JP2011101297A - Ofdm demodulation apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To appropriately determine a pass band of a frequency interpolation filter part, for allowing to estimate transmission path characteristics with sufficient accuracy under radio wave environment in which many incoming waves with small power exist when a pilot signal is interpolated in the frequency direction to estimate the transmission path characteristics. <P>SOLUTION: A pass band setting part (6) which sets the pass band of the frequency interpolation filter part (7) determines the pass band so as to pass the incoming waves within the range below boundary time, and determines the boundary time so that ratio of a power integral value (B) of the incoming waves for a range exceeding the boundary time to a power integral value (A) of the incoming waves for a range below the boundary time among delay profiles estimated by a delay profile estimation part (5) is within a range below a predetermined value (THba), and becomes as large as possible. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a demodulator that demodulates a signal that has been modulated and transmitted by an OFDM (Orthogonal Frequency Division Multiplexing) modulation method.

  In terrestrial digital broadcasting systems such as ISDB-T and DVB-T, an SP (Scattered Pilot) signal, which is one of the reference pilot signals, is transmitted once every 12 carriers in the subcarrier direction and 1 every 4 symbols in the symbol direction. Inserted once. The SP signal is a “known signal” in that the content is determined by the standard when transmitted from the transmitting station side, and the amplitude and phase are known. Here, in the digital terrestrial broadcasting using the OFDM modulation system, in addition to the subcarriers directly received by the receiver antenna among the subcarriers emitted from the transmitting station, the signals are received after being reflected by the buildings or mountains. Since there are various delayed waves that are received by the antenna of the machine, as a result of these delayed waves directly interfering with the waves, each subcarrier carrying data etc. will be distorted by passing through the transmission path. . Thus, using the point that the SP signal is a known signal as described above, it is estimated how the subcarriers of the received SP signal are distorted in the transmission path, that is, the transmission path characteristics are estimated. The receiver obtains the transmission path characteristic of the subcarrier carrying the data by interpolating the transmission path characteristic of the SP signal, and performs equalization processing using the obtained estimation result of the transmission path characteristic. In general, in the estimation method of transmission path characteristics, the transmission path characteristics are estimated by performing interpolation processing in the carrier direction after interpolation processing in the symbol direction. At this time, in the interpolation process in the carrier direction, the maximum delay time in the delayed wave component in the received OFDM signal is separately calculated, and the maximum delay time in the calculation result is taken into consideration, and the filter in the appropriate passband is used in the frequency direction. Filtering is performed to remove unnecessary noise components. In the detection of the delayed wave at this time, one having a power value larger than a predetermined threshold is detected.

Japanese Patent No. 3654646 (paragraph 0029) Japanese Patent Laid-Open No. 10-257013 (paragraph 0071) JP 11-17642 A (paragraphs 0055 to 0057)

  In the conventional method, in a reception environment such as a mountainous area where a large number of reflected waves having a relatively small power value arrive, these reflected wave sets exist outside the frequency interpolation filter band and have a predetermined threshold value. If it is smaller than the power, it is removed as a noise component. As a result, the transmission path estimation result ignores the influence of these many reflected waves, and the equalization result based on the estimated transmission path characteristics deteriorates.

The present invention has been made to solve the above-described problems, and the OFDM demodulator according to the present invention provides:
A Fourier transform unit for transforming an OFDM signal into a frequency spectrum signal;
A pilot extraction unit that extracts a pilot signal from the subcarrier component output from the Fourier transform unit;
A transmission pilot generator for generating a known signal corresponding to the pilot signal;
A division unit that calculates transmission path characteristics for subcarriers used for transmission of the pilot signal by dividing the pilot signal extracted by the pilot extraction unit by the transmission pilot signal;
A time interpolation filter unit that generates a transmission line characteristic by performing interpolation in the time direction on the transmission line characteristic of the pilot signal calculated by the division unit;
A frequency interpolation filter unit that performs interpolation in the frequency direction with respect to the transmission path characteristic output from the time interpolation filter unit;
A delay profile estimation unit that estimates a delay profile based on the transmission path characteristics of the pilot signal calculated by the division unit;
A passband setting unit that sets a passband of the frequency interpolation filter unit based on the delay profile estimated by the delay profile estimation unit;
The passband setting unit determines the passband so as to pass an incoming wave in a range equal to or less than a boundary time, and the boundary time is determined from the delay profiles estimated by the delay profile estimation unit. The ratio of the power integral value of the arriving wave over the range exceeding the boundary time to the power integral value of the arriving wave over the boundary time or less range is within a predetermined value or less and as large as possible. It is characterized by defining.

  According to the present invention, it is possible to accurately estimate transmission path characteristics in a radio wave environment where there are many incoming waves with small power.

It is a block diagram which shows the structure of the OFDM demodulation apparatus in Embodiment 1 of this invention. 3 is a block diagram illustrating a configuration example of a transmission path estimation unit 104 in Embodiment 1. FIG. It is a figure which shows the structural example of the frequency interpolation filter part 7 of FIG. 6 is a diagram showing an arrangement of pilot signals of OFDM signals received by the OFDM demodulator according to Embodiment 1. FIG. (A) thru | or (c) is a figure which shows the relationship between the delay profile calculated | required by the delay profile estimation part 5 of FIG. 2, and a frequency interpolation filter. 6 is a diagram showing how to determine an integration range of power of an incoming wave in the first embodiment. FIG. 6 is a flowchart illustrating an example of a bandwidth setting process in the first embodiment. It is a figure which shows the structural example of the transmission path estimation part 104 in Embodiment 2 of this invention. It is a figure which shows the structural example of the frequency interpolation filter part 8 of FIG. FIG. 10 is a diagram illustrating how to determine an integration range of power of an incoming wave in the second embodiment. 10 is a flowchart illustrating an example of a band setting process in the second embodiment. 12 is a flowchart illustrating another example of bandwidth setting processing according to the second embodiment.

  Embodiments of the present invention will be described below with reference to the drawings.

Embodiment 1 FIG.
1 is a diagram showing a configuration of an OFDM demodulator according to Embodiment 1 of the present invention. In FIG. 1, 100 is a receiving antenna that receives broadcast radio waves, 101 is a tuner, 102 is an A / D converter circuit that converts an analog signal into a digital signal, 103 is an orthogonal demodulator, and an I-axis signal is output on its output side. A complex baseband signal of the Q-axis signal is obtained. Reference numeral 104 denotes a Fourier transform unit for converting a time signal into a frequency signal, for example, FFT (consisting of a fast Fourier transform circuit). Reference numeral 106 denotes a transmission path estimation unit that estimates the frequency characteristics of the transmission path, and reference numeral 105 denotes an equalization unit that performs equalization based on the estimation result of the transmission path estimation unit 106.

  FIG. 2 is a block diagram showing an example of the transmission path estimation unit 106 according to the first embodiment of the present invention. 2 includes a pilot extraction unit 1 that extracts an SP (scattered pilot) signal from the subcarrier component output from the Fourier transform unit 104, and a known signal (transmission) corresponding to the pilot signal. Transmission pilot generation unit 2 for generating a pilot signal), and the SP signal extracted by pilot extraction unit 1 is divided by the SP signal generated by transmission pilot generation unit 2 to transmit to the subcarrier used for transmission of the pilot signal. And a complex division unit 3 for obtaining path characteristics (transmission path characteristics for SP signals).

