JP2011149871A - Frequency detector, composite band radar equipped with the frequency detector, and missile guiding device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a frequency detector, a composite band radar equipped with the frequency detector, and a missile guiding device. <P>SOLUTION: The frequency detector includes a base band conversion part which converts a received signal of frequency hopping composed of a plurality of inputted frequency bands into a plurality of samples of base band signals, a Fourier transform part which performs the Fourier transform of the base band signals obtained by conversion by the base band conversion part into a plurality of frequency bins, a norm conversion part which converts a value of each frequency bin being a conversion result obtained by the Fourier transform part into a norm, an addition part which adds the norm for each frequency bin, and a peak frequency detecting part which detects a peak frequency being the result of addition by the addition part. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a frequency detector, a synthetic band radar including the frequency detector, and a flying object guidance apparatus.

  A step frequency type synthetic band radar (hereinafter referred to as a synthetic band radar) is a modified version of a chirped pulse radar. As shown in FIG. 23, the chirped pulse radar sweeps the frequency in time from a frequency at the end of the band with respect to a certain frequency bandwidth B. Since it has a wide bandwidth, if the received signal is phase-adjusted and subjected to inverse Fourier transform, a waveform close to an impulse is obtained, and the distance to the target can be detected based on the peak of the impulse.

  The synthetic band radar is obtained by temporally discretizing the sweep frequency of the chirped pulse radar, and the frequency changes stepwise. The operations other than the discretization are almost the same, and a waveform close to an impulse is obtained by phase adjustment of the received signal and inverse Fourier transform, and the distance from the peak of the impulse to the target can be detected.

  However, in the synthetic band radar, the pulse continues to be sent for a period sufficiently longer than the pulse length of the chirp pulse. Therefore, when the target or the radar device is moving, the relative distance changes during the detection period, and it is clear. Peak does not appear and correct detection is not possible. Therefore, it is known to detect the target moving speed from the Doppler frequency of the received signal and perform distance detection after correcting the moving amount during the target detection period (for example, Patent Document 1, Patent Document 2, And Patent Document 3).

Japanese Patent No. 3709698 JP 2005-308723 A JP 2009-180666 A

  The conventional synthetic band radar is a method for obtaining an optimum moving speed in decision feedback (Patent Document 2), the interval between transmitted pulses and the frequency interval are limited to specific conditions (Patent Document 3), etc. There were problems such as problems in calculation time and convergence, and unnecessary conditions added to the specifications, so it was not always convenient.

  Accordingly, an object of the present invention is to provide a convenient frequency detector, a synthetic band radar including the frequency detector, and a flying object guidance apparatus.

  In order to achieve the above object, a frequency detector according to claim 1, wherein a baseband converter that converts a received signal of frequency hopping composed of a plurality of frequency bands into a baseband signal of a plurality of samples, and the baseband A Fourier transform unit that Fourier transforms the baseband signal transformed by the transform unit into a plurality of frequency bins; a norm transform unit that transforms the value of each frequency bin of the transform result transformed by the Fourier transform unit into a norm; An addition unit that adds the norm for each frequency bin, and a peak frequency detection unit that detects a peak frequency as a result of addition by the addition unit.

  The synthetic band radar according to claim 9 is converted by the baseband conversion unit that converts a frequency hopping reception signal composed of a plurality of input frequency bands into a baseband signal of a plurality of samples, and the baseband conversion unit. A Fourier transform unit that Fourier-transforms the baseband signal into a plurality of frequency bins, a norm transform unit that transforms the value of each frequency bin of the transform result transformed by the Fourier transform unit into a norm, and the norm as a frequency bin An addition unit that adds each time, a peak frequency detection unit that detects a peak frequency as a result of addition by the addition unit, and detects a relative speed with a target from the peak frequency detected by the peak frequency detection unit, A distance from the received signal corrected using the relative speed to a target is detected.

  The flying object guidance device according to claim 10 is a preprocessing unit that converts a signal input from a sensor into a form suitable for processing, a signal processing unit that performs signal processing on an output of the preprocessing unit, and A guidance signal generation unit that generates a guidance signal from an output of the signal processing unit and outputs the guidance signal to a steering device, and the signal processing unit samples a plurality of input signals of frequency hopping received from a plurality of frequency bands. A baseband conversion unit that converts the baseband signal into a baseband signal, a Fourier transform unit that Fourier transforms the baseband signal converted by the baseband conversion unit into a plurality of frequency bins, and a conversion result converted by the Fourier transform unit A norm conversion unit that converts the value of each frequency bin into a norm, an addition unit that adds the norm for each frequency bin, and a result obtained by addition by the addition unit A peak frequency detection unit for detecting a peak frequency, and a relative speed to the target is detected from the peak frequency detected by the peak frequency detection unit, and a distance from the received signal corrected using the relative speed to the target is detected. It is characterized by including a synthetic band radar characterized by.

