JP6643770B2 - Frequency offset estimation device and wireless communication device - Google Patents

Description

  The present invention relates to a wireless communication device, and more particularly, to a technology for detecting a frequency offset for carrier recovery.

  In a wireless communication device, it is known that there is a deviation between the frequency on the transmitting device side and the frequency on the receiving device side, and the bit error rate (BER) characteristic is deteriorated by the influence of the frequency deviation. Therefore, the receiving side demodulator removes the frequency deviation between transmission and reception by an automatic frequency control (AFC) circuit or the like. Similarly, there is a phase error or a phase shift between transmission and reception, and the demodulator removes the phase error by a carrier recovery (CR) circuit or the like.

  The AFC circuit forms an AFC loop including a phase rotator, a frequency discriminator, and a voltage-controlled oscillator (local oscillator). The frequency discriminating circuit obtains a frequency deviation component from the received modulated signal that has been frequency-converted to the baseband band. A voltage controlled oscillator (VCO) generates an AFC reference signal having a frequency corresponding to the frequency deviation component, and a phase rotator performs a phase rotation corresponding to the frequency deviation from the received modulation signal using the AFC reference signal. It is assumed that the frequency deviation component is removed.

  In such an AFC circuit, Patent Document 1 discloses a technique capable of detecting a synchronous / asynchronous state of the circuit and automatically establishing synchronization when the circuit is in an asynchronous state.

  Further, in Patent Document 2, in a wireless communication such as a satellite communication or a mobile communication, in a demodulation circuit for realizing synchronous detection, an AFC circuit and a CR circuit are made to converge in a state where a phase error remains. By detecting the residual phase error by correlating the received modulation signal with the known pattern by the phase uncertainty elimination circuit and correcting the detected phase error, the adjustment parameter for the feedback loop is reduced, A technique for simplifying a feedback loop is disclosed.

  On the other hand, in recent years, with the widespread use of wireless communication systems, shortages of frequency resources have become apparent mainly in the microwave band, and transmission technologies for achieving high frequency use efficiency have been demanded. The orthogonal polarization multiplexing technique focuses on the wavefront direction of a radio wave radiated from an antenna and transmits independent signals having mutually orthogonal wavefronts at the same frequency.

  When this orthogonal polarization multiplexing technology is applied, in the case of linear polarization used in fixed wireless communication or the like, V and H polarization multiplexing using vertical (V) polarization and horizontal (H) polarization can be realized. In this case, the frequency use efficiency is doubled as compared with the case where the orthogonal polarization multiplexing technique is not applied. The V and H polarization multiplexed signals can be transmitted and received by, for example, arranging two linear radiating elements orthogonally in a cross shape.

  Further, in satellite communication, a polarization multiplexing method is known in which a signal is multiplexed on three or more polarizations that are not always orthogonal to improve frequency use efficiency (for example, see Non-Patent Document 1). In this method, a bit is allocated to each polarization, and a signal point is multiplexed (combined) by mapping a value obtained by mapping each polarization component before combination on the horizontal (H) / vertical (V) polarization IQ plane. Form. At this time, signal point arrangement is optimized so that the minimum Euclidean distance between symbols on each H / V polarization is maximized.

  In such satellite communication, reception is generally performed in a state where the S / N ratio is low. Therefore, a frequency error (frequency offset) between the transmitting side and the receiving side is detected to remove the frequency deviation (frequency reproduction). In such a circuit, high accuracy is required even in such a situation. In order to respond to such a demand, Non-Patent Document 2 discloses a case where the interval D between pilot symbols of a unique word is variable when detecting a frequency offset by correlation of unique words of a predetermined pattern inserted into a transmission signal. The results of evaluation of the bit error rate (BER) of the above are disclosed.

JP-A-2006-217054 JP 2009-267714 A

Night Ship, Webber, Yano, Ban, Kobayashi, "Proposal of Multi-Polarization Spatial Multiplexing Technology in Satellite Communication", IEICE Technical Report, Vol. 112, No. 191, pp. 1-5, Aug. 2012. J. Webber, M. Yofune, K. Yano, H. Ban, and K. Kobayashi, “Performance of frequency recovery algorithms for a poly-polarization multiplexing satellite system”, 11th IEEE Malaysia International Conference on Communications (MICC2013), 27- 29th Nov. 2013.

  Generally, if the interval D between the pilot symbols of the unique word is increased, the frequency offset can be estimated with a smaller number of frames (shorter acquisition period), but in this case, the range of the frequency offset that can be evaluated is narrowed. On the other hand, to detect a coarse frequency offset, it is necessary to reduce the interval D between the pilot symbols of the unique word. However, reducing the interval D between the pilot symbols of the unique word causes an increase in the number of frames required for estimating the frequency offset.

  Here, the maximum allowable interval D is limited by an unknown frequency offset. For this reason, in a general communication environment, it is difficult to know in advance how to set the interval D between the symbols of the unique word.

  SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and an object of the present invention is to provide a demodulation circuit for realizing synchronous detection, which adaptively optimizes the performance of offset estimation and corrects frequency offset accurately. It is an object of the present invention to provide a frequency offset estimating apparatus capable of estimating and capturing a carrier, and a wireless communication apparatus using the same.

  According to one aspect of the present invention, there is provided a frequency offset estimating apparatus for estimating a frequency offset between a carrier frequency of a received signal and a local oscillation frequency, wherein each of the received signals includes a plurality of unique words. A plurality of frames, including an error detection code, comprising a correlator for calculating the correlation between the received signal and the unique word, the correlator can calculate the correlation with a variable correlation length, An offset calculator for integrating the calculated correlations over the number of frames of the set frame average size and calculating an estimated value of the frequency offset, and an error detection code for detecting a transmission error by using an error detection code. And a control unit for controlling the correlator, the offset calculation unit and the error detection unit, the control unit comprising: Receiving on the receiving side, and sequentially performing the error detection by sequentially changing the correlation length among a set of predetermined values. Ii) According to the result of the error detection, the transmission error is equal to or less than a predetermined value. The correlation value used for estimating the frequency offset is specified within the range of.