  The transmission path estimation unit 106 shown in FIG. 2 further includes a time interpolation filter section 4 that performs interpolation in the time direction on the transmission path characteristics of the SP signal calculated by the complex division section 3, and a time interpolation filter. A delay profile estimator 5 for calculating a delay time and power of the delay wave based on transmission path characteristics corresponding to each of the subcarrier frequency components output from the unit 4, thereby obtaining a delay profile; A passband setting unit 6 that sets a passband (passband of a filter for frequency interpolation) of the frequency interpolation filter unit 7 based on the delay profile obtained by the unit 5, and a time interpolation filter unit 4 And a frequency interpolation filter unit 7 for performing interpolation in the frequency direction (carrier direction) with respect to the output transmission path characteristics.

The passband setting unit 6 includes an integrated value calculating unit 10, a delayed wave detecting unit 11, an integrated value comparing unit 12, and a band determining unit 13.
The integral value calculation unit 10 receives the delay profile and determines an integral value of power for the integral range that is determined and changed as described later.
The delayed wave detector 11 detects a delayed wave that is equal to or greater than a predetermined threshold value THp.
The integral value comparison unit 12 compares the two integral values obtained by the integral value calculation unit 10.
The band determination unit 13 generates a signal CPB that specifies the pass band of the frequency interpolation filter unit 7 based on the comparison result of the integral value comparison unit 12.

FIG. 3 is a detailed block diagram of the frequency interpolation filter unit 7 according to the first embodiment of the present invention. The frequency interpolation filter unit 7 shown in FIG. 3 has an LPF (low-pass filter) 20 that passes a frequency component of a pass band set as described later, and a coefficient table 21 that prepares a plurality of tap coefficients of the LPF 20 as a table. And have. The pass band of the LPF 20 changes according to which of the plurality of tap coefficients stored in the coefficient table 21 is selected and supplied to the LPF 20.
Of the plurality of tap coefficients stored in the coefficient table 21, a tap coefficient for realizing the passband specified by the passband specifying signal CPB supplied from the passband setting unit 6 is selected and supplied to the LPF 20. .

  As a filter characteristic of the frequency interpolation filter unit 7, it is sufficient if there is a pass band through which an incoming wave passes. If the pass band is unnecessarily wide, unnecessary noise components also pass through the filter. Performance will be degraded. In order to avoid such a decrease in demodulation performance, it is necessary to minimize the passband of the frequency interpolation filter unit 7. The present invention is characterized in that a new method for determining the pass band of the frequency interpolation filter unit 7 is used. In this embodiment, the passband is set by selecting the coefficient table 21. Hereinafter, a method of selecting the coefficient table 21 will be described in detail.

First, the OFDM demodulator in Embodiment 1 will be described.
An OFDM-modulated radio signal (OFDM modulated signal) is received by the receiving antenna 100, and a tuner 101 converts an RF frequency signal into an intermediate frequency band signal. The intermediate frequency signal (IF) is converted into a digital signal by the A / D conversion circuit 102, and the IF signal converted into a digital signal is separated into an I-axis signal and a Q-axis signal, which are baseband signals, by the orthogonal demodulation unit 103. . Further, the Fourier transform unit 104 converts the I-axis signal and the Q-axis signal, which are time signals, into frequency components. As a result, complex data of all subcarriers in one symbol can be obtained.

  In the signal converted into the frequency component, the SP signal is arranged as shown in FIG. 4 in the Japanese terrestrial digital broadcasting system ISDB-T. In FIG. 4, the horizontal direction indicates the frequency direction, and the vertical direction indicates the time direction. The hatched circle indicates a pilot signal, and the white circle indicates subcarrier components other than the pilot signal. Also, the part surrounded by a broken line and marked “i-th symbol” indicates a subcarrier constituting one symbol (i-th symbol), surrounded by a solid line, and marked “k-th subcarrier” The portion indicates the kth subcarrier in each symbol.

  As shown in FIG. 4, the SP signal is inserted every 12 subcarriers in the frequency direction and every 4 symbols in the time direction. Therefore, in order to calculate the channel characteristics for all carriers from the channel characteristics obtained from the pilot signal, interpolation processing in the time direction and the frequency direction is required.

The pilot extraction unit 1 of the transmission path estimation unit 106 inserts a zero value for the complex data converted to the frequency axis, except for the SP signal, and extracts only the SP signal.
The complex division unit 3 divides the SP signal extracted by the pilot extraction unit 1 by the SP signal generated by the transmission pilot generation unit 2, and calculates the transmission path characteristic of the subcarrier used for transmission of the SP signal. Here, if the received SP signal is Y, the SP signal that is a known signal generated by the transmission pilot generator 2 is R, the transmission path characteristic is H, and the noise is N, the received signal Y is expressed by Expression (1). Is done.
Y = H · R + N (1)

Accordingly, the transmission path characteristics are estimated by the equation (2).
H = Y / R-N / R (2)

Furthermore, when the noise N is small,
H = Y / R
Can be estimated as The transmission path characteristics H calculated at this time are for the subcarriers used in the transmission of the SP signal, and the subcarriers of the data signal have unknown transmission path characteristics. Therefore, the time interpolation filter and the frequency interpolation filter The transmission path characteristics are obtained by interpolation by the interpolation processing according to. The data signal is equalized based on the estimation result of the transmission path characteristic obtained by interpolation in this way.

  First, the transmission path characteristic is interpolated in the time direction, and the delay profile estimation unit 5 outputs a delay profile based on the time interpolation result. That is, the time interpolation filter unit 4 interpolates the transmission channel characteristics of the pilot signal calculated by the division unit 3 in the time direction to generate transmission channel characteristics for all symbols, and delay profile estimation The unit 5 estimates delay profiles for all subcarriers based on the transmission path characteristics output from the time interpolation filter unit 4. Assume that an estimation result as shown in FIG. The horizontal axis represents delay time, and the vertical axis represents power.

FIG. 5A shows a case where there is no delayed wave.
The bandwidth of the frequency interpolation filter unit 7 is determined in advance so as to take one of a plurality of stepwise or discrete values corresponding to the delay time range of the incoming wave.
In the example shown in FIG. 5A, a filter whose pass band is a band corresponding to the range DR0 of delay time 0 to τ0, a filter whose pass band is a band corresponding to the range DR1 of delay time 0 to τ1, and the delay time A filter whose pass band is a band corresponding to the range DR2 of 0 to τ2, a filter whose pass band is a band corresponding to the range DR3 of the delay time 0 to τ3, and a band corresponding to the range DR4 of the delay time 0 to τ4 It is configured to be able to act as a filter having a total of five different (passbands) and a filter as a band, and as one of the five filters depending on the detected delay time range Whether to act is selected. Which filter to act on is selected by the selection of filter coefficients.

As shown in FIG. 5B, assuming that a delayed wave having a power larger than a certain threshold value THp is detected, the bandwidth corresponding to the range including the delay time τa of the detected delayed wave is the conventional method. The filter is selected and the noise component is removed.
However, in such a method, as indicated by hatching in FIG. 5C, when there are a large number of delayed waves whose power is equal to or lower than the threshold value THp, these components are also removed, resulting in an error. The transmission path is estimated.
If the delay wave component in the shaded area is picked up by lowering the threshold value THp, a filter with a wide band is selected unnecessarily, and the noise component may be erroneously detected as a delay wave. .

  In order to solve such a problem, in the first embodiment, a range including a delayed wave and a range including only a noise component are determined based on an integral value in the delay time direction of incoming wave power. That is, the integrated value in the range from zero delay time to a certain delay time (boundary candidate time) is compared with the integrated value in the range exceeding the boundary candidate time, and the ratio of the latter to the former is a predetermined value (predetermined value). The boundary candidate time is determined as the boundary time so as to be equal to or less than the threshold value), and the band corresponding to the range equal to or less than the boundary time is determined as the pass band. Specifically, while gradually increasing the boundary candidate time, search for a time when the ratio is less than or equal to the predetermined value, determine the found time as the boundary time, and pass the band corresponding to the range up to the boundary time And decide.

