JP2012049768A - Receiver - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve precision of a training process of a receiver.SOLUTION: An FFT window deviation amount τ setting unit 50 sets an FFT start time to be a second natural frequency which is larger than a first natural number, for a training signal that is provided to a received frame and is continuously allocated to a unit time (T) which is a first natural multiple equal to 2 or larger. A detection unit 20, based on the set FFT start time, deviates an FFT start time to perform a detection process of the training signal in parallel. A synthesizing unit performs a weighing calculation process which introduces a weighing factor which causes an error power that is introduced to be smaller, based on the detected training signal and a reference signal which is predetermined according to the training signal. A signal conversion unit 40 performs a demodulation process based on the synthesized signal synthesized with weighing.

Description

  The present invention relates to a receiving apparatus that receives a modulated signal.

  2. Description of the Related Art An OFDM (Orthogonal Frequency Division Multiplexing) system that can secure transmission capacity using a plurality of subcarriers is known in a receiving apparatus that receives a modulated signal. As an example to which the OFDM system is applied, there is a wireless LAN according to IEEE802.11a.

  In a wireless LAN conforming to IEEE802.11a, data is transmitted divided into frames. For each subcarrier assigned to the transmission of the frame, a training process at the time of reception is performed based on the training symbol in the preamble included in the frame.

  In addition, there is a method of receiving an OFDM signal with an adaptive array antenna configuration, improving the RLS (Recursive Least Square) algorithm, and converging the antenna weight with a small number of training symbols (see, for example, Non-Patent Document 1).

Japanese Patent Laid-Open No. 10-322144 JP 2000-031753 A

Fujii, Suzuki, “OFDM RLS adaptive array antenna using repetitive weight update with adjacent subcarrier signals”, IEICE Tech. Bulletin DSP2003-78, WBS2003-46, July 2003.

  By the way, in a frame transmitted by being divided into frame units, an allocation time as a training time is limited in order to increase transmission efficiency. In IEEE802.11a, since there are only two symbols assigned to training per frame, if training is performed for each symbol, it can be performed only twice. Further, this symbol is defined as a signal in which a known signal (reference signal) is assigned to every subcarrier.

  In addition, the reception quality is improved by combining radio signals received from a plurality of antennas. For example, a case is assumed in which a weighting coefficient that minimizes error power between a prescribed reference signal and a weighted combined output is derived using a MMSE (Minimum Mean Square Error) standard to perform a weighting operation. In order to converge the weighting coefficient based on signals received from a plurality of systems, it is necessary to update the weights at least twice as many times as the number of systems to be combined (the number of branches). For example, in the case of two systems, it is necessary to update the weight four times or more.

  However, when the number of symbols assigned to a frame is smaller than the number of times, the weighting coefficient does not converge and a weighting coefficient including an error is derived. If an error is included in the weighting coefficient, the training process is terminated in an incomplete state. In such a state, there is a problem that the communication quality may be deteriorated because subcarriers whose weighting coefficients cannot be stochastically converged may occur.

  The present invention has been made to solve the above problems, and an object of the present invention is to provide a receiving apparatus that can improve the accuracy of training processing of the receiving apparatus.

  In order to solve the above problem, the present invention is a wireless communication system that performs frequency division multiplexing transmission using a subcarrier frequency, and receives a frame to which a predetermined predetermined unit time training signal is received. A second natural number of times greater than the first natural number for the training signal assigned to the received frame and continuously allocated in the unit time that is two or more times the first natural number times A setting unit for setting the FFT start time, a detection unit for detecting the training signal in parallel based on the set FFT start time, the detected training signal, and the training signal A combining unit that performs a weighting calculation process that derives a weighting coefficient that reduces the derived error power based on a predetermined reference signal; A signal conversion unit for the demodulation processing based on serial weighted synthesized composite signal to a receiving apparatus, characterized in that it comprises a.

  Further, the present invention is the above invention, wherein the setting unit divides a time obtained by multiplying the unit time by a number obtained by subtracting 1 from the first natural number by a number obtained by subtracting 1 from the second natural number. The FFT start time is set at the starting point.

  Moreover, the present invention is characterized in that, in the above-mentioned invention, the synthesizing unit includes a phase correcting unit that corrects a phase of the reference signal in accordance with the set FFT start time.