Preferably, the error detection code is provided over a plurality of frames, the error detection unit, and integrating the error detection results for the plurality of frames, and calculates the error information, the control unit, based on the error information, The maximum correlation length is specified among the correlation lengths corresponding to the frame with the minimum transmission error.

Preferably, the error detection code is provided over a plurality of frames, the error detection unit, and integrating the error detection results for the plurality of frames, and calculates the error information, the control unit, based on the error information, The minimum correlation length is specified from among the correlation lengths corresponding to the frame with the minimum transmission error.

  Preferably, the apparatus further includes a storage device for storing a correlation value used for calculating a frequency offset, wherein the control unit determines a correlation length of each of the number of received frames having the average size among a set of predetermined values. The correlation values calculated by sequentially changing the correlation values are stored in the storage device, and the offset calculation unit calculates the estimated value of the frequency offset based on the correlation values stored in the storage device.

  According to another aspect of the present invention, there is provided a wireless communication apparatus for receiving a wireless signal including a plurality of frames each including a plurality of unique words and an error detection code, the wireless communication apparatus comprising: Receiving means for estimating a frequency offset between the carrier frequency of the received signal from the receiving means and the local oscillation frequency, and correcting the frequency error of the received signal based on the estimated frequency offset. Regenerating means, the carrier frequency reproducing means includes frequency offset estimating means for estimating a frequency offset, and the frequency offset estimating means includes a correlator for calculating a correlation between the received signal and a unique word. , The correlator can calculate the correlation with a variable correlation length, and calculate the calculated correlation with the set frame average size. Integrating over a number of frames, an offset calculation unit for calculating an estimated value of the frequency offset, and an error detection code, further including an error detection unit for detecting a transmission error, a correlator, an offset calculation unit and A control unit for controlling the error detection unit, the control unit i) receiving a signal transmitted at a predetermined unique word interval on a reception side and setting a correlation length in a set of predetermined values. The error detection is executed by sequentially changing the values. Ii) According to the result of the error detection, a correlation value used for estimating the frequency offset is specified within a range in which the transmission error is equal to or less than a predetermined value, and the correlation value is corrected by the carrier frequency recovery means. Decoding means for performing a decoding process on the decoded signal.

Preferably, the error detection code is provided over a plurality of frames, the error detection unit, and integrating the error detection results for the plurality of frames, and calculates the error information, the control unit, based on the error information, The maximum correlation length is specified among the correlation lengths corresponding to the frame with the minimum transmission error.

Preferably, the error detection code is provided over a plurality of frames, the error detection unit, and integrating the error detection results for the plurality of frames, and calculates the error information, the control unit, based on the error information, The minimum correlation length is specified from among the correlation lengths corresponding to the frame with the minimum transmission error.

  Preferably, the apparatus further includes a storage device for storing a correlation value used for calculating a frequency offset, wherein the control unit determines a correlation length of each of the number of received frames having the average size among a set of predetermined values. The correlation values calculated by sequentially changing the correlation values are stored in the storage device, and the offset calculation unit calculates the estimated value of the frequency offset based on the correlation values stored in the storage device.

  According to the present invention, in a demodulation circuit for realizing synchronous detection, by selecting an optimum correlation length N according to a propagation path state, it is possible to adaptively and accurately estimate a frequency offset and capture a carrier. It becomes possible.

FIG. 3 is a functional block diagram for describing a configuration of transmitter 1000 according to the present embodiment. FIG. 3 is a functional block diagram for describing a configuration of receiver 2000 of the present embodiment. FIG. 9 is a diagram for explaining the arrangement of symbols in a unique word. FIG. 3 is a functional block diagram for explaining a configuration of a frequency offset estimating device 2070 in a carrier frequency reproducing unit 207 and a CRC decoding processing unit 216. FIG. 9 is a diagram illustrating a simulation result of a relationship between a correlation length N and a maximum frequency offset that can be evaluated. It is a figure showing an example of different arrangement of a unique word in unique word interval D and correlation length N. FIG. 9 is a diagram for explaining a mode in which unique words arranged at different unique word intervals D are arranged in a frame of a transmission signal. FIG. 3 is a diagram for explaining the arrangement of unique words and CRC codes transmitted from a transmission side to a reception side. It is a conceptual diagram which shows the evaluation of the adjustable maximum frequency offset at the time of changing the correlation length N adaptively. FIG. 10 is a diagram illustrating a simulation result of a relationship between a bit energy to noise power density ratio Eb / No and a bit error rate BER. FIG. 10 is another diagram showing a simulation result of a relationship between a bit energy-to-noise power density ratio Eb / No and a bit error rate BER. It is a flowchart for demonstrating the processing flow of an adaptive correlation length. FIG. 7 is a diagram illustrating a relationship between the number of received frames and a remaining frequency offset. FIG. 4 is a conceptual diagram illustrating an example of a correlation value stored in a storage device.

  Hereinafter, a wireless communication system according to an embodiment of the present invention will be described with reference to the drawings. In the following embodiments, components and processing steps denoted by the same reference numerals are the same or equivalent, and the description thereof will not be repeated unless necessary.

  As described below, the wireless communication system of the present embodiment employs a polarization multiplexing system.

  Preferably, in the present embodiment, in a communication device in a system for transmitting information using two polarizations, a state used in a system having no obstruction between transmission and reception such as satellite communication is more preferable. is there. Note that a communication device having only the transmission function may have only the transmission function of the present embodiment, and a communication device having only the reception function may have only the reception function of the present embodiment. Further, the transceiver may be configured to have a transmission / reception function.