  The boundary candidate time may be gradually increased from zero. For example, when a delay wave having a predetermined threshold value THp or more is detected, the delay time τa of the delay wave or a delay time longer than that (for example, a discrete wave in advance) One of a plurality of delay times determined in (1) may be used as a search start point.

In the following, the range in which the incoming wave can appear (the range where the possibility of arrival of delay cannot be ignored) is set as the analysis target, divided into a plurality of sections in advance, and one of the division points is set as the initial boundary candidate time (boundary A case where the initial value of the candidate time is used will be described. For example, as shown in FIG. 6, the upper limit of the range of delay time that cannot be ignored is τ4, and the range from delay time 0 to upper limit τ4 is divided into five sections Sq (q = 0 to 4). The subscript q of the code Sq representing the section is added in order from the delay time 0 side, and is also referred to as a section number. The division points between the sections are indicated by symbols τ0, τ1, τ2, and τ3. The widths of the sections S _{0 to} S ₄ may be equal to each other, but may be different from each other. For example, the width of the section may be narrowed as the delay time is shorter.

  As will be described in detail below, one of the division points τ0 to τ3 is used as the boundary candidate time. When a delay wave equal to or greater than the predetermined threshold THp is detected, the upper limit division point (upper limit end of the section) including the delay time of the delay wave is set as the boundary time search start point.

In Figure 6, or more delayed wave threshold THp is detected, the delay time of the delay wave is .tau.a, search start point division points τ1 upper side of the section S ₁ including a delay time .tau.a (first boundary candidate time ).

In this case, first, an integral value (first integral value) of power in the range of _{0 to} τ1 (range consisting of the sections S ₀ and S ₁ ) is set with the upper limit dividing point τ ₁ of the section S ₁ as the boundary candidate time τx. ) And the integral value (second integral value) of the power in the range of τ1 to τ4 (range consisting of the sections S ₂ , S ₃ , S ₄ ) by the integral value calculation unit 10, The integrated value comparison unit 21 compares these integrated values. That is, the first obtains the ratio (integration ratio) of the second integral value to the integral value, and determines whether the integral ratio is larger than a predetermined value. That is, when the first integral value is A, the second integral value is B, and the integral ratio is RTba,
RTba = B / A (3)
And whether or not the integration ratio RTba is larger than a predetermined value THba is determined.

The integration ratio RTba is larger than the predetermined value THba, that is,
RTba = B / A> THba
In this case, it means that there are a large number of delayed waves in the delay time sections (S ₂ , S ₃ , S ₄ ) longer than the boundary candidate time (τx = τ1) used for calculating the integral value. To do.

In this case, the band determination unit 13 issues an instruction to change the boundary candidate time τx (and thus change the integration range) to the integration value calculation unit 10. Specifically, an instruction to select as a step greater division point (upper side end portion of the section S ₂ of the lower end of .tau.1) a is τ2 boundary candidate time τx than .tau.1. In response to this, the integral value calculation unit 10 obtains an integral value in a range of _{0 to} τ2 (range consisting of sections S ₀ , S ₁ , S ₂ ) as a new first integral value A, and τ 2 to τ 4 The integral value of the range (range consisting of sections S ₃ and S ₄ ) is obtained as a new second integral value B, and the ratio (integration ratio) RTba of the second integral value B to the first integral value A is a predetermined value. It is determined whether or not it is greater than THba.

Also at this time, if it is determined that the integration ratio is greater than the predetermined value THba (RTba = B / A> THba), the band determination unit 13 instructs to change the boundary candidate time τx (and thus change the integration range). Is output to the integral value calculation unit 10. Specifically, an instruction to select as a step greater division point (upper side end portion of the section S ₃ of the lower end of .tau.2) a is τ3 boundary candidate time τx than .tau.2. In response to this, the integral value calculation unit 10 obtains the integral value in the range of _{0 to} τ3 (range consisting of the sections S ₀ , S ₁ , S ₂ , S ₃ ) as the new first integral value A, and τ 3 obtains the integral value of the range of ～Tau4 (range consisting section S ₄₎ as the second integral value B new, the ratio of the second integrated value B for the first integrated value a (integration ratio) RTba predetermined value It is determined whether or not it is greater than THba.

At this time, when it is determined that the integration ratio is equal to or less than the predetermined value THba, the band determination unit 13 determines the boundary candidate time τx = τ3 at this time as the boundary time τc, and a range (interval S _{0) that is} equal to or less than the boundary time τc. , S ₁ , S ₂ , S ₃ ) is determined to be a pass band, and a signal (pass band designation signal) CPB indicating the determination result is supplied to the frequency interpolation filter unit 7. To do.

  In the frequency interpolation filter unit 7, the band determination unit 13 selects the tap coefficient for realizing the pass band specified by the passband specification signal CPB from the band determination unit 13. CPB is output.

  As described above, in the above embodiment, when the boundary candidate time is gradually increased, the boundary time in which the integration ratio RTba is smaller than the predetermined value THba is searched for, and the filter bandwidth (passband) of the LPF 10 is determined. .

As described above, as a method of obtaining the integration value in each integration range while gradually moving the boundary candidate time τx, integration over each integration range is performed every time the boundary candidate time τx is changed. Although it is good, an integral value is obtained and stored for each of the sections S _{0 to} S ₄ , and every time the integration range is changed, the sum of the integration values of the section belonging to the changed integration range is obtained, It may also be used as an integral value in the integral range. In this case, the integral value of the interval added by the change is added to the sum of the integral values of the integration range before the change, or the integral value of the interval reduced by the change is subtracted from the sum of the integral values of the integration range before the change. By doing so, you may obtain | require the sum total of the integrated value of the integration range after a change.
Note that “sum” is not necessarily the sum of a plurality of integral values, and there is only one section included in the integration range, and therefore a single integral value itself may be “sum”.

  The above operation will be described in a more general manner with reference to the flowchart of FIG. In the following description, it is assumed that the integral value for each section is obtained and stored, and the sum of the integral values for each section belonging to a new integral range is obtained every time the integral range changes.

First, the integral value calculation unit 10 obtains power integration values T _{0 to} T ₄ for each of the sections S _{0 to} S ₄ and stores them in the memory 10m in the integral value calculation unit 10 (ST11). When this process ends, the bandwidth determination unit 13 is notified of the end.
In parallel with this, the delay wave detection unit 11 detects a delay wave equal to or greater than the threshold THp, and informs the band determination unit 13 of information (section number) indicating the section to which the peak delay time τa belongs.
In the band determination unit 13, when the delay wave detection unit 11 detects a delay wave equal to or greater than the threshold value THp, and transmits information (section number) indicating the section to which the delay time τa belongs (ST12), the section number Is set to the value of the parameter p (ST13), and if it is not a layer, 0 is set to the value of the parameter p (ST14), and the integral value calculation unit 10 is notified.

  Note that when the delay wave equal to or greater than the predetermined threshold is not detected and the division point τ0 closest to 0 is set as the first boundary candidate time, steps ST12 and ST13 are omitted, and step ST11 is followed by step ST14. Just go ahead.

を求め、同時に区間Ｓ_ｐ＋１からＳ_４までの積分値Ｔ_ｐ＋１〜Ｔ_４の総和

を求めて出力する（ＳＴ１５）。 In the integral value calculation unit 10, the sum of integral values T ₀ = T _p from the sections S ₀ to S _p

At the same time, the sum of the integral values T _{p + 1 to} T ₄ from the interval S _{p + 1} to S ₄

Is obtained and output (ST15).