  According to the present invention, it is possible to provide a convenient frequency detection device, a synthetic band radar including this frequency detector, and a flying object guidance device.

The figure which shows the structure of the frequency detector 1 which concerns on Embodiment 1 of this invention. The flowchart which shows the process of the frequency detector 1 which concerns on Embodiment 1 of this invention. The figure for demonstrating synthetic | combination zone | band radar. The figure for demonstrating operation | movement of the frequency detector 1 which concerns on Embodiment 1 of this invention. The figure for demonstrating operation | movement of the frequency detector 1 which concerns on Embodiment 1 of this invention. The figure for demonstrating short time FFT. The figure for demonstrating short time FFT. The figure for demonstrating operation | movement of the frequency detector 1 which concerns on Embodiment 1 of this invention. The figure which shows the structure of the frequency detector 1 which concerns on Embodiment 2 of this invention. The figure for demonstrating frequency selective control fading. The figure for demonstrating frequency selective control fading. The figure which shows the structure of the frequency detector 1 which concerns on Embodiment 3 of this invention. The figure which shows the structure of a peak matching type Fourier-transform part. The figure which shows the structure of a peak matching type Fourier-transform part. The figure for demonstrating the principle of a peak matching type Fourier-transform. The figure for demonstrating the principle of a peak matching type Fourier-transform. The figure for demonstrating the principle of a peak matching type Fourier-transform. The figure which shows the structure of the frequency detector 1 which concerns on Embodiment 4 of this invention. The figure for demonstrating the principle of a peak matching type Fourier-transform. The figure for demonstrating the principle of a peak matching type Fourier-transform. The figure which shows the structure of the synthetic | combination zone | band radar 20 which concerns on Embodiment 5 of this invention. The figure for demonstrating a flying body guidance device. The figure for demonstrating synthetic | combination zone | band radar.

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

(Embodiment 1)
Embodiment 1 of the present invention will be described. FIG. 1 is a diagram showing a configuration of a frequency detector 1 according to Embodiment 1 of the present invention.

  The frequency detector 1 includes a baseband conversion unit 2, a Fourier transform unit 3, a norm extraction unit 4, an addition unit 5, and a peak detection unit 6.

  The baseband conversion unit 2 converts a received signal that is sequentially input in time series for a plurality of frequency bands into a baseband signal in which one frequency band is composed of a plurality of samples.

  The Fourier transform unit 3 performs Fourier transform on the baseband signal converted by the baseband transform unit 2.

  The norm extraction unit 4 converts the result (Fourier transform result) converted by the Fourier transform unit 3 into an amplitude (or power) from which the phase component has been removed, and extracts the norm component.

  The adding unit 5 adds the amplitude (or power) extracted by the norm extracting unit 4 for each frequency bin.

  The peak detector 6 detects the peak frequency from the result (addition result) added by the adder 5.

  FIG. 2 is a flowchart showing processing for detecting a peak frequency by the frequency detector 1 according to Embodiment 1 of the present invention.

  The frequency detector 1 receives received signals that are sequentially received in time series for a plurality of frequency bands.

Here, the pulse structure of the received signal is composed of N _p frequency bands similar to the transmitted pulse structure as shown in FIG. 3, and the interval between the center frequencies of adjacent frequency bands is f _st , _{Suppose that there} are N _f frequency bands.

  The input received signal is converted into a baseband signal in which one frequency band is composed of a plurality of samples by the baseband conversion unit 2 (St1).

  The baseband signal converted in St1 is sent to the Fourier transform unit 3 and subjected to Fourier transform (St2).

  The result of the Fourier transform in St3 is sent to the norm extraction unit 4 and converted into an amplitude (or power) obtained by removing the phase component from the Fourier transform result (St3).

  FIG. 4 shows the amplitude component of the result of fast Fourier transform (FFT) of Np pulses in each frequency band for each frequency band. The horizontal axis is the frequency bin number, and the vertical axis is the linear scale amplitude. Since the SNR (Signal Noise Ratio) of one pulse is very low at 0 dB, the peak heights are varied, and the peak positions that should be arranged while being shifted little by little are also varied. Even if peak detection is performed for each frequency band in this state, it is difficult to estimate the correct peak position because the influence of noise is too strong.