  Further, the wireless communication system of the present embodiment has a wireless communication device that simultaneously uses two orthogonal polarizations (orthogonal at a realistic level, and the cross polarization component need not be 0) for information transmission at the same time. The subject. However, in the transmitter, the receiver, and the transceiver of the present embodiment, the function of the polarization multiplexing communication as described below is temporarily stopped so that it is possible to switch to the communication using the conventional communication method. It is also possible to configure the system.

  Further, the present embodiment can also be used for a polarization multiplexing system that multiplexes a signal into three or more polarizations.

(Configuration of transmitter and receiver)
FIG. 1 is a functional block diagram illustrating a configuration of a transmitter 1000 according to the present embodiment.

  Referring to FIG. 1, a transmitter 1000 receives a digital data signal (information bits) to be transmitted, and performs CRC (Cyclic Redundancy Check) coding for a check calculated from the data in order to capture a transmission error. A CRC encoding unit 100 for performing error correction encoding processing on the information bits from the CRC encoding unit 100 and converting the information bits into transmission symbols. Note that not only the error correction coding process but also an “interleave process” or the like may be performed.

  Transmitter 1000 further stores S / P conversion section 102 for performing serial / parallel conversion (S / P conversion) of the transmitted symbol and I / Q mapping data storage section 106 for the transmission symbol converted to the parallel signal. V / H mapping processing for mapping each polarization component of horizontal polarization (H polarization) and vertical polarization (V polarization) to a signal point in a signal space diagram (constellation) based on the held information. Unit 104.

  Here, the I component means an in-phase component at the time of quadrature modulation, the Q component means a quadrature phase component at the time of quadrature modulation, and the I / Q mapping means an I component and a Q component. Means that signal points are arranged on a plane spanned by.

  The transmitter 1000 further includes a unique word insertion unit (hereinafter referred to as a UW insertion unit) 107a for inserting a unique word into the I / Q component of the V polarization from the V / H mapping processing unit 104, and a V / H. A UW insertion unit 107b for inserting a unique word into the I / Q component of the H polarization from the mapping processing unit 104.

  Transmitter 1000 further performs orthogonal modulation section 108a for orthogonally modulating the I / Q component of the V polarization from UW insertion section 107a using a corresponding modulation scheme (for example, QPSK modulation scheme) and UW insertion section 107b. A quadrature modulator 108b that quadrature-modulates the I / Q component of the H-polarization with a corresponding modulation scheme (for example, a QPSK modulation scheme), and a D / A for digital-to-analog conversion of the output of the quadrature modulator 108a. It includes a conversion unit 110a and a D / A conversion unit 110b for digital-to-analog conversion of the output of the quadrature modulation unit 108b.

  The output of the D / A converter 110a is amplified by a power amplifier (not shown), filtered by a transmission filter processor 112a for suppressing unnecessary frequency components, and then transmitted from a vertically polarized antenna 114a. The output of the D / A converter 110b is amplified by a power amplifier (not shown), filtered by the transmission filter processor 112b, and then transmitted from the horizontal polarization antenna 114b.

  As will be described later, UW insertion sections 107a and 107b insert a unique word into a transmission signal at an interval D between unique word symbols known in advance on the transmission side and the reception side.

  In the following, the “unique word symbol interval” is simply referred to as “unique word interval”.

  FIG. 2 is a functional block diagram for describing a configuration of receiver 2000 according to the present embodiment.

  Referring to FIG. 2, receiver 2000 includes a vertical polarization antenna 200a, a horizontal polarization antenna 200b, and a reception filter that filters a signal amplified by a low noise amplifier (not shown) from a reception signal from vertical polarization antenna 200a. A processing unit 202a, a reception filter processing unit 202b that filters a signal amplified by a low-noise amplifier (not shown) from a reception signal from the horizontal polarization antenna 200b, and an analog-to-digital conversion of a signal from the reception filter processing unit 202a. An analog / digital conversion unit (A / D conversion unit) 204a and an A / D conversion unit 204b for converting a signal from the reception filter processing unit 202b to analog / digital conversion are provided.

  Receiver 2000 further includes quadrature detectors 206a and 206b that receive signals from A / D converters 204a and 204b, respectively, and separate I / Q components on a constellation.

  The carrier frequency recovery unit 207 performs an estimation of a value of a frequency offset between the signal from the quadrature detection unit 206a and the quadrature detection unit 206b with respect to a local oscillator as described later, and Based on the offset, the input signal is corrected to a state where the frequency error approaches zero by the same configuration as the above-described AFC circuit or the like.

  Subsequently, the received modulated signal whose frequency error has been corrected by the carrier frequency reproduction unit 207 is input to the carrier phase reproduction unit 208. Even after the correction of the frequency error by the carrier frequency reproduction unit 207, the frequency error remains, so that the received symbol point rotates slowly, causing deterioration of characteristics. Thus, the carrier frequency reproduction unit 208 corrects the remaining frequency error (phase error). The carrier phase recovery unit 208 brings the residual phase error in the corrected received modulated signal closer to zero.

  Maximum likelihood determination processing section 209 calculates the likelihood of a signal from carrier phase recovery section 208 with respect to a predetermined signal point on a signal space diagram, based on mapping information from I / Q mapping data storage section 210. Then, the maximum likelihood decoding by the MLD (Maximum Likelihood Detection) method is performed. In the MLD method, a metric is calculated for a received signal using all combinations of transmission signals that can be transmitted from a transmission antenna. Then, a combination of transmission signals that gives the minimum distance is selected.