In the integral value comparison unit 12, the ratio Bp / Ap between the integral values Ap and Bp output from the integral value calculation unit 10 is larger than a predetermined value THba, that is, Bp / Ap> THba.
Is satisfied (ST16), and the determination result is notified to the band determination unit 13.

  If the determination result is greater than the predetermined value THba, the band determination unit 13 increases p by 1 (ST17), and notifies the integrated value calculation unit 10 of the increased p value. Thereafter, the process of step ST15 is performed again.

If it is determined in step ST16 is equal to or less than the predetermined value THba, the process proceeds to step ST18, the band determining section 13, the section _S 0 to S _p to determine a pass band of band corresponding signal indicating the determined passband (Passband signal) CPB is output.

As described above, the passband setting unit 6 determines the passband of the frequency interpolation filter unit 7 as a passband corresponding to a range equal to or less than the boundary time (τc).
Among the delay profiles obtained by the delay profile estimation unit 5, the boundary time is
For the integrated power A of the incoming wave over the boundary time (τc) or less,
It is determined that the ratio of the power integral value B of the incoming wave over the range exceeding the boundary time (τc) is within the range of the predetermined value THba and as large as possible.
Here, “as much as possible” is due to the restriction that the possible value of the boundary candidate time is discrete, and is maximized when the boundary candidate time is changed within the range of possible values. Means value.

  More specifically, the band determination unit 13 of the passband setting unit 6 first sets a time (τ1 in the illustrated example) longer than the delay time τa of the delay wave equal to or greater than the threshold as the boundary candidate time τx, or relatively The dividing point τ0 closest to 0 is set as the boundary candidate time τx, for example, and the integral value calculation unit 10 of the passband setting unit 6 sets the power integration value of the incoming wave over the range equal to or less than the boundary candidate time τx as the first integration. While obtaining as the value A, the integral value of the incoming wave over the range exceeding the boundary candidate time τx is obtained as the second integral value B, and the integral value comparison unit 12 calculates the second integral value with respect to the first integral value A. It is determined whether the ratio of B is greater than the predetermined value THba, and the band determination unit 13 increases the boundary candidate time τx every time it is determined that the ratio (B / A) is greater than the predetermined value THba. Integrated value calculation unit The new first integral value A and the new second integral value B are calculated at the boundary candidate time τx increased to 10, and the integral value comparison unit 12 performs the determination, and the ratio ( When B / A) becomes equal to or less than the predetermined value THba, the boundary candidate time τx at that time is set as the boundary time τc, and the band corresponding to the range up to the boundary time τc is determined as the pass band. A signal indicating the pass band (pass band signal) CPB is output.

Based on the pass band designation signal CPB output from the band judgment unit 13, the coefficient table 21 in the frequency interpolation filter unit 7 selects a tap coefficient for realizing the pass band designated by the band designation signal CPB. And output to the LPF 10.
By the above processing, proper filtering that can originally remove the delayed wave component and remove unnecessary noise components is realized.

  In the above example, the integrated power value of the incoming wave over the range below the boundary candidate time is the first integrated value A, and the integrated power value of the incoming wave over the range beyond the boundary candidate time is the second integrated value B. While the boundary candidate time is gradually lengthened, a point where the ratio RTba of the second integral value B to the first integral value A is equal to or smaller than the predetermined value THba is searched for. The power integral value over the range to be analyzed) is obtained as the first integral value C, the power integral value over the range exceeding the boundary candidate time is obtained as the second integral value B, and the first integral value C remains fixed. While gradually increasing the boundary candidate time (while narrowing the range for obtaining the second integral value B), the ratio of the second integral value B to the first integral value C, that is, the integral ratio RTbc = B / C And the integration ratio RTb Is searched for a point (boundary candidate time) that is equal to or less than the predetermined value THbc, the boundary candidate time found in this way is adopted as the boundary time, and a band corresponding to the range of the boundary time or less is determined as the pass band. Even in such a configuration, it is possible to achieve appropriate filtering that leaves the originally required delayed wave component and can remove unnecessary noise components, as in the above example.

  In the above example, while the boundary candidate time is gradually increased, the state where the ratio RTba of the integral value B to the integral value A is not more than a predetermined value and is as large as possible is found. On the contrary, it is also possible to search for a state in which the condition that the ratio RTba of the integral value B to the integral value A is not more than a predetermined value and is as large as possible is satisfied while the boundary candidate time is gradually decreased. In this case, the band determination unit 13 of the passband setting unit 6 first sets the boundary candidate time τx as a relatively large value, for example, the division point τ3 closest to the upper limit, and the integral value A in a range equal to or less than the boundary candidate time τx. When the ratio of the integral value B in the range exceeding the boundary candidate time τx is less than or equal to the predetermined value THba, it is determined whenever the ratio B / A is determined to be less than or equal to the predetermined value THba. When the time τx is decreased and the ratio B / A of the new integral value B to the new integral value A becomes larger than the predetermined value THba, the band corresponding to the range up to the boundary candidate time before the last decrease is obtained. The signal is determined as a pass band, and a signal (pass band signal) CPB indicating the determined pass band is output.

  Similarly, a range (all analysis target range) integral value C in the interval 0 to τ4 is obtained, and the ratio of the integral value B in the range exceeding the boundary candidate time τx to the integral value C while gradually shortening the boundary candidate time τx. A state in which the condition that B / C is equal to or smaller than the predetermined value THbc and is as large as possible may be found.

  Furthermore, comparing the ratio B / A of the two integrated values A and B with the predetermined value THba is equivalent to comparing the ratio A / B, which is the reciprocal thereof, with the predetermined value THab (= 1 / THba). There are modifications based on such considerations and are also within the scope of the present invention. The same applies to the ratio of the integral values B and C.

  In the first embodiment, the integral value is calculated by the integral value calculator 10 using the delay profile estimated by the delay profile estimator 5 as it is. However, the delay profile estimator 5 determines the predetermined number of times. The estimation results are averaged, an averaged delay profile is output, and the integral value calculation unit 10 integrates in the delay time direction based on the averaged delay profile output from the delay profile estimation unit 5 It is also possible to calculate the value, and in this way, unnecessary noise components included in the delay profile estimation result can be suppressed in advance, and the passband can be determined with high accuracy.

Embodiment 2. FIG.
In the first embodiment, the frequency interpolation filter unit 7 including the LPF is used. However, instead of the configuration of the first embodiment, the second embodiment uses the LPF and the BPF as the frequency interpolation filter unit 7. A combination of (bandpass filters) is used. Hereinafter, the present embodiment will be described with reference to FIGS.

  Although the overall configuration of the OFDM demodulator according to the second embodiment is as shown in FIG. 1, the configuration of the transmission path estimation unit 106 is different. FIG. 8 is a block diagram illustrating a configuration example of the transmission path estimation unit 106 in the present embodiment. The circuits denoted by reference numerals 1 to 5 in FIG. 8 are the same as those in the first embodiment, and thus description thereof is omitted.

  Unlike the one shown in FIG. 2, the passband setting unit 6 includes an integration value calculation unit 40, an extraction / addition unit 41, an integration value comparison unit 42, and a band determination unit 43. Further, a frequency interpolation filter unit 8 is provided instead of the frequency interpolation filter unit 7 of FIG.