  Therefore, an amplitude component is added for each frequency band. The reason why the addition is not performed in the complex number state where the spectrum is in phase is that the addition result in the complex number is smaller than the amplitude addition in terms of noise alone, but the addition is performed in the complex number when there is a phase shift for each frequency band. This is because the signals also cancel each other.

  The conversion result converted in St4 is sent to the adder 5, and the adder 5 adds the amplitude (or power), which is the Fourier transform result of all frequency bands, for each frequency bin (St5).

  The addition result at St5 is shown in FIG. It is clear that the spectrum using all the pulses improves the SNR, and the peak position is between frequency bin numbers 2 and 3.

  The addition result added in St5 is sent to the peak detector 6, and the peak detector 6 detects the peak frequency (St6).

  The peak detector 6 detects a peak from FIG. In FIG. 5, the peak is clearly between frequency bin numbers 2 and 3, and a simple peak search, that is, a method that compares the heights of each bin and outputs the highest bin number, finds it with high accuracy. I wo n’t.

Therefore, a waveform close to the waveform of FIG. 5 is prepared in advance and fitting is performed. Each spectrum in FIG. 4 is a Fourier transform result of N _p pulses in the same frequency band. FIG. 4 shows a case where Fourier transformation is performed without applying a window. That is, since it is a rectangular window, each spectrum has the shape of the absolute value of the sinc function, and the addition result thereof has a shape close to the absolute value of the sinc function. Therefore, a waveform of the sinc function waveform, particularly a peak position waveform, is prepared in advance, and the peak portion of FIG. 5 is combined with the amplitude as appropriate, and the center frequency of the sinc function is swung so that the peak frequency matches.

The best matching frequency, for example, the frequency of the sinc function that minimizes the square error is the peak frequency to be obtained. The peak frequency obtained here is a Doppler frequency with respect to the center frequency of the bandwidth composed of f _st · N _f .

  According to Embodiment 1 of the present invention, peak detection is performed by adding amplitudes, but each frequency stool component is more likely to perform peak detection using power expressed by the square of the amplitude instead of the amplitude. It can be said that the amount of calculation when calculating the norm of is small and efficient.

  Note that it is desirable that the waveform to be fitted is matched with the spectral shape of the window when Fourier transform is performed, if possible. When spectrum addition is performed with amplitude, it is possible to use the shape of the window spectral shape as it is, for example, a sinc function. However, when addition is performed with power, when generating a fitting waveform, the spectrum of the window is used. It is necessary to square the shape, which slightly increases the amount of calculation. However, the amount of calculation can be reduced by generating fitting waveforms at very fine intervals in advance and storing them in a memory.

  However, if fitting is performed in a very narrow range around the peak, a similar convex waveform may be used without necessarily matching the spectral shape of the window. In this case, at the time of amplitude adjustment before fitting, it is preferable to adjust not only the coefficient to be applied to the fitting waveform but also the offset that is added to the fitting waveform.

  In FIG. 5, the peak is between the frequency bin numbers 2 and 3, and in such a case, the initial amplitude adjustment is difficult, and the accuracy of the fitting result may be deteriorated. Therefore, when performing Fourier transform, a short time FFT (ST-FFT: Short Time-Fast Fourier Transform) is used for the purpose of complementing between the points of the spectrum.

ST-FFT is a method of performing FFT by adding zero to the original signal, as shown in FIG. Since the score of the spectrum of the FFT result is determined by the score of the FFT frame, the score of the spectrum can be increased so as to interpolate between the original points by adding zero and increasing the score. FIG. 7 shows the result of ST-FFT of the original signal of FIG. 7, a ST-FFT length as 4N _p, is the result of increasing the number of four times. Compared with the case of FIG. 4, it can be seen that the spectrum is interpolated between the points to obtain a smooth spectrum shape.

  FIG. 8 shows the result of adding the same in amplitude, and the peak position is clearer than the case shown in FIG.

  By doing so, the peak position becomes clear, and even if the peak position is between the frequency bins in the original number of points, no sensitivity deterioration occurs.

  In addition, the method of calculating | requiring a peak frequency from the spectrum of FIG. 8 may perform fitting with the waveform corresponding to the spectrum shape of a window similarly to the case where ST-FFT is not performed. Looking at the peak position in FIG. 8, it can be seen that it is between bin numbers 10 and 11, and the ST-FFT length is high when the number of samples is relatively small, such as twice or four times the original number of samples. It can be seen that fitting is required when it is desired to obtain a peak with accuracy.

However, if the ST-FFT length e.g. 16N _p and sufficiently long, the shape of the near peak is sufficiently smooth, it is sufficient to detect peaks in the peak search.