  The "signal point" refers to a reference position defined on a constellation by a modulation method, and the "symbol" is a unit of information that is modulated on a transmission side and transmitted by a reference clock. Means a “sign”.

  The bit information of the transmission signal calculated by the maximum likelihood determination processing unit 209 passes through a P / S conversion unit 212 that performs parallel / serial conversion (P / S conversion), and is subjected to error correction by an error correction decoding processing unit 214. , And are output as reception data.

  Note that, according to the configuration on the transmitter side, convolutional decoding and deinterleave processing may be performed in the error correction decoding processing unit 214.

Further, the CRC decoding unit 216 performs a CRC decoding process on the received signal, detects whether or not a transmission error has occurred in the current combination of the unique word interval D and the correlation length N, The frequency offset estimating process in the frequency reproducing unit 207 is controlled.
(Embodiment 1)
FIG. 3 is a diagram for explaining the arrangement of symbols in a unique word.

  In FIG. 3, the symbols are BPSK modulated.

  FIG. 3A is a diagram for explaining a situation in which the phase is estimated.

  The received signal is represented as follows.

Here, zn represents noise.

  At this time, the estimated frequency is expressed as follows.

As shown in FIG. 3A, the phase estimation value θ _iD hat at the unique word position (i-D) (hereinafter, when the symbol “＾” is added to the character X, “X hat” Let us consider a case in which the phase estimation value θ _i hat of the unique word position i is obtained based on the following description). The estimated value θ _i hat includes the additive white Gaussian noise Δe _i . At this time, in FIG. 3A, an area that fluctuates due to the additive white Gaussian noise Δe _i is indicated by an up-down arrow at the unique word position i. As shown in FIG. 3A, the phase estimation error due to noise becomes relatively smaller as the interval D between unique words becomes larger, and the influence thereof becomes smaller.

  FIG. 3B shows the symbol position (symbol number) on the horizontal axis and the Pi of the symbol together with the amplitude on the vertical axis when the unique word interval D = 1 (when unique words are continuously arranged). In a noise environment, estimation is possible even when the offset range is large, but accuracy is reduced.

  FIG. 3C shows the symbol position (symbol number) on the horizontal axis and the Pi of the symbol along with the amplitude when the unique word interval D = 4. Since the symbols are spread over a wider range of the frame, higher accuracy can be obtained, but the range of the offset that can be estimated is smaller.

  FIGS. 3B and 3C illustrate a case where the number of unique word symbols is eight for simplicity.

  As shown in FIG. 3C, when the unique word interval is D = 4, data symbols exist between the unique word symbols.

  On the other hand, in a noise environment, the noise error is included in the estimation of the phase from one unique word symbol to the next unique word symbol. Therefore, as the unique word interval D increases, the influence of the noise error on the phase estimation decreases, and the estimation accuracy of the frequency offset improves.

  FIG. 4 is a functional block diagram for explaining the configurations of the frequency offset estimating device 2070 in the carrier frequency reproducing unit 207 and the CRC decoding processing unit 216.

  In the present embodiment, it is assumed that the unique word interval D is fixed as a large value in advance on the transmitting side. Here, the “large value” means a value sufficient for the accuracy of the required frequency offset estimation, and is not particularly limited. For example, any one of D = {8, 16, 32} Set to value. The receiving side adaptively changes the correlation length N on the assumption that the unique word interval D is fixed, as described below.

  In estimating the frequency offset, a modified L & R (MLR) algorithm is a widely used and well-known frequency offset estimation algorithm in DVB-S2 satellite receivers. Such a modified L & R algorithm is based on calculating the average correlation for L frames in order to improve the frequency estimation characteristics in a noisy environment.

  The modified L & R algorithm is disclosed in the following literature.

Known Document 1: E. Casini, R. De Gaudenzi, and A. Ginesi, "DVB-S2 modem algorithms design and performance over typical satellite channels," Proceedings of the Int. Journal of Sat. Commun. And Networking, pp. 281 --318, June 2004.
Known Document 2: M. Luise. And R, Reggiannini, "Carrier frequency recovery in all-digital modems for burst-mode transmissions," IEEE Trans. Commun., Vol. 43, No. 2/3/4, pp. 1169 --1178, Feb-Apr. 1995.
Further, in the following document, the inventor of the present patent application proposes a method (DMLR method) modified so that the interval D between UW symbols can be 1 or more.

Known Document 3: J. Webber, M. Yofune, K. Yano, H. Ban, and K. Kobayashi, “Performance of frequency recovery algorithms for a poly-polarization multiplexing satellite system”, 11th IEEE Malaysia International Conference on Communications (MICC2013 ), 27-29th Nov. 2013.
Here, the frequency offset estimating device 2070 is hardware for executing the “modified L & R algorithm modified so that the interval D between UW symbols can be 1 or more” as described above. Therefore, the estimation of the frequency offset is detailed in the publicly known document 3, and the outline thereof will be described below.

  The frequency offset estimating apparatus 2070 controls the reception modulation signal Z input from the quadrature detection unit 206a or 206b (hereinafter, collectively referred to as the quadrature detection unit 206) and the memory under the control of a control unit (not shown), as described later. Is calculated by a correlator 3020 having a correlation length N (N: natural number) in accordance with the following equation, and the correlation value is calculated by an integration processing unit 3030 by a moving average size L (frame number). ).

That is, the integration processing unit 3030 integrates the correlation values over the L frames, and the offset calculation unit 3040 converts the integrated correlation values into a frequency offset f _LR-D hat according to the following equations (3) and (4). I do.

  Where l is the current frame number, m is the delay, Lp is the unique word length, N is the correlation length, L is the average frame size, and D is the unique word interval. It is.

Z (k) means the product of the k-th symbol of the l-th frame and the conjugate of the k-th symbol of the known unique word correlated with the k-th symbol. By taking the product with the conjugate of a known unique word, the effects of modulation can be removed. z ^* (k) is the complex conjugate of z (k).