  FIG. 9 is a diagram illustrating a configuration of the frequency interpolation filter unit 8. The illustrated frequency interpolation filter unit 8 includes an LPF 30, a LPF coefficient table 31 prepared by using a plurality of tap coefficients of the LPF 30, a BPF 32, and a BPF coefficient table 33 prepared by using a plurality of tap coefficients of the BPF 32 as a table, An adder 34 for adding the result of the LPF 30 and the result of the BPF 32 is provided.

It is assumed that a delay profile as shown in FIG. 10 is obtained by the estimation by the delay profile estimation unit 5.
In FIG. 10, sections S _{0 to} S ₄ are formed by dividing a range (analysis target range) in which an incoming wave can appear, as described in the first embodiment, and are continuous with each other. ing.

The integral value calculation unit 40 calculates the power integral values T ₀ , T ₁ , T ₂ , T ₃ , T ₄ of the incoming waves in each of the sections S ₀ , S ₁ , S ₂ , S ₃ , S _4. And stored in the internal memory 40m.

The extraction and addition unit 41 divides the plurality of sections S _{0 to} S ₄ into a first set Gd composed of a relatively small power integrated value and a second set Ge composed of a relatively large power integrated value. The sum of the power integral values of the section belonging to the first set Gd is obtained as the first sum D, and the sum of the power integral values of the section belonging to the second set Ge is obtained as the second sum E. The above division is performed such that the maximum power integral value in the section belonging to the first set Gd is smaller than the minimum power integral value Ge in the section belonging to the second set Ge. Further, the division is performed so that the ratio of the first sum D to the second sum E is within a range equal to or less than a predetermined value THde and as large as possible. The integral value comparison unit 42 determines whether the ratio D / E of the first sum D to the second sum E is within a predetermined value or less.

In the process of reaching the state where the condition “the ratio D / E of the first sum D to the second sum E is within the predetermined value THde and as large as possible” is satisfied,
The section belonging to the first set Gd is gradually increased (accordingly, the section belonging to the second set Ge is gradually decreased), or the section belonging to the second set Ge is gradually increased (according to this, the first set The section belonging to the set Gd is gradually decreased), and the calculation of the first sum D and the second sum E and the comparison with the predetermined value THde are repeated.

  When the above-mentioned condition “the ratio D / E of the first sum D to the second sum E is within the predetermined value THde and as large as possible” is satisfied. At that time, a band corresponding to a set of sections belonging to the second set Ge is determined as a pass band of the frequency interpolation filter unit 8, and a signal (band specifying signal) CPB specifying the determined pass band is included in the frequency. This is supplied to the insertion filter unit 8. In the above processing, the calculation of the sums D and E, the control for comparison with the predetermined value THde, the output of the band designation signal CPB, and the like are performed by the band determination unit 43.

  Hereinafter, the case where the section belonging to the first set Gd is gradually increased will be described in detail.

When the integral value calculation unit 40 obtains the power integration values T _{0 to} T ₄ of the sections S _{0 to} S ₄ and stores them in the memory 40m, the extraction / addition unit 41 determines the smallest value among these integration values. The integral value of the section belonging to one set Gd is extracted, the other integral value is extracted as the integral value of the section belonging to the second set Ge, and the integral values of the respective sets are added. In the example shown in FIG. 10, if it is determined that the section S ₁ has the minimum integral value, the extraction adder 41 converts the integral value of the section S ₁ into the sum of the integral values of the sections belonging to the first group Gd ( First sum) D, and the sum of the integral values of the sections S ₀ , S ₂ , S ₃ , S _{4 is obtained} as the sum of the integral values (second sum) E of the sections belonging to the second set Ge. . The ratio (integration ratio) of the first sum D to the second sum E
RTde = D / E (4)
Is calculated.

If the calculated integral ratio RTde is less than a predetermined value THde, segment S ₁ is band determination unit 43 determines the are removable section by the filter of the frequency interpolation filter unit 8. Then, the extraction adding unit 41 is instructed to shift the section having the smallest integral value next to the integral value of the section S ₁ from the second set Ge to the first set Gd.

The extraction / addition unit 41 detects a section where the integral value is the smallest among the four sections belonging to the second set Ge. In the example of FIG. 10, if the integral value of the section S ₄ is determined to be the smallest among the four sections belonging to the second set of Ge, the section S ₄ from the second set Ge to a first set Gd The sum of integrated values (first sum) D of the sections S ₁ and S ₄ belonging to the first set Gd and the integrated values of the sections S ₀ , S ₂ and S ₃ belonging to the second set Ge The sum (second sum) E is calculated, and the ratio (integration ratio) RTde = D / E between the two is calculated. If the calculated integral ratio RTde is less than a predetermined value THde, band determination unit 43 determines the even section S ₄ is a removable section similar to the filter. Then, to move the next smallest in a section integral value of the section S ₄ from the second set Ge to a first set Gd, band determination unit 43 instructs the extraction addition unit 41.

The extraction / addition unit 41 detects a section where the integral value is the smallest among the three sections belonging to the second set Ge. In the example of FIG. 10, if the integral value of the section S ₃ is determined to be the smallest among the three sections belonging to the second set Ge, section S ₃ of the second set Ge from the first set Gd , The sum (first sum) D of the integrated values of the sections S ₁ , S ₃ , and S ₄ belonging to the first set Gd and the two sections S ₀ and S ₂ belonging to the second set Ge The sum (second sum) E is calculated, and the ratio (integration ratio) RTde = D / E between the two is calculated. If the calculated integral ratio RTde is greater than a predetermined value THde, section S ₃ is intended a band determining section 43 not to be removed by the filter is determined, returning the segment S ₃ to a second set Ge.

Then, based on the above determination result, the band determination unit 43 determines a band corresponding to a set of sections S ₀ , S ₂ , and S ₃ (sections belonging to the second set Ge) other than the sections S ₁ and S _4. A signal (band designation signal) CPB instructing to set the pass band is supplied to the frequency interpolation filter unit 8.

The above operation will be described using the flowchart of FIG. 11 in a more general form.
First, the integral value calculating unit 40, for each segment _S 0 to S ₄ obtains a power integral value _T 0 through T _4, the storing in the memory 40m of the integral value calculating unit 40 (ST21).
In the next integral value calculating section 40, arranged integral value _T 0 through T ₄ sections _S 0 to S ₄ in ascending order of value, in ascending order a sign of _{U q (q = 0~4) (} ST22).

  When this processing is finished, the bandwidth determination unit 43 is notified of the end. The band determination unit 43 sets the parameter r to 0 and notifies the extraction / addition unit 41 (ST23).

を求め、同時に積分値Ｕ_ｒ＋１〜Ｕ_４の総和

を求めて出力する（ＳＴ２４）。 In the extraction / addition unit 41, the total sum of the integral values U _{0 to} U _r

At the same time, the sum of the integral values U _{r + 1 to} U ₄

Is obtained and output (ST24).

In the integral value comparison unit 42, the ratio Dr / Er of the integral values Dr and Er output from the extraction / addition unit 41 is equal to or less than a predetermined value THde, that is, Dr / Er ≦ THde (5)
Is determined (ST25), and the band determination unit 43 is notified of the determination result.
If the determination result indicates that it is equal to or less than the predetermined value THde, the band determination unit 43 increments r by 1 (ST26), and informs the extraction / addition unit 41 of the increased value of r. Thereafter, the process of step ST24 is performed again.

If it is determined in step ST25 that it is larger than the predetermined value THde, the process proceeds to step ST27, where the band determining unit 43 determines the band corresponding to the section having the integral values U _{0 to} U _r−1 as the pass band. (Band designation signal) CPB is output.

