n = 0 to 15 and not on n = 16 to 63, it is shown in the first line of equation (24). s ′ _k + am _k is am _k = 0 when n = 16 to 63. Therefore, in equation (24), calculations of n = 16 to 63 are omitted.

  Here, the second term in the second row from the bottom of Expression (24) is 0 because n = 0 to 15, m = 4 to 11, and there is no condition for n = 4m.

From this, when the interference wave am _k is placed on the low frequency side of both end frequency bands of n = 0 to 15, the time axis waveform pattern of the STS stored in the known symbol storage unit 42 is the time in the center frequency band. Since this is an axial waveform pattern, the interference wave am _k can be reduced to zero by performing correlation processing using this time axis waveform pattern. Therefore, the correlators 314 to 316 shown in FIG. 8 can obtain the correlation value with higher accuracy than the correlators 311 to 313.

  On the other hand, when the interference wave is in the central frequency band of n = 15 to 48, the STS time axis waveform pattern stored in the known symbol storage unit 41 is the time axis waveform pattern of both end frequency bands. The correlators 311 to 313 shown in FIG. 8 that perform correlation processing using this time axis waveform pattern can obtain the correlation value more accurately than the correlators 314 to 316.

  FIG. 10 shows a correlation result of a received signal on which an AM-modulated interference wave having a carrier frequency of 3.4 MHz is superimposed. FIG. 11 shows a correlation result in a signal vector on which an AM-modulated interference wave having a carrier wave of 20.2 MHz is superimposed.

  10 and 11, the first column shows the correlation results when the STS time axis waveform pattern generated by the subcarriers of FIG. 9A is used, and the second column shows the sub results of FIG. 9C. FIG. 9B shows the correlation results when the STS time axis waveform pattern generated by the carrier (subcarrier in the central frequency band) is used, and the third column shows the subcarriers in FIG. 9B (subcarriers in both frequency bands). The correlation result at the time of using the time-axis waveform pattern of the produced | generated STS is shown.

  10 and 11, the first row shows the correlation results when the window function processing is not performed, and the second to fourth rows show the correlations when the number of samples in the reference period is 64, 32, and 16, respectively. Results are shown.

  That is, in FIGS. 10 and 11, the correlation results in the second column to the second to fourth rows indicate the correlation results by the correlators 314 to 316 shown in FIG. 8, respectively. The correlation results show the correlation results by the correlators 311 to 313 shown in FIG.

  In FIG. 10, the frequency of the jamming wave is 3.4 MHz, and the frequency band of the jamming wave belongs to the lower frequency side of both end frequency bands of n = 0 to 15, and therefore the second row 2 surrounded by a square frame In the fourth to fourth lines, a good correlation result is obtained, and it can be seen that the symbol timing has been successfully detected.

  In FIG. 11, since the frequency of the disturbing wave is 20.2 MHz and the frequency band of the disturbing wave belongs to the central frequency band of n = 16 to 48, in the 2nd to 4th rows of the third column surrounded by a square, It can be seen that a good correlation result is obtained and the symbol timing is successfully detected.

  As described above, according to the receiving apparatus according to the present embodiment, for the received signal that has been subjected to the window function processing by the three window functions each having a different number of samples, the STS time axis waveform pattern in the central frequency band, Since the correlation value with the time axis waveform pattern of the STS in both frequency bands is obtained and the symbol timing is detected using the correlation value with which the peak interval can be detected most accurately among these correlation values, the symbol timing is It is possible to detect with higher accuracy.

  In the third embodiment, two types of STS time-axis waveform patterns are used, but three or more types may be used. In this case, a time axis waveform pattern may be generated by combining subcarriers having different frequencies.

  In the second and third embodiments, three types of window function processing units are used, but the present invention is not limited to this, and four or more types may be used. In this case, the reference period of each window function processing unit may be set to a different value that is an integral multiple of one cycle of STS.