  In the equation (1), the correlation between symbols shifted by m is obtained for the l-th frame. Equation (2) means that integration of correlation values over L frames is performed.

The frequency offset f _LR-D hat is converted into complex frequency error information by a table or the like, and a complex multiplier (not shown) performs complex multiplication of the input received modulation signal and the frequency error information to correct the frequency error. The obtained received modulated signal is obtained. The carrier frequency reproduction section 207 repeats the above operation to make the residual frequency error in the corrected received modulated signal close to zero.

  On the other hand, under the control of a control unit (not shown), the CRC decoding unit 216 operates, the decoding unit 2162 performs a CRC decoding process on the received signal from the error correction decoding unit 214, and the CRC checking unit 2164 , CRC code to determine whether there is a transmission error. The correlation length control unit 2166 detects whether or not an error has occurred in the current unique word interval D, the correlation length N, and the frame average size L, updates the correlation length N according to a flow described later, and determines the optimum correlation length. Determine N. Note that the correlation length control unit 2166 updates the correlation length N by referring to the look-up table (LUT) 2168.

  Further, since a correlation value for a different correlation length N can be calculated for the same set of received signals, a correlation value is calculated for the same set of received signals for each element of the set of predetermined correlation lengths N. By calculating and storing in the storage device, the correlation length control unit 2166 may be configured to determine the optimum correlation length N based on the correlation value thus stored in the storage device. it can. With this configuration, at the time when the signal of the L frame is received and the correlation value is calculated, the calculation of the correlation value for each element of the set of possible correlation lengths N has been completed. , It is possible to reduce the time required to determine the optimum correlation length N.

  Although FIG. 4 illustrates that the correlation length control unit 2166 is different from the control unit that controls the entire receiver 2000, they may be integrated as one control unit.

  With this configuration, the transmitting side transmits a signal having a fixed unique word interval D, as will be described in more detail later. On the receiving side, in the initial phase, the CRC check is performed by an algorithm under the control of the control unit while sequentially selecting a specific correlation length N from a set predetermined as a set of possible values of the correlation length N. Then, the optimum correlation length N is specified in a range where the transmission error is equal to or less than a predetermined value.

  Note that the error detection code is not limited to the CRC code, and another error detection code may be used.

  FIG. 5 is a diagram illustrating a simulation result of the relationship between the correlation length N and the maximum frequency offset that can be evaluated.

  That is, FIG. 5 shows a change in the maximum frequency offset that can be pulled in with respect to the correlation length N using the unique word interval D as a parameter.

  In FIG. 5, the simulation is performed on the assumption that the number of symbols of the unique word is 32 and the data transfer speed is 10 MBd.

  As described above, for the same correlation length, the maximum frequency offset that can be evaluated decreases as the unique word interval D increases.

  On the other hand, when the unique word interval D is fixed, if the correlation length N is reduced, the maximum evaluable offset frequency can be increased.

  Therefore, even when the transmitting side is transmitting with the unique word interval D fixed, the maximum frequency offset that can be evaluated can be made variable by adaptively changing the correlation length N on the receiving side. is there.

  FIG. 6 is a diagram illustrating an example of different arrangement of unique words at the unique word interval D and the correlation length N.

  FIG. 7 is a diagram for explaining an aspect in which unique words arranged at different unique word intervals D are arranged in a frame of a transmission signal.

  Also in FIG. 5, as described, the maximum frequency offset that can be evaluated is determined by two parameters of the unique word interval D and the correlation length N.

  Here, if the effective unique word interval D ′ is defined by D ′ = D · N, the maximum frequency offset that can be evaluated is the same for the same effective unique word interval D ′.

  In FIG. 6, the unique word symbols distributed and arranged in a predetermined area of the frame are indicated by oblique lines. The unique word (UW) symbol index indicates the order of a certain unique word symbol starting from the first unique word symbol in a frame.

  For example, (D, N, D ') = (1,4,4) as shown in FIG. 6A and (D, N, D') = (D) as shown in FIG. 6B. In the case of (4, 1, 4), the effective unique word interval D 'is 4, and the maximum frequency offset that can be evaluated is the same.

  However, for the purpose of offset evaluation, the interval between any two unique words needs to be set so that the corresponding phase difference is 2π or less, and the value of the unique word interval D has an upper limit.

  Next, referring to FIG. 7, as shown in the upper part of FIG. 7, when the unique word interval D = 1, a predetermined number of consecutive symbols (the symbol The predetermined number includes, for example, a portion (hereinafter referred to as a UW portion) in which a unique word consisting of 64) is arranged.

  On the other hand, as shown in the lower part of FIG. 7, in the arrangement where the unique word interval D exceeds 1, a predetermined unique word interval (for example, D = 2 to 64 in the lower part of FIG. 7) from the head of each frame. , Unique words of a predetermined number of symbols are dispersedly arranged.

  In this embodiment, the transmitting side selects and fixes the unique word interval D from a set of predetermined values, and the selected unique word interval is also used for data communication.

  Here, the number of symbols used as the unique word is not necessarily limited to the above value.

  Here, the unique word interval D = n (n: natural number) means that when the ith symbol (UW symbol index = i) is a unique word, the (i + n) th symbol is also a unique word. It means there is.

  FIG. 8 is a diagram for explaining the arrangement of unique words and CRC codes transmitted from the transmission side to the reception side.

  FIG. 8A shows an arrangement of unique words distributed in one frame, and FIG. 8B shows L frames having a frame average size, which is the number of frames to be subjected to frame averaging in frequency offset evaluation. 8 shows an example of the arrangement of the CRC code provided corresponding to the frame, and FIG. 8C shows another example of the arrangement of the CRC code.