As described above, in the above example, the passband setting unit 6 assumes that the one with the smallest power integral value among the plurality of sections S _{0 to} S ₄ belongs to the first set Gd, and the other sections. Each time the integration ratio RTde = D / E is determined to be less than or equal to the predetermined value THde, the power integral value in the second group Ge is the smallest. Is repeated until it is determined that the integration ratio RTde is greater than the predetermined value THde. When it is determined that the integration ratio RTde is greater than the predetermined value THde, The section moved from the second set Ge to the first set Gd (the last section moved from the second set Ge to the first set Gd) is returned to the second set Ge, and the second set is returned to the second set Ge. The band corresponding to the set of sections belonging to Ge is determined as the pass band. And a signal indicating the determined pass band (passband signal) and outputs the CPB.

  The band designation signal CPB output from the band determination unit 43 designates a frequency band that the frequency interpolation filter unit 8 should pass. In the second embodiment, unlike the first embodiment, the case where the passband is dispersed into two or more groups is considered. In the frequency interpolation filter unit 8, the delay time is closer to zero. The LPF is applied and the BPF is applied to other sections. In the example of FIG. 10, the LPF 30 is used to pass incoming waves in the range of 0 to τ0, and the BPF 32 is used to pass incoming waves in the range of τ1 to τ3.

  The band designation signal CPB output from the band determination unit 43 is a signal (tap coefficient selection signal) CPBa for selecting a tap coefficient for allowing the LPF 30 to have a desired pass band, and a desired pass band for the BPF 32. And a signal (tap coefficient selection signal) CPBb for selecting a tap coefficient.

The LPF coefficient table 31 of the frequency interpolation filter unit 8 selects a tap coefficient according to the selection signal CPBa and supplies it to the LPF 30, and the BPF coefficient table 33 selects a tap coefficient according to the selection signal CPBa and supplies it to the BPF 32. Supply.
The tap coefficient supplied to the LPF 30 is a tap coefficient for causing the LPF 30 to act as a filter having a section (selected section) of 0 to τ0 as a pass band, and the tap coefficient supplied to the BPF 32 is obtained by converting the BPF 32 to τ1. It is a tap coefficient for making it act as a filter which uses the area of -tau3 as a pass band.
As a result, it is possible to achieve proper filtering that leaves the originally required delayed wave component and that can remove unnecessary noise components.

As described above, when the frequency interpolation filter unit 8 is composed of one LPF and one BPF, the sections belonging to the second set Ge are two or less groups composed of sections that are continuous with each other. Must be configured. For example, when there are three sections belonging to the second set Ge and they are not continuous with each other and are divided into three groups (for example, the sections S ₀ , S ₂ , S ₄ belong to the second set Ge. and if), located between the sections belonging to the second set Ge, a section of the first set Gd, transferred to a second set Ge (e.g., transferred to section S ₃ to the second set Ge, or transferred to section S ₁ to the second set Ge), thereby, the number of the group consisting of consecutive intervals is so 2 becomes bract. Such processing is also performed by the band determination unit 43.

  Furthermore, as shown in FIG. 10, when a delay wave of a predetermined threshold value THp or more is detected, a section including the delay time τa of the delay wave is given priority in addition to the order of the magnitude of the integral value. It is good also as putting in a 2nd group and making a corresponding zone | band become (one part) pass band.

In the above example, the sum total of the integral values of the section belonging to the first group Gd composed of the relatively small integral value is defined as the first sum D, and the second group composed of the relatively large integral value. The ratio RTde of the first sum D to the second sum E is set while the sum of the integral values of the section belonging to Ge is set as the second sum E, and the section belonging to the second set Ge is moved to the first set Gd. A state in which the value is larger than the predetermined value THde is searched for. The power integrated value over the range from _{0 to} τ4 (analysis target range), that is, the total sum of the integrated values in the intervals S _{0 to} S ₄ is calculated as the second total C. The second sum C of the first sum D is obtained by updating the first sum D while increasing the number of sections belonging to the first set Gd while the second sum C is fixed. To find a state in which the ratio to is greater than the predetermined value THdc Immediately before entering such a state, the section added to the first set Gd (the section added last to the first set Gd) is removed from the first set Gd, whereby the integration ratio RTdc becomes a predetermined value. The condition that it is equal to or less than THdc is satisfied, and a state in which the largest number of sections belong to the first group Gd is reached. In this state, a band corresponding to a section other than the section belonging to the first group Gd Even in such a case, as in the above example, appropriate filtering can be realized that leaves the originally required delayed wave component and can remove unnecessary noise components.

  In the example described with reference to FIG. 11, the sections are extracted in ascending order of the power integration value, and the sections belonging to the first set Gd are gradually increased. , And the number of sections belonging to the second set Ge may be gradually increased. Hereinafter, processing in this case will be described.

When the integral value calculation unit 40 obtains the power integration values T _{0 to} T ₄ of the sections S _{0 to} S ₄ and stores them in the memory 40m, the extraction / addition unit 41 sets the largest value among these integration values. The integration values of the sections belonging to the second set Ge are extracted, the other integration values are extracted as the integration values of the sections belonging to the first set Gd, and the integration values of the respective sets are added. In the example shown in FIG. 10, if it is determined that the section S ₀ has the maximum integral value, the extraction adder 41 _converts the integral value of the section S ₀ into the sum of the integral values of the sections belonging to the second set Ge ( 2nd sum) E, and the sum of the integral values of the sections S ₁ , S ₂ , S ₃ , S _{4 is obtained} as the sum of the integral values (first sum) D of the sections belonging to the first group Gd. . The ratio (integration ratio) of the first sum D to the second sum E
RTde = D / E
Is calculated.

If the calculated integral ratio RTde is greater than a predetermined value THde, band determination unit 43, the intervals with the integrated value becomes maximum in the next integral value of the interval S ₀ from the first set Gd to a second set Ge An instruction is given to the extraction / addition unit 41 to shift.

The extraction / addition unit 41 detects a section where the integral value is maximum among the four sections remaining in the first set Gd. In the example of FIG. 10, when it is determined that the integral value of the section S ₂ is the maximum among the sections remaining in the first set Gd, the section S ₂ is changed from the first set Gd to the second set Ge. The total sum (first sum) D of the integrated values of the sections S ₁ , S ₃ and S ₄ belonging to the first set Gd, and the sections S ₀ and S ₂ belonging to the second set Ge The total sum (second sum) E of the integral values is calculated, and the ratio (integration ratio) RTde = D / E of both is calculated. When the calculated integration ratio RTde is larger than the predetermined value THde, the band determination unit 43 determines the section having the maximum integral value from the first group Gd to the second group among the sections remaining in the first group Gd. The band determination unit 43 instructs the extraction / addition unit 41 to shift to Ge.

The extraction / addition unit 41 detects a section having the maximum integral value among the three sections remaining in the first set Gd. In the example of FIG. 10, when it is determined that the integral value of the section S ₃ is the maximum among the sections remaining in the first set Gd, the section S ₃ is changed from the first set Gd to the second set Ge. , And the total sum (first sum) D of the integrated values of the sections S ₁ and S ₄ belonging to the first set Gd and the two sections S ₀ and S ₂ belonging to the second set Ge. , And S ₃ (second sum) E is calculated, and a ratio (integration ratio) RTde = D / E between them is calculated. When the calculated integration ratio RTde is less than or equal to the predetermined value THde, the band determination unit 43 corresponds to a set of _three sections S ₀ , S ₂ , and S ₃ belonging to the second set Ge at that time. A signal (band designation signal) CPB for instructing the band to be a pass band is supplied to the frequency interpolation filter unit 8.