  Further, the present invention is the above invention, wherein the detection unit adds information on a range obtained using a window function that defines a range of information to be extracted from the received frame according to the set FFT start time. Detection processing is performed based on this.

  Also, the present invention is characterized in that, in the above-mentioned invention, the combining unit derives a weighting coefficient for reducing the error power by using MMSE.

  By the means, there is an effect that the accuracy of training processing of the receiving apparatus can be improved.

It is a schematic block diagram which shows the receiver in this embodiment of this invention. It is a figure which shows the setting of the FFT window in this embodiment. It is a figure which shows a process in case an interference signal with a large level fluctuation | variation is contained in the training signal in this embodiment. It is a block diagram which shows the window control part in this embodiment. It is a figure which shows the phase control of the reference signal according to the setting of the FFT window in this embodiment. It is a block diagram which shows the structure of the phase correction part in this embodiment. It is a graph which shows the difference in the square error by the difference in training processing. It is a graph which shows the difference in the CNR-BER characteristic by the difference in training processing. It is a block diagram which shows the window control part in 2nd Embodiment. It is a block diagram which shows the window control part in 3rd Embodiment. It is a block diagram which shows the window control part in 4th Embodiment.

(First embodiment)
Hereinafter, an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a schematic block diagram illustrating a receiving apparatus according to the present embodiment. The receiving apparatus 1 shown in the figure includes a radio unit 10, a detection unit 20, a synthesis unit 30, a signal conversion unit 40, and an FFT window shift amount τ setting unit 50.

  The radio unit 10 includes an orthogonal detection unit 11, an analog / digital conversion (A / D) unit 12, and a guard interval (GI) removal unit 13. The quadrature detection unit 11 performs quadrature detection on the radio wave received from the antenna. An analog / digital conversion (A / D) unit 12 converts the quadrature-detected analog signal into a digital signal. The guard interval (GI) removal unit 13 detects a frame from information converted into a digital signal, removes the guard interval signal included in the frame, and generates a reception frame.

  Based on the generated reception frame, the detection unit 20 detects a training signal included in a predetermined time range for each subcarrier, and generates a subcarrier signal. The detection unit 20 includes a window control unit 21, a serial / parallel conversion (S / P) unit 22, and an FFT unit 23. The window control unit 21 extracts a range in which detection processing for detecting a frequency component included in the training signal included in the generated preamble of the frame is extracted according to the control signal from the FFT window shift amount τ setting unit 50. Functions as a window function. The serial / parallel conversion (S / P) unit 22 performs serial / parallel conversion on a signal whose extraction range is determined by a window function. The FFT unit 23 performs FFT processing based on the parallel-converted time domain information, converts it to a frequency domain signal, and generates a subcarrier signal for each subcarrier.

  The combining unit 30 is provided for each subcarrier, and performs combining processing for each subcarrier based on the result of FFT processing performed by each detection unit 20. Here, the combining unit 30 describes the m-th subcarrier. The synthesis unit 30 includes a synthesis processing unit 31, a reference signal unit 32, a phase correction unit 33, a subtractor 34, and a weight coefficient calculation (MMSE) unit 35.

The synthesis processing unit 31 includes multipliers 31a and 31b and an adder 31c. The multiplier 31a is to calculate the weighting coefficient _{W mn1} been multiplied by the extracted subcarrier signals _{X mn1.} The multiplier 31b may Calculation is weight coefficient _{W mn2} was multiplied by the extracted subcarrier signals _{X mn2.} The adder 31c adds the calculation results calculated by the multiplier 31a and the multiplier 31b, and generates a combined signal _Ymn of the subcarrier signal. Here, n is an integer of 1 ≦ n ≦ N, and N indicates the number of pseudo training symbols described later.

The reference signal unit 32 stores reference signal information determined according to the training signal, and is referred to according to the arithmetic processing. The phase correction unit 33 performs phase control on the reference signal D _mn held in the reference signal unit 32 according to the control signal from the FFT window shift amount τ setting unit 50. The subtractor 34 calculates an error signal e _mn between the combined signal Y _mn calculated by the combining processing unit 31 and the phase-adjusted reference signal D _mn .