  Note that, as shown in FIG. 8B or FIG. 8C, a set of frames that is a unit for performing error detection using a CRC code or the like is referred to as a block. In the present embodiment, it is assumed that the unique word interval D is constant within the same block.

  As shown in FIG. 8A, the unique words are distributed in one frame at a designated interval D.

  As shown in FIG. 8 (b), it is possible to arrange, for example, a 16-bit CRC symbol immediately after transmission of L frames (frame numbers 1 = 1 to L), which is the average frame size. .

Alternatively, as shown in FIG. 8 (c), immediately after the transmission of the L (frame number l = 1 to L) is the frame average size and the frame for each of the L _N-number of subsequent (L _N ≧ 0) For example, a configuration in which a 16-bit CRC symbol is arranged is also possible.

However, the arrangement of the CRC symbols is not limited to the arrangement shown in FIGS. 8B and 8C, and may be, for example, provided in each frame. It is sufficient that at least one is arranged at a position where a transmission error can be checked after receiving L frames having a predetermined unique word interval D on the side. Desirably, (L _N +1) positions may be arranged at positions where transmission errors can be checked after receiving L frames at a predetermined unique word interval D. When (L _N +1) CRC codes are arranged, for example, the configuration is such that error detection results for each of these CRC codes are integrated, and the presence or absence of an error is detected based on the integrated results. The effect can be reduced.

  Therefore, for example, the following procedure can be used as a flow for adaptively changing the correlation length N on the receiving side.

  i) The average correlation is calculated by using the correlator shown in FIG. 4 with the average frame size being L frames.

  At this time, when the number of received frames is less than L, the correlation length N is set to a predetermined value stored in the lookup table 2168. By setting the initial value of the correlation length N in this way, it is possible to evaluate the frequency offset with the correlation length N suitable for the unique word interval D fixed on the transmission side.

  ii) When the number of received frames is L or more, the value of the correlation length N is adaptively changed according to the result of error detection by the CRC code according to the flow described later.

At this time, when the CRC codes are arranged as shown in FIG. 8C, by performing a process of integrating (averaging) the error detection results by the CRC codes for (L _N +1) CRC codes. As described above, the effect of noise can be reduced.

  FIG. 9 is a conceptual diagram showing the evaluation of the adjustable maximum frequency offset when the correlation length N is changed adaptively.

  FIG. 9A shows the evaluation of the maximum frequency offset by the above-mentioned conventional modified L & R (MLR) algorithm, and FIG. 9B shows the algorithm (fine-tuning DMLR algorithm) in the present embodiment in a state of high accuracy. 9) shows the evaluation of the maximum frequency offset by the algorithm (coarse DMLR algorithm) in the present embodiment in a state of low accuracy.

  It is assumed that the data transfer rate fsym is 10 MBd.

  As shown in FIG. 9A, in the conventional MLR algorithm, when (D, N, D ′) = (1, 16, 16), the maximum frequency offset Δfmax is 〜588 kHz.

  On the other hand, in the fine-tuning DMLR algorithm as shown in FIG. 9B, for example, 64 is set as a sufficiently large unique word interval D, and (D, N, D ') = (64, 16, 1024). In this case, the maximum frequency offset Δfmax becomes 〜9.19 kHz. On the other hand, as shown in FIG. 9C, when the correlation length N is reduced and D, N, D ') = (64, 2, 128) as the coarse DMLR algorithm, the maximum frequency offset Δfmax is , To 52.1 kHz.

  FIG. 10 is a diagram showing a simulation result of the relationship between the bit energy-to-noise power density ratio Eb / No and the bit error rate BER.

  FIG. 10 shows a case where the unique word interval D and the correlation length N are changed as parameters.

  For the number of unique word symbols Nuw = 32, the correlation length employed in the conventional MLR algorithm is Nuw / 2 = 16.

  The modulation scheme is 8PSK for each of the two polarizations, the frequency offset is Δf = ± 2 kHz, the data transfer rate fsym is 10 MBd, the average frame size L = 256, and the forward error correction (FEC) is The simulation is performed under the condition that the simulation is not performed.

  FIG. 10 shows the results of the following algorithms in comparison.

-Conventional MLR algorithm: (D, N, D ') = (1, 16, 16)
DMLR algorithm: (D, N, D ') = (2,16,32)
(D, N, D ') = (4,16,64)
(D, N, D ') = (32, 16, 512)
-Adaptive correlation length algorithm: (D, N, D ') = (32, 8, 256)
(D, N, D ') = (32,4,128)
The adaptive correlation length algorithm of (D, N, D ') = (32,4,128) is larger than the DMLR algorithm of (D, N, D') = (32,16,512). Comparable accuracy can be achieved while supporting the maximum frequency offset.

  FIG. 11 is another diagram showing a simulation result of the relationship between the bit energy to noise power density ratio Eb / No and the bit error rate BER.

  The simulation conditions are the same as in FIG.

  Even if the correlation length N is reduced from 16 to 1 by setting the unique word interval D to 16, no significant deterioration is observed in the BER. In other words, setting D = 16 indicates that the correlation length N can be changed on the receiving side to make the maximum frequency offset variable.

  In actual communication, the variable ranges of the unique word interval D and the correlation length N can be set in advance by such simulations and experiments so as to be adapted to the communication environment. 2168.

  FIG. 12 is a flowchart for explaining the processing flow of the adaptive correlation length.

  As described above, in the present embodiment, the unique word interval D is fixed at a fixed value on the transmitting side, and the correlation length N is changed on the receiving side to specify the optimum correlation length N.

  As described below, the maximum correlation length N is specified by using the CRC check. Here, it is assumed that the arrangement of the CRC codes is, for example, as shown in FIG. 8C.