Hereinafter, the above operation will be described in a more general manner with reference to the flowchart of FIG.
First, in the same manner as described with reference to FIG. 11, the integral value calculation unit 40 obtains power integration values T _{0 to} T ₄ for each of the sections S _{0 to} S ₄ , and the memory 40 m in the integral value calculation unit 40. (ST21).
In the next integral value calculating section 40, arranged integral value _T 0 through T ₄ sections _S 0 to S ₄ in the order of decreasing, a sign of _V q (q = _0~4) in the ascending order (ST32).

  When this processing is finished, the bandwidth determination unit 43 is notified of the end. The band determination unit 43 sets the parameter r to 0 and notifies the extraction / addition unit 41 (ST23).

を求め、同時に積分値Ｖ_ｒ＋１〜Ｖ_４の総和

を求めて出力する（ＳＴ３４）。 The extraction addition unit 41, the sum of the integrated value _V 0 ~V _r

At the same time, the sum of the integral values V _{r + 1 to} V ₄

Is obtained and output (ST34).

In the integral value comparison unit 42, the ratio Dr / Er of the integral values Dr and Er output from the extraction / addition unit 41 is greater than a predetermined value THde, that is,
Dr / Er> THde
Is satisfied (ST35), and the determination result is notified to the band determination unit 43.
If the determination result is greater than the predetermined value THde, the band determination unit 43 increases r by 1 (ST26), and informs the extraction / addition unit 41 of the increased r value. Thereafter, the process of step ST34 is performed again.

If it is determined in step ST35 that the value is equal to or less than the predetermined value THde, the process proceeds to step ST37, where the band determination unit 43 determines the band corresponding to the section having the integral values V _{0 to} V _r as the pass band, and indicates this. A signal (band designation signal) CPB is output.

As described above, the passband setting unit 6 sets the one having the largest power integration value among the plurality of sections S _{0 to} S ₄ as belonging to the second group Ge, and sets the other sections as the first section. Each time it is determined that the integration ratio RTde = D / E is larger than the predetermined value THde, the power integration value in the section of the first group Gd is the second one that belongs to the group Gd. The process of moving to the set Ge is repeated until it is determined that the integration ratio RTde is equal to or less than the predetermined value THde. When it is determined that the integration ratio RTde is equal to or less than the predetermined value THde, the process belongs to the second set Ge. A band corresponding to the set of sections is determined as a pass band, and a signal (pass band signal) CPB indicating the determined pass band is output.

  In the above example, the sum total of the integral values of the sections belonging to the first set Gd (first sum) is transferred to the second set Ge in descending order of the sections belonging to the first set Gd. A state in which the ratio RTde to the total sum (second total) E of the intervals belonging to the second set Ge in D is equal to or less than a predetermined value THde is searched for. A state in which the ratio of the second sum E to the first sum C is larger than the predetermined value THec while obtaining the power integral value over the target range) as the first sum C and increasing the section belonging to the second set Ge. Thus, the condition that the integration ratio RTec is equal to or greater than the predetermined value THec is satisfied, and a state in which the minimum interval belongs to the second group Ge is reached, and in this state, the second group Ge Corresponds to the section to which it belongs May also possible to determine the pass band, even in this way, leaving a likewise originally required delayed wave component and the above example, proper filtering is achieved capable of removing yet unnecessary noise components.

  Furthermore, comparing the ratio D / E of the two integral values D and E with the predetermined value THde is equivalent to comparing the ratio E / D, which is the reciprocal thereof, with the predetermined value THed (= 1 / THde). There are modifications based on such considerations and are also within the scope of the present invention. The same applies to the ratio of the integrated values D and C.

  In the second embodiment, the integral value is calculated by the integral value calculator 40 using the delay profile estimated by the delay profile estimator 5 as it is. However, the delay profile estimator 5 determines a predetermined number of times. The estimation results are averaged, an averaged delay profile is output, and the integration value calculation unit 40 integrates in the delay time direction based on the averaged delay profile output from the delay profile estimation unit 5 It is also possible to calculate the value, and in this way, unnecessary noise components included in the delay profile estimation result can be suppressed in advance, and the passband can be determined with high accuracy.

  In the second embodiment, the frequency interpolation filter unit 8 is composed of one LPF and one BPF. However, a plurality of BPFs are prepared according to the number of sections, and all filter output results are added and transmitted. Similarly, even if the path characteristics are estimated, appropriate delaying wave components can be left and appropriate filtering that can remove unnecessary noise components can be realized.

  DESCRIPTION OF SYMBOLS 1 Pilot extraction part, 2 Transmission pilot generation part, 3 Complex division part, 4 Time interpolation filter part, 5 Delay profile estimation part, 6 Band setting part, 7 Frequency interpolation filter part, 10 Integral value calculation part, 11 Delay wave Detection unit, 12 integral value comparison unit, 13 band determination unit, 20 LPF, 21 coefficient table, 30 LPF, 31 LPF coefficient table, 32 BPF, 33 BPF coefficient table, 34 adder, 40 integral value calculation unit, 41 Extraction and addition unit, 42 integral value comparison unit, 43 band determination unit, 44 frequency interpolation filter unit.

Claims (12)