Weighting factor calculation (MMSE) unit 35, guided error signal _{e mn,} and, based on the sub-carrier signal _{X mn1} subcarrier signal _{X mn2} extracted respectively, the mean square error _{(E [| e mn |} ^2]) is calculated weighting factor _{W mn1} and _{W mn2} to minimize. The weighting factor calculation (MMSE) unit 35 performs this calculation based on the MMSE standard. The combining process performed by the combining unit 30 is expressed as Expression (1), and the mean square error (E [| e _mn | ² ]) is expressed as Expression (2).

In Equation (1), Y _mn represents a weighted composite signal, W _mn represents a weight vector, X _mn represents a subcarrier signal vector, and H represents a complex conjugate transpose. In the equation (2), D _mn represents a reference signal, and e _mn represents an error signal between the reference signal D _mn and the weighted combined output Y _mn . Here, the weight vector W _mn and the subcarrier signal vector X _mn are shown as Expression (3) and Expression (4), respectively.

  However, T in Formula (3) and Formula (4) indicates transposition.

  The signal conversion unit 40 includes a demodulation unit 41 and a parallel-serial conversion (P / S) unit 42. The demodulator 41 performs demodulation processing based on the combined signal of each subcarrier signal. The parallel-serial conversion (P / S) unit 42 performs parallel-serial conversion processing based on the demodulated signal.

  The FFT window shift amount τ setting unit 50 sets the FFT window shift amount τ in order to cause the window control unit 21 to function as a window function that determines the range of information for which the FFT unit 23 performs the FFT calculation in the detection unit 20. Moreover, the FFT window shift amount τ setting unit 50 causes the phase correction unit 33 to perform phase conversion of the reference signal held in the reference signal unit 32, so that it corresponds to the FFT window shift amount τ set in the window control unit 21. Phase control.

  The FFT window shift amount τ setting unit 50 regards the training signal as a continuous training signal regardless of the basic period of the training signal included in the received frame, and determines the time based on the shift amount τ with respect to the training signal regarded as continuous. Information from the shifted starting time is used as a pseudo training signal (hereinafter referred to as a pseudo training symbol), and training processing is performed. When the shift amount τ is a predetermined constant value, for example, it is derived by Expression (5).

  In Equation (5), τ represents the FFT shift amount, N represents the number of pseudo training symbols, K represents the number of continuous training signals, and T represents the FFT period.

  A method for setting an FFT window will be described with reference to the drawings. FIG. 2 is a diagram showing the setting of the FFT window in the present embodiment. As an aspect of the present embodiment, a description will be given using a training signal conforming to IEEE802.11a. The preamble of the IEEE802.11a frame is composed of a “short preamble” transmitted first and a “long preamble” transmitted thereafter. “Short preamble” is mainly used for timing detection and coarse adjustment of AFC. The “long preamble” is mainly used for AFC fine adjustment and channel estimation.

  In the range shown in this figure, the “long preamble” portion of the IEEE802.11a frame preamble is enlarged. This “long preamble” consists of two long training symbols (TL1 and TL2) following the guard interval GI2. Long training symbols that are repetitions of the same time waveform are used for so-called training signals, each having the same length as an OFDM symbol. Therefore, the total length of TL1 and TL2 is the length of two OFDM symbols. Here, the period of one OFDM symbol, which is the length of each of TL1 and TL2, is shown as an FFT cycle T.

  In conventional general training processing, the FFT processing is performed once every TL1 and TL2. In the training process shown in this embodiment, a phase difference is set between the symbol period and the FFT process period, and a plurality of FFT processes are performed in parallel. Here, N FFT processes are performed. The first FFT processing is indicated as FFT_1, the second FFT processing is indicated as FFT_2, and the Nth FFT processing is indicated as FFT_N. FFT_2 starts processing from a time delayed by a predetermined period (τ) from FFT_1. The start times of a plurality of FFT processes that are sequentially delayed by a predetermined period (τ) are defined.

  Since the phase continuity of each subcarrier of the training signal is ensured at the boundary between TL1 and TL2, even if the FFT period is performed across the boundary between TL1 and TL2, the OFDM delayed wave is within the guard interval GI2. As long as it falls within the range, linear convolution based on propagation path characteristics can be regarded as cyclic convolution.

  With reference to the figure, other effects expected by performing the FFT processing a plurality of times will be described. FIG. 3 is a diagram illustrating processing when an interference signal having a large level variation is included in the training signal according to the present embodiment. In the period of the training process illustrated in FIG. 3, a case where interference waves interfere with each other during the TL2 period among the continuous periods of TL1 and TL2 is illustrated.

  First, the operation of the conventional method will be described. When FFT processing is performed according to the symbol period, channel estimation of a desired signal can be performed in the TL1 processing, but interference waves interfere with the next TL2, and the interference waves cannot be suppressed only by the TL2 processing. This is because only the TL2 process results in one weight update count for suppressing interference, and the weight update count for converging the weight coefficient is insufficient. Therefore, there is a problem that the weighting coefficient for suppressing the interference wave cannot be converged in the training process by the conventional method.

  Therefore, an operation when the present embodiment is applied to TL1 and TL2 including the same interference wave will be described. Each period from FFT_2 to FFT_5 includes a period in which interference waves interfere with each other. Therefore, multiple detections can be performed for the same interference wave. Thereby, it is possible to obtain information supplied to the calculation of the weighting coefficient four times from the information of the interference wave having a large influence, and further, it is possible to calculate the weighting coefficient that cancels the interference wave. Training can be performed in a state in which a weighting factor that cancels the interference wave is applied. Therefore, even if the training signals have the same length, the interference wave included in a part of the training signal period can be effectively used and converged to the weighting coefficient for suppressing the interference wave.

  With reference to the drawings, a more specific embodiment of the configuration shown in FIG. 1 will be described. FIG. 4 is a block diagram showing the window control unit in the present embodiment. The window control unit 21A and the FFT window shift amount τ setting unit 50A shown in the figure are one mode of the window control unit 21 and the FFT window shift amount τ setting unit 50 shown in FIG. The window control unit 21A includes an FFT window setting unit 211 and an FFT start timing notification unit 212.

  The FFT window setting unit 211 extracts a signal included in a predetermined time range from the signal supplied from the wireless unit 10 according to the set control information. The FFT start timing notification unit 212 generates a control signal for setting a time window for FFT processing so that the FFT window setting unit 211 extracts information in a predetermined time range, and supplies the control signal to the FFT window setting unit 211. The FFT start timing notifying unit 212 generates a control signal to the FFT window setting unit 211 based on the control signal supplied from the FFT window shift amount τ setting unit 50A.

  Further, the control signal to the FFT window setting unit 211 is a signal indicating the required FFT start time, regarding the FFT window setting unit 211 as an FFT window with a fixed FFT start time. The FFT window shift amount τ setting unit 50 </ b> A supplies control information for instructing the FFT start time to the FFT start timing notification unit 212. With this configuration, the FFT window can be set with a simple configuration.

The reference signal phase control according to the setting of the FFT window will be described with reference to the drawings. FIG. 5 is a diagram illustrating the phase control of the reference signal according to the setting of the FFT window in the present embodiment. This figure shows a method of controlling the phase of the reference signal in accordance with IEEE802.11a as previously shown in FIGS. Reference signals d ₁ , d ₂ , and reference signals corresponding to the first FFT process, FFT_1, the second FFT process, FFT_2, and the Nth FFT process, FFT_N, respectively. the signal _{d N.} The reference signal D _mn corresponding to the mth subcarrier of the nth pseudo training symbol is derived by Equation (6).

In Equation (6), d ₁ represents the original reference signal as a reference, FFT [·] _m represents the symbol of the m th subcarrier, and D _m represents the m th of the original reference signal d ₁ as a reference. The subcarrier signal is shown. Term exponential operation of the right-hand side of equation (6) is a phase correction term to perform phase correction operation on the reference signal d ₁ is the original reference signal referenced. By performing this calculation, a reference signal subjected to necessary phase control can be generated.

  The configuration of the phase correction unit will be described with reference to the drawing. FIG. 6 is a block diagram showing the configuration of the phase correction unit in this embodiment. The phase correction unit 33 shown in the figure includes a phase correction term calculation unit 331 and a multiplier 332. The phase correction term calculation unit 331 generates a phase correction signal for controlling the phase of the reference signal in accordance with the control signal supplied from the FFT window shift amount τ setting unit 50A.

The multiplier 332 multiplies the supplied reference signal (D _m ) by the phase correction signal generated by the phase correction term calculation unit 331, and outputs a phase-corrected reference signal. For example, the phase correction term calculation unit 331 performs the calculation of the exponent operation term on the right side in Equation (6), and the phase correction unit 33 can perform the calculation processing of Equation (6).

The effect of the training process in the present embodiment will be shown with reference to the drawings. FIG. 7 is a graph showing a difference in square error due to a difference in training processing. The vertical axis in the figure represents the square error (| e | ² ), and the horizontal axis represents the number of training calculations.

FIG. 7A is a graph showing an error occurrence state when using a pseudo training symbol for performing a plurality of training processes according to the present embodiment. The vertical axis of this graph indicates the square error (| e | ² ), and the horizontal axis indicates the number of trainings that are repeatedly performed. This error occurrence situation is when the number of FFT processes (N times) shown in FIG. 2 is set to 5 in the configuration of FIG. 1, and the waveform shown for each line indicates the convergence status for each subcarrier. Show. As the number of times gradually increases from 1, the square error when received using information on the result of the training process up to that number tends to gradually decrease. When the number of pseudo trainings reaches 5, the square error becomes small, the square error converges to 10 ⁻² or less, and the effect of training is sufficiently exhibited.

FIG. 7B shows a result obtained by the conventional method, and shows a square error of the result of the training processing of the number of times (two times) corresponding to the two symbols TL1 and TL2. In this graph, as shown in FIG. 7A, the square error of many subcarriers tends to decrease as the number of times increases, but the square error of some subcarriers does not decrease, and 10 Some have a square error higher than ^-2 .

  In FIG. 7B, when the number of times is 3 or more, it is not a period of training processing but a period in which actual data is transmitted, but is required for a subcarrier in a graph in which the square error does not decrease. This is a state in which the square error is not secured, and the communication quality is not sufficiently secured. Thus, by comparing two graphs with different methods, the present embodiment shows that the square error is improved and the communication quality is ensured.

  The CNR-BER characteristic in this embodiment is shown with reference to the figure. FIG. 8 is a graph showing differences in CNR-BER characteristics due to differences in training processing. In the figure, the vertical axis represents the bit error rate (BER), and the horizontal axis represents CNR (dB). In this graph, three cases with different DUR conditions are shown together. The pseudo training symbol addition method according to the present embodiment shows that the bit error rate is improved by 3 dB even if the DUR condition is different from the conventional two-symbol weight update method.

  Further, in the present embodiment, interference suppression can be performed as an effect by using MMSE for calculation of the weighting coefficient for performing the synthesis process. In addition, by configuring a reception system using a plurality of antennas, it is possible to adapt the diversity gain according to the reception situation.

(Second embodiment)
Referring to the drawings, different aspects of the receiving apparatus shown in FIG. 1 are shown. FIG. 9 is a block diagram showing the window control unit in the present embodiment. The window control unit 21B and the FFT window shift amount τ setting unit 50B shown in the figure are one mode of the window control unit 21 and the FFT window shift amount τ setting unit 50 shown in FIG.

  The window control unit 21B includes FFT window setting units 211-1 to 211 -N, delay units 213-2 to 213 -N, and a connection unit 214. The FFT window setting units 211-1 to 211 -N set phase differences for the signals supplied from the radio unit 10 by delay units 213-2 to 213 -N having fixed delay amounts, respectively. The received signal is input, and a signal included in a predetermined time range is extracted according to each input signal.

  The delay units 213-2 to 213 -N have a fixed delay amount with respect to the signal supplied from the radio unit 10. The delay amount to be set increases in increments of the reference delay amount τ in the order of the delay units 213-2 to 213-N. The concatenation unit 214 concatenates and outputs the results obtained by the FFT window setting units 211-1 to 211-N extracting information in a predetermined time range. The FFT window shift amount τ setting unit 50B supplies control information for setting the delay amount τ to the delay units 213-2 to 213-N.

  This configuration simplifies the control from the FFT window shift amount τ setting unit 50B because the functional units of a plurality of systems to which the delay amount τ is fixedly set process in parallel.

(Third embodiment)
Referring to the drawings, different aspects of the receiving apparatus shown in FIG. 1 are shown. FIG. 10 is a block diagram showing the window control unit in the present embodiment. The window control unit 21C and the FFT window shift amount τ setting unit 50C shown in the figure are one mode of the window control unit 21 and the FFT window shift amount τ setting unit 50 shown in FIG.

  The window control unit 21C includes a training 2 symbol buffer 215C, an FFT window shifting unit 216, and an FFT window setting unit 217. The training 2 symbol buffer 215C stores training symbols that are received in succession.

  The FFT window shifting unit 216 generates a control signal for setting an FFT processing time window so that the FFT window setting unit 217 extracts information in a predetermined time range, and supplies the control signal to the FFT window setting unit 217. Based on the set control signal, the FFT window setting unit 217 sets a time window for FFT processing so as to extract information in a predetermined time range. Then, the FFT window setting unit 217 extracts and outputs the set time window range from the accumulated training symbols. The FFT window shift amount τ setting unit 50C supplies control information indicating the FFT start time to the FFT window shift unit 216.

(Fourth embodiment)
Referring to the drawings, different aspects of the receiving apparatus shown in FIG. 1 are shown. FIG. 11 is a block diagram showing a window control unit in the present embodiment. The window control unit 21D and the FFT window shift amount τ setting unit 50D shown in the figure are one mode of the window control unit 21 and the FFT window shift amount τ setting unit 50 shown in FIG.

  The window control unit 21D includes FFT window setting units 211-1 to 211-N, delay units 213D-2 to 213D-N, a concatenation unit 214, and a training 2 symbol buffer 215. The same components as those in FIG. 9 are denoted by the same reference numerals.

  Delay devices 213D-2 to 213D-N have a fixed delay amount with respect to the signal supplied from training 2 symbol buffer 215D. The delay amount to be set increases by the reference delay amount τ in the order of the delay units 213D-2 to 213D-N. The FFT window shift amount τ setting unit 50D supplies control information for setting the delay amount τ to the delay units 213D-2 to 213D-N.

  In this configuration, training processing is performed on the training symbols temporarily buffered by the training 2 symbol buffer 215D. In the training process, as shown in the second embodiment, since the functional units of a plurality of systems in which the delay amount τ is fixedly processed are processed in parallel, the control from the FFT window shift amount τ setting unit 50D It can be simplified.

  Even with the configurations shown in the second to fourth embodiments, the same effects as the reception characteristics shown in the first embodiment can be obtained. Depending on the selected configuration, a difference in mounting form occurs, but any configuration can be selected as appropriate.

  Note that the receiving apparatus 1 shown in the present embodiment shifts the FFT window with respect to a training signal that is assigned to a received frame and continuously assigned to a unit time (T) that is two or more times the first natural number. The τ setting unit 50 sets the FFT start time to a second natural number greater than the first natural number. The detection unit 20 shifts the FFT start time based on the set FFT start time, and performs the training signal detection processing in parallel. The synthesizer performs a weighting calculation process for deriving a weighting coefficient that reduces the derived error power based on the detected training signal and a reference signal that is determined in advance according to the training signal. The signal conversion unit 40 performs demodulation processing based on the combined signal that has been weighted and combined. Here, the FFT window shift amount τ setting unit 50 sets the FFT start time to the second natural number times larger than the first natural number, thereby setting the difference between the FFT start times to a predetermined time. Can do.

  The FFT window shift amount τ setting unit 50 starts from the time when the time obtained by multiplying the unit time by the number obtained by subtracting 1 from the first natural number is divided by the number obtained by subtracting 1 from the second natural number. Set the start time. By setting the difference between the FFT start times to a constant value, it is possible to set time intervals at equal intervals. This time difference can be set as a delay amount τ that can be set by a delay device or the like, and the configuration of the repeated processing can be simplified.

  Further, the synthesis unit 30 of the present embodiment includes a phase correction unit 33 that corrects the phase of the reference signal in accordance with the FFT start time set as the starting point. Here, it is possible to generate a reference signal corresponding to the FFT start time set as the starting point.

  In addition, the detection unit 20 according to the present embodiment performs detection processing based on information on a range obtained using a window function that defines a range of information extracted from the received frame according to the set start time. As a result, a larger number of training signals can be detected than in the detection process corresponding to the symbol period, regardless of the number of training signals assigned to the received frame.

  Further, the combining unit 30 of the present embodiment uses MMSE to derive a weighting coefficient that reduces the error power. As a result, the SINR can be maximized for each subcarrier.

  As described above, in the configuration shown in the embodiment of the present invention, when the known symbol is a repetition of the same signal such as the long training symbol of IEEE802.11a, the FFT window can be set by overlapping the training symbols. , A desired signal for one cycle can be acquired and used as a pseudo training symbol to increase the number of training (update) times.

  If the fluctuation of the propagation path characteristic is small, the antenna weight is almost constant. That is, the antenna weight does not vary depending on the cutout position on the time axis. Therefore, by applying phase rotation according to the cut-out position to the known signal, the antenna weight is not phase-rotated and can be converged to a true value based on the MMSE standard.

  In addition, as shown in the present embodiment, if the training symbol is cut out to generate a pseudo training symbol, and the shift amount (delay amount τ) at the time of cutting out is too small, the number of pseudo training symbols can be increased. In some cases, a correlation is generated in the thermal noise attached to the symbol and the antenna weight deviates from the optimum value. It is desirable that the shift amount (delay amount τ) at the time of extraction is set to a value within a range in which the correlation of thermal noise does not occur, so as to earn the number of pseudo training symbols.

  In addition, this invention is not restrict | limited to the structure shown in said embodiment, A structure, quantity, etc. can be changed suitably in the range which does not change the summary of invention. For example, the number of antennas (number of elements), the number of training symbols, and the like have been shown taking specific eigenvalues as an example, but any number can be selected.

  Further, a method such as LMS or RLS can be selected as a method for realizing MMSE.

  Further, although the wireless LAN standard has been exemplified by the IEEE802.11a system, it can be applied to wireless LANs of other standards. For example, by applying to the IEEE802.11g system, the IEEE802.11n system, and the IEEE802.11p system, it is possible to obtain an effect on interference suppression.

  Further, the FFT window shift amount (delay amount) τ has been described as being equally spaced. However, the shift amount (delay amount) may be set at unequal intervals without taking a constant value. Is possible.

DESCRIPTION OF SYMBOLS 1 Receiver 10 Radio | wireless part, 11 Quadrature detection part, 12 Analog digital conversion (A / D) part,
13 Guard interval (GI) removal unit,
20 detection units, 21 window control units, 22 serial-parallel (S / P) conversion units, 23 FFT units,
30 synthesis unit, 31 synthesis processing unit, 32 reference signal unit, 33 phase correction unit,
34 subtractor, 35 weight coefficient calculation (MMSE) unit,
31a, 31b multiplier, 31c adder,
40 signal converter, 41 demodulator, 42 parallel-serial converter (P / S),
50 FFT window shift amount τ setting section

Claims (5)

A wireless communication system that performs frequency division multiplex transmission using a subcarrier frequency, and a receiving device that receives a frame to which a predetermined predetermined unit time training signal is provided,
An FFT start time is set to a second natural number greater than the first natural number with respect to the training signal assigned to the received frame and continuously assigned to the unit time that is two or more times the first natural number. A setting section to be set;
Based on the set FFT start time, a detection unit that performs the training signal detection processing in parallel by shifting the FFT start time;
Based on the detected training signal and a reference signal that is predetermined according to the training signal, a synthesis unit that performs a weighting calculation process that derives a weighting coefficient that reduces the derived error power;
A signal converter for performing demodulation processing based on the weighted synthesized signal,
A receiving apparatus comprising: The setting unit
The FFT start time is set starting from a time obtained by dividing a time obtained by multiplying the unit time by a number obtained by subtracting 1 from the first natural number by a number obtained by subtracting 1 from the second natural number. The receiving device according to claim 1. The synthesis unit is
The receiving apparatus according to claim 1, further comprising: a phase correction unit that corrects a phase of the reference signal in accordance with the set FFT start time. The detector is
The detection process is performed based on information on a range obtained using a window function that defines a range of information to be extracted from the received frame according to the set FFT start time. The receiving device according to claim 3. The synthesis unit is
The receiving apparatus according to claim 1, wherein a weighting coefficient that reduces the error power is derived using MMSE.