  Referring to FIG. 12, first, as initialization processing, the transmitting side sets the value of unique word interval D to a predetermined value, and this value D is kept constant throughout the entire transmission. It is assumed that the unique word interval D is also known on the receiving side. Here, D = 16 (S100).

  Further, as initialization processing, it is assumed that a possible combination of correlation lengths Nset = {Nset (1), Nset (2),..., Nset (n)} is set in advance on the receiving side (S100). ).

  Here, as an example, Nset = {Nset (1), Nset (2),..., Nset (5)} = {1, 2, 4, 8, Nuw / 2 (= 16)}. Although not particularly limited, as described below, it is desirable that the elements of Nset are arranged in ascending order. In the following, it is assumed that the elements of Nset are arranged in ascending order.

N _N is the number of elements of the Nset to be tested. The frame average size L is assumed to be 256, for example. _LN shown in FIG. 8C is assumed to be 1. Therefore, the number of times Lc for performing error detection by the CRC code per block is Lc = (L _N +1). Further, during processing, n, which is a working variable for designating an element in Nset, is initialized to n = 0.

Further, referring to FIG. 12, the receiving side receives (L + L _N ) frames including the unique word (S102).

  Next, the value of n is incremented by 1 and the correlation length N is set to N = Nset (n) (S104).

A CRC check is performed for the first L frames and subsequently for a total of L _N frames. If the CRC check does not pass, CRC (l) = 1, and if it passes, CRC (l) = 0 (S106).

  The failure index Fail (n) is calculated by the following equation (S108).

Then, the current value of variable n, whether there are the number of elements N _N or more in the Nset is discriminated (S110), is less than N _N, process returns to step S104, the value of n Is incremented by 1, the next value is set as the correlation length N (S104), and the subsequent processing is repeated.

On the other hand, in step S110, if n is equal to or more than N _N , the maximum correlation length N among the correlation lengths having the minimum failure index Fail (n) is selected as the correlation length for evaluating the frequency offset. (S112).

  That is, when priority is given to accuracy, it is desirable to select the element having the largest value of N when two or more elements among the elements in Nset have the same value of the defect number index.

  However, by giving priority to the maximum frequency offset that can be evaluated, the minimum correlation length N is selected as the correlation length for evaluating the frequency offset among the correlation lengths having the minimum failure number index Fail (n). Alternatively, the configuration may be as follows.

As described above, the failure index is defined as an integrated value of the results of error detection for Lc = ( _LN + 1) times. As a result, the influence of noise or the like can be reduced, but the index (error information) indicating the error situation in a unified manner for a plurality of frames is not limited to this. For example, by using the ratio of the number of frames in which an error is detected to Lc, a method of determining whether or not the ratio is equal to or less than a predetermined value may be employed.

  FIG. 13 is a diagram illustrating the relationship between the number of received frames and the remaining frequency offset.

  FIG. 13 shows a case where the test is performed while sequentially changing the correlation length N from a smaller value to a larger value in the flowchart of FIG.

  It is assumed that the unique word interval D = 8, the frame average size L = 64, the bit energy to noise power density ratio Eb / No = 4.8 dB, the data transfer rate fsym is 5 MBd, and the initial frequency offset Δf = 64 kHz. .

  In the initial phase, the remaining frequency offset tends to decrease as the correlation length N is sequentially increased from N = 1 to N = 8 according to the flowchart of FIG. However, when the correlation length N = 16, the value of the maximum frequency offset that can be evaluated becomes too small, and the residual frequency offset rapidly increases.

For this reason, in the range of N = 1 to N = 8, and in the range of the minimum defect index, for example, the maximum correlation length N is selected as the optimum correlation length N _OPT , and thereafter, the correlation length in the communication phase Will be used as

  By adaptively setting the value of N in the initial operation start phase, stable communication can be performed even in the subsequent communication phase.

(Storing correlation data in storage device)
In the following, while the correlation value is stored in a storage device (not shown) during the evaluation of the optimum value of the correlation length on the receiving side, when the optimum correlation length is determined, the optimum value is immediately handled. A configuration that enables calculation of a correlation value to be calculated will be described.

That is, although the processing shown in FIG. 12 may be executed as the processing for selecting the correlation length N, after the correlation length N is selected, as described below, (L + L _N ) Need not be received. Such a configuration will be described below.

  FIG. 14 is a conceptual diagram illustrating an example of a correlation value stored in the storage device.

In step S102 of FIG. 12, while sequentially increasing the frame number l by one, the frames are sequentially received, and the correlation length N is changed by the correlator 3020 for the L frames from the frame l to the frame (l + L-1). The stored values are stored in a storage device not shown in FIG. 4, as shown in FIG. At this time, the size of the data to be stored is a matrix of L rows and _NN columns.

If the optimum correlation length N _OPT is determined in step S112, the correlation value corresponding to the correlation length N _OPT is read from the storage device, integrated by the integration processing unit 3030, and the offset calculation unit 3040 outputs the frequency offset. Can be calculated. As a result, in the evaluation of the frequency offset, l = 1 to L are targets for taking a moving average.

  As described above, during the operation start phase, a predetermined number of frames are transmitted while selecting a specific correlation length from the set of correlation lengths on the reception side, and the reception side performs decoding processing of the CRC code. The transmission error is evaluated, and the correlation length is specified in a range where the transmission error is minimized.

  At this time, if the maximum correlation length among the minimum transmission error correlation lengths is selected as the correlation length for evaluating the frequency offset, the frequency offset can be estimated with high accuracy.

  Alternatively, if the smallest correlation length among the smallest transmission error correlation lengths is selected as the correlation length for evaluating the frequency offset, the range of the frequency offset that can be evaluated can be made as wide as possible.