1 is a block diagram of a receiving device according to Embodiment 1. FIG. It is a wave form diagram which shows an example of a window function. It is a block diagram which shows the detailed structure of a window function process part. It is explanatory drawing of the window function process by a window function process part. (A) is a conceptual diagram when adding reception signals not subjected to window function processing for the FFT size, and (b) is for adding reception signals subjected to window function processing for the FFT size. It is a conceptual diagram. FIG. 6 shows a block diagram of a receiving apparatus according to a second embodiment. It is a graph which shows the calculation result of the correlation value by a correlator. FIG. 9 is a block diagram of a receiving apparatus according to Embodiment 3. The power spectrum of the subcarrier used for generation | occurrence | production of STS is shown. The correlation result in the received signal on which the AM-modulated interference wave having a carrier wave of 3.4 MHz is superimposed is shown. The correlation result in the signal vector on which the AM-modulated interference wave having a carrier wave of 20.2 MHz is superimposed is shown. It is a frame block diagram of the multicarrier signal of 1 packet in 802.11a. The block diagram of the synchronous detection circuit which detects the symbol timing in the conventional receiver is shown.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Reception part 2,201-203 Window function processing part 3 Timing detection part 4 Known symbol memory | storage part 5 OFDM demodulation circuit 21 Window function generation part 22 Processing signal generation part 23 Output part 31,311-316 Correlator 32,321-326 Peak detection unit 33 Threshold storage unit 34 Timing determination unit 35 Counters 4, 41, 42 Known symbol storage units 351 to 356 Peak counter

Claims (7)

A receiving apparatus for detecting symbol timing of a multicarrier signal including a preamble portion in which known symbols are repeated,
A window function processing unit that performs window function processing on a received signal using a window function having a finite interval shorter than a reference period that is an integer multiple of the period of the known symbol;
A timing detection unit that obtains a correlation peak between the received signal subjected to the window function processing by the window function processing unit and the time-axis waveform pattern of the known symbol, and detects the symbol timing of the multicarrier signal based on the correlation peak; With
The window function processing unit
Taking out the left half section or the right half section of the window function, and based on the extracted window function, a window function generating unit that generates a left section window function and a right section window function;
From the predetermined reference time to the reference period, from the time point traced back to the past, the finite interval of the left interval window function, the time point of the received signal retroactive to the past, and the time interval of the left interval window function to the received signal To generate a first processing signal, and the right interval window function is used as the reception signal starting from the finite interval of the right interval window function from the reference time, and the time of the received signal retroactive to the past. A processing signal generator that multiplies the sequence to generate a second processing signal;
A signal obtained by adding the first processed signal and the second processed signal in time series is output to the timing detection unit as a received signal from the reference time point to the past and before the finite interval of the right interval window function. And a receiving device.   The window function generation unit extracts the window function of the left half section, generates the extracted window function as the left section window function, and generates the right section window function by subtracting the left section window function from 1 The receiving apparatus according to claim 1, wherein:   The window function generation unit extracts the window function of the right half interval, generates the extracted window function as the right interval window function, and generates the left interval window function by subtracting the right interval window function from 1 The receiving apparatus according to claim 1, wherein: There are a plurality of window function processing units,
Each window function processing unit has a different reference period,
The timing detection unit obtains a peak period of a correlation peak of each received signal output from each window function processing unit, and based on the correlation peak of the received signal whose peak period is closest to the period of the known symbol, The receiving apparatus according to claim 1, wherein symbol timing of a carrier signal is detected. A known symbol storage unit for storing a plurality of types of time-axis waveform patterns having different frequency bands generated in advance by combining a plurality of subcarriers used for generating the known symbol;
The timing detection unit obtains a peak period of a correlation peak between the reception signal output from the window function processing unit and each time axis waveform pattern, and the peak period is based on a time axis waveform pattern closest to the period of the known symbol. The receiving apparatus according to claim 1, wherein a symbol timing of the multicarrier signal is detected based on a correlation peak of the received signal.   The reception according to claim 1, wherein the timing detection unit detects a correlation value having a level within a predetermined range as a correlation peak, and obtains the peak period using the correlation peak. apparatus. A symbol timing detection method for detecting a symbol timing of a multicarrier signal including a preamble portion in which known symbols are repeated,
A window function processing step for performing window function processing on a received signal using a window function having a finite interval shorter than a reference period that is an integral multiple of the period of the known symbol;
A timing detection step of obtaining a correlation peak between the reception signal processed by the window function processing step and the time-axis waveform pattern of the known symbol, and detecting a symbol timing of the multicarrier signal based on the correlation peak. ,
The window function processing step includes:
A window function generating step of extracting a left half section or a right half section of the window function and generating a left section window function and a right section window function based on the extracted window function;
From the predetermined reference time to the reference period, from the time point traced back to the past, the finite interval of the left interval window function, the time point of the received signal retroactive to the past, and the time interval of the left interval window function to the received signal To generate a first processing signal, and the right interval window function is used as the reception signal starting from the finite interval of the right interval window function from the reference time, and the time of the received signal retroactive to the past. A processing signal generation step of multiplying the sequence to generate a second processing signal;
An output step of outputting a signal obtained by adding the first processed signal and the second processed signal in time series as a received signal retroactively from the reference time point to the finite interval of the right interval window function; A symbol timing detection method comprising:

Images