  Note that discrete Fourier transform (DFT: Discrete Fourier Transform) may be used instead of ST-FFT. ST-FFT is a method for obtaining a smooth spectrum as described above. According to DFT, the reference frequency at the time of Fourier transform can be a number other than an integer.

(Embodiment 2)
Next, a second embodiment of the present invention will be described. FIG. 9 is a diagram showing a configuration of the frequency detector 1 according to Embodiment 2 of the present invention. The same components as those in the first embodiment are denoted by the same reference numerals, and redundant description is omitted. It is different from the first embodiment in that it is configured by the DFT transform unit 27 instead of the Fourier transform unit 3 of the first embodiment.

  The predicted value of the peak position is input to the DFT converter 27, and the peak frequency is detected only in the vicinity of the predicted value.

  By setting the reference frequency interval at the time of DFT to an appropriate fine value, a spectrum equivalent to that in FIG. 7 can be obtained.

In the present embodiment, the reference frequency at the time of Fourier transform is f (or each frequency: 2πf) of the following Fourier transform equation.

Here, k is a sample number, Y (k) is a signal to be subjected to Fourier transform, and N is a Fourier transform frame length.

  However, since the DFT has a large amount of calculation compared to the ST-FFT, it is desirable to input a predicted value of the peak position in advance and calculate only the spectrum in the vicinity.

  For example, by calculating up to the waveform of FIG. 5 by FFT and searching for the peak position, it can be seen that the peak position is between frequency bin numbers 2 and 3, so what are the frequencies between bin numbers 2 and 3? DFT is performed by designating it as a reference frequency.

  The predicted value of the peak position may be extracted from the signal in this way. For example, when the target Doppler frequency and relative speed are estimated to some extent in advance, the predicted Doppler frequency and relative The DFT may be performed only in the vicinity of the Doppler frequency corresponding to the speed.

When the peak is detected from the DFT result, the spectrum obtained by adding the amplitude (or power) for all frequency bands is used as before. Thereafter, an appropriate waveform is fitted to the spectrum after addition to obtain a peak frequency, or a peak search may be performed if DFT reference frequencies are set at sufficiently fine intervals.

Thus, by performing peak frequency detection using a smooth spectrum, more accurate estimation is possible.

  In the above method, the detected peak frequency is a value corresponding to the center frequency of all bands including all frequency bands. However, the above method may increase an error when there is frequency selective fading.

  FIG. 10 shows an amplitude spectrum after Fourier transform of each frequency band when the target is a point target of only one point. Since the SNR at the time of reception is very large as 100 dB, unlike FIG. 7, the peak frequency for each frequency band is clear and the peak frequencies are aligned while being slightly shifted.

  On the other hand, FIG. 11 is an example in the case where there is frequency selective fading, and the height of the peak after Fourier transform varies greatly depending on the frequency band.

  In the example of FIG. 11, the frequency band that gives the peak with the lowest frequency is significantly larger than the other. Therefore, even if such a spectrum is added to the amplitude to obtain the peak frequency, It approaches the Doppler frequency and becomes a value far from the Doppler frequency with respect to the overall center frequency.

(Embodiment 3)
Next, Embodiment 3 of the present invention will be described. FIG. 12 is a diagram showing a configuration of the frequency detector 1 according to Embodiment 3 of the present invention. The same components as those in the first embodiment are denoted by the same reference numerals, and redundant description is omitted. The fourth embodiment differs from the first embodiment in that the Fourier transform unit 3 is a peak matching Fourier transform unit 28.

  The baseband converter 2 receives transmission frequency information, and the received signal is converted into a baseband signal composed of a plurality of samples by a local signal having the transmission frequency. The peak matching type Fourier transform described above in the Fourier transform unit 28 is applied to the baseband signal. The output spectrum of the Fourier transform unit 28 is converted into amplitude (or power) by the norm extraction unit 4.

  Subsequently, the spectrum of the amplitude (or power) of all frequency bands is added by the adding unit 4 for each identical frequency bin. The peak frequency is detected by the peak detector 6 from the result.

  13 and 14 are diagrams showing details of the peak matching Fourier transform unit 28. FIG.

  FIG. 13 shows an embodiment corresponding to a method using ST-FFT.

  The zero length determination unit 9 determines an appropriate zero length that is different for each frequency band using the input transmission frequency information, and the signal adjustment unit 10 has the length determined by the zero length determination unit 9. Zero is added to the input multiple sample sequence. The ST-FFT unit 11 performs ST-FFT on this and outputs the result.

  FIG. 14 shows an embodiment corresponding to a method using DFT.

  Based on the input