  Of the correlation lengths that minimize the transmission error described above, the maximum or the minimum is selected depending on the frequency band to be used, the modulation method to be used, the data transfer speed, and the assumed frequency offset. It shall be determined in advance according to the range.

  In the data communication phase, data reception is performed with the correlation length specified in this manner, thereby maximizing the range of frequencies that can be pulled in the carrier frequency recovery unit on the receiving side, or estimating the frequency offset with high accuracy. It is possible to do. In this case, since the correlation length is selected so that the transmission error is minimized, the frequency offset can be estimated with a smaller number of frames (shorter acquisition period).

  The embodiment disclosed this time is an exemplification of a configuration for specifically implementing the present invention, and does not limit the technical scope of the present invention. The technical scope of the present invention is defined by the terms of the claims, rather than the description of the embodiments, and includes changes within the wording and equivalents of the claims. Is intended.

  100 error correction coding processing section, 102 S / P conversion section, 104 V / H mapping processing section, 106, 210 I / Q mapping data storage section, 107a, 107b UW insertion section, 108a, 108b quadrature modulation section, 110a, 110b D / A converter, 112a, 112b Transmission filter processor, 114a, 200a Vertical polarization antenna, 114b, 200b Horizontal polarization antenna, 202a, 202b Reception filter processor, 204a, 204b A / D converter, 206a, 206b Quadrature detection section, 207 Carrier frequency recovery section, 208 Carrier phase recovery section, 209 Maximum likelihood determination processing section, 212 P / S conversion section, 214 Error correction decoding processing section, 216 CRC decoding processing section, 1000 transmitter, 2000 reception Machine.

Claims (8)

A frequency offset estimating device for estimating a frequency offset between a carrier frequency of a received signal and a local oscillation frequency,
The received signal includes a plurality of frames each having a plurality of unique words arranged therein, and an error detection code,
A correlator for calculating a correlation between the received signal and the unique word, wherein the correlator can calculate the correlation with a variable correlation length;
The calculated correlation is integrated over the number of frames of the set frame average size, an offset calculation unit for calculating an estimated value of the frequency offset,
An error detection unit for detecting a transmission error by the error detection code,
The correlator further includes a control unit for controlling the offset calculation unit and the error detection unit, the control unit,
i) receiving a signal transmitted at a predetermined unique word interval at a receiving side, and performing error detection by sequentially changing the correlation length among a set of predetermined values;
ii) A frequency offset estimating apparatus that specifies a correlation value used for estimating a frequency offset within a range in which a transmission error is equal to or less than a predetermined value in accordance with a result of the error detection. The error detection code is provided over a plurality of frames,
The error detection unit, and integrating the error detection results for the plurality of frames, calculates an error information,
The frequency offset estimating device according to claim 1, wherein the control unit specifies a maximum correlation length among correlation lengths corresponding to a frame with a minimum transmission error based on the error information. The error detection code is provided over a plurality of frames,
The error detection unit, and integrating the error detection results for the plurality of frames, calculates an error information,
The frequency offset estimating apparatus according to claim 1, wherein the control unit specifies a minimum correlation length among correlation lengths corresponding to a frame with a minimum transmission error based on the error information. Further comprising a storage device for storing a correlation value used for calculating the frequency offset,
The control unit, for each frame of the number of the received frame average size, stored in the storage device a correlation value calculated by sequentially changing the correlation length among a predetermined value set,
The frequency offset estimating device according to claim 1, wherein the offset calculating unit calculates an estimated value of the frequency offset based on the correlation value stored in the storage device. A plurality of frames each having a plurality of unique words, and a wireless communication device for receiving a wireless signal including an error detection code,
Receiving means for receiving and detecting the wireless signal,
Estimating a frequency offset between the carrier frequency of the received signal from the receiving means and the local oscillation frequency, based on the estimated frequency offset, a carrier frequency reproducing means for correcting the frequency error of the received signal. Prepare,
The carrier frequency reproduction means includes a frequency offset estimating means for estimating the frequency offset, the frequency offset estimating means,
A correlator for calculating a correlation between the received signal and the unique word, wherein the correlator can calculate the correlation with a variable correlation length,
The calculated correlation is integrated over the number of frames of the set frame average size, an offset calculation unit for calculating an estimated value of the frequency offset,
The error detection code further includes an error detection unit for detecting a transmission error,
The correlator further comprises a control unit for controlling the offset calculation unit and the error detection unit, the control unit,
i) receiving a signal transmitted at a predetermined unique word interval at a receiving side, and performing error detection by sequentially changing the correlation length among a set of predetermined values;
ii) Specifying a correlation value to be used for estimating a frequency offset within a range in which a transmission error is equal to or less than a predetermined value according to a result of the error detection;
Respect corrected signal in the previous SL carrier frequency reproducing means further comprises a decoding means for performing decoding processing, the wireless communication device. The error detection code is provided over a plurality of frames,
The error detection unit, and integrating the error detection results for the plurality of frames, calculates an error information,
The wireless communication device according to claim 5, wherein the control unit specifies, based on the error information, a maximum correlation length among correlation lengths corresponding to a frame with a minimum transmission error. The error detection code is provided over a plurality of frames,
The error detection unit, and integrating the error detection results for the plurality of frames, calculates an error information,
The wireless communication device according to claim 5, wherein the control unit specifies a minimum correlation length among correlation lengths corresponding to a frame with a minimum transmission error based on the error information. Further comprising a storage device for storing a correlation value used for calculating the frequency offset,
The control unit, for each frame of the number of the received frame average size, stored in the storage device a correlation value calculated by sequentially changing the correlation length among a predetermined value set,
The wireless communication device according to claim 5, wherein the offset calculation unit calculates the estimated value of the frequency offset based on the correlation value stored in the storage device.