A Fourier transform unit for transforming an OFDM signal into a frequency spectrum signal;
A pilot extraction unit that extracts a pilot signal from the subcarrier component output from the Fourier transform unit;
A transmission pilot generator for generating a known signal corresponding to the pilot signal;
A division unit that calculates transmission path characteristics for subcarriers used for transmission of the pilot signal by dividing the pilot signal extracted by the pilot extraction unit by the transmission pilot signal;
A time interpolation filter unit that generates a transmission line characteristic by performing interpolation in the time direction on the transmission line characteristic of the pilot signal calculated by the division unit;
A frequency interpolation filter unit that performs interpolation in the frequency direction with respect to the transmission path characteristic output from the time interpolation filter unit;
A delay profile estimation unit that estimates a delay profile based on the transmission path characteristics of the pilot signal calculated by the division unit;
A passband setting unit that sets a passband of the frequency interpolation filter unit based on the delay profile estimated by the delay profile estimation unit;
The passband setting unit determines the passband so as to pass an incoming wave in a range equal to or less than a boundary time, and the boundary time is determined from the delay profiles estimated by the delay profile estimation unit. The ratio of the power integral value of the arriving wave over the range exceeding the boundary time to the power integral value of the arriving wave over the boundary time or less range is within a predetermined value or less and as large as possible. An OFDM demodulator characterized by defining. The passband setting unit
An integral value calculation unit that obtains a power integral value of an arriving wave over a range that is less than or equal to the boundary candidate time as a first integral value, and obtains a power integral value of the arriving wave over a range that exceeds the boundary candidate time as a second integral value; ,
An integral value comparison unit that determines whether a ratio of the second integral value to the first integral value is greater than the predetermined value;
Each time it is determined that the ratio is greater than the predetermined value, the boundary candidate time is increased, and the integral value calculation unit increases the first and second integral values at the boundary candidate time. When the ratio falls below the predetermined value, the boundary candidate time at that time is set as the boundary time, and the time until the boundary time is calculated. The OFDM demodulator according to claim 1, further comprising: a band determination unit that determines a band corresponding to a range as the pass band and outputs a signal representing the determined pass band. A Fourier transform unit for transforming an OFDM signal into a frequency spectrum signal;
A pilot extraction unit that extracts a pilot signal from the subcarrier component output from the Fourier transform unit;
A transmission pilot generator for generating a known signal corresponding to the pilot signal;
A division unit that calculates transmission path characteristics for subcarriers used for transmission of the pilot signal by dividing the pilot signal extracted by the pilot extraction unit by the transmission pilot signal;
A time interpolation filter unit that generates a transmission line characteristic by performing interpolation in the time direction on the transmission line characteristic of the pilot signal calculated by the division unit;
A frequency interpolation filter unit that performs interpolation in the frequency direction with respect to the transmission path characteristic output from the time interpolation filter unit;
A delay profile estimation unit that estimates a delay profile based on the transmission path characteristics of the pilot signal calculated by the division unit;
A passband setting unit that sets a passband of the frequency interpolation filter unit based on the delay profile estimated by the delay profile estimation unit;
The passband setting unit determines the passband so as to pass an incoming wave in a range equal to or less than a boundary time, and the boundary time is determined from the delay profiles estimated by the delay profile estimation unit. The ratio of the integrated power value of the incoming wave over the range exceeding the boundary time to the integrated value of the power of the incoming wave over the entire delay time range is within a predetermined value or less and is as large as possible. An OFDM demodulator characterized by the above. The passband setting unit
An integral value calculation unit that obtains the integral value of the power of the incoming wave over the entire delay time range as the first integral value, and obtains the power integral value of the incoming wave over the range exceeding the boundary candidate time as the second integral value;
An integral value comparison unit that determines whether a ratio of the second integral value to the first integral value is greater than the predetermined value;
Each time it is determined that the ratio is greater than the predetermined value, the boundary candidate time is increased, and the integral value calculation unit is configured to calculate the second integral value with the increased boundary candidate time, and the integration When the value comparison unit performs the determination, and the ratio is equal to or less than the predetermined value, the boundary candidate time at that time is set as the boundary time, and a band corresponding to the range up to the boundary time is defined. The OFDM demodulator according to claim 3, further comprising: a band determination unit that determines the passband and outputs a signal representing the determined passband. Divide the entire delay time range in which the incoming wave can appear into multiple sections using multiple division points,
The division point is set as the boundary candidate time,
3. When a delay wave having a predetermined threshold value or more is detected, a division point that is longer than the delay time of the delay wave and closest to the delay time is first selected as the boundary candidate time. Or the OFDM demodulator according to 4. The frequency interpolation filter unit has an LPF and a LPF coefficient table including a plurality of coefficients,
6. The method according to claim 1, further comprising: selecting a coefficient corresponding to the pass band from among a plurality of coefficients included in the LPF coefficient table in accordance with a signal representing the determined pass band. An OFDM demodulator according to claim 1. A Fourier transform unit for transforming an OFDM signal into a frequency spectrum signal;
A pilot extraction unit that extracts a pilot signal from the subcarrier component output from the Fourier transform unit;
A transmission pilot generator for generating a known signal corresponding to the pilot signal;
A division unit that calculates transmission path characteristics for subcarriers used for transmission of the pilot signal by dividing the pilot signal extracted by the pilot extraction unit by the transmission pilot signal;
A time interpolation filter unit that generates a transmission line characteristic by performing interpolation in the time direction on the transmission line characteristic of the pilot signal calculated by the division unit;
A frequency interpolation filter unit that performs interpolation in the frequency direction with respect to the transmission path characteristic output from the time interpolation filter unit;
A delay profile estimation unit that estimates a delay profile based on the transmission path characteristics of the pilot signal calculated by the division unit;
A passband setting unit that sets a passband of the frequency interpolation filter unit based on the delay profile estimated by the delay profile estimation unit;
The passband setting unit
The delay time range in which the delayed arrival wave can appear is divided into a plurality of sections, and the power integration value of the arrival wave is obtained for each of the plurality of sections, and the power integration value of the section in which the power integration value is relatively large among the plurality of sections. A band corresponding to the set is determined as a pass band of the frequency interpolation filter unit,
The plurality of sections are divided into a first set and a second set, and the division is divided into a maximum one of the integral values of the sections belonging to the first set and the sections of the sections belonging to the second set. The sum of the power integration values of all the sections belonging to the first set so as to be smaller than the minimum integral value of the power integration values of all the sections belonging to the second set. The ratio of the sum to the sum is within a predetermined value and is as large as possible,
An OFDM demodulator characterized in that a band corresponding to a set of sections belonging to the second set divided in this way is determined as a pass band. A Fourier transform unit for transforming an OFDM signal into a frequency spectrum signal;
A pilot extraction unit that extracts a pilot signal from the subcarrier component output from the Fourier transform unit;
A transmission pilot generator for generating a known signal corresponding to the pilot signal;
A division unit that calculates transmission path characteristics for subcarriers used for transmission of the pilot signal by dividing the pilot signal extracted by the pilot extraction unit by the transmission pilot signal;
A time interpolation filter unit that generates a transmission line characteristic by performing interpolation in the time direction on the transmission line characteristic of the pilot signal calculated by the division unit;
A frequency interpolation filter unit that performs interpolation in the frequency direction with respect to the transmission path characteristic output from the time interpolation filter unit;
A delay profile estimation unit that estimates a delay profile based on the transmission path characteristics of the pilot signal calculated by the division unit;
A passband setting unit that sets a passband of the frequency interpolation filter unit based on the delay profile estimated by the delay profile estimation unit;
The passband setting unit
The delay time range in which the delayed arrival wave can appear is divided into a plurality of sections, and the power integration value of the arrival wave is obtained for each of the plurality of sections, and the power integration value of the section in which the power integration value is relatively large among the plurality of sections. A band corresponding to the set is determined as a pass band of the frequency interpolation filter unit,
The plurality of sections are divided into a first set and a second set, and the division is divided into a maximum one of the integral values of the sections belonging to the first set and the sections of the sections belonging to the second set. The total sum of the power integral values of all the sections belonging to the first set is smaller than the minimum integral value, and all the sections within the delay time range in which the delayed arrival wave can appear. The ratio of the power integrated value to the total sum is within a predetermined value or less and is as large as possible.
An OFDM demodulator characterized in that a band corresponding to a set of sections belonging to the second set divided in this way is determined as a pass band. The passband setting unit
Of the plurality of sections, the one with the smallest integrated power value belongs to the first set, and the other sections belong to the second set.
Each time it is determined that the ratio is less than or equal to the predetermined value, the process of moving the smallest integrated power value of the sections of the second set to the first set, Repeat until it is determined that it is greater than the predetermined value,
When it is determined that the ratio is greater than the predetermined value, the section that has just moved from the second set to the first set is returned to the second set;
The OFDM demodulator according to claim 7 or 8, wherein a band corresponding to a set of sections belonging to the second set in the returned state is determined as a pass band. The passband setting unit
Of the plurality of sections, the one with the largest power integration value belongs to the second set, and the other sections belong to the first set.
Each time it is determined that the ratio is greater than the predetermined value, the process of moving the largest integrated power value of the sections of the first set to the second set, Repeat until it is determined that the value is below the specified value,
The band corresponding to the set of sections belonging to the second set is determined as a pass band when it is determined that the ratio is equal to or less than the predetermined value. OFDM demodulator. The frequency interpolation filter unit is
An LPF, a BPF, an LPF coefficient table including a plurality of coefficients, and a BPF coefficient table including a plurality of coefficients;
Of the plurality of coefficients included in the LPF coefficient table, the one corresponding to the passband and the plurality of coefficients included in the BPF coefficient table according to the signal representing the determined passband. 11. The OFDM demodulator according to claim 7, wherein a device corresponding to the pass band is selected.   The delay profile estimation unit averages the delay profiles obtained by the plurality of estimations, outputs the averaged delay profile, and the integral value calculation unit calculates the power based on the averaged delay profile. The OFDM demodulator according to claim 1, wherein an integral value is obtained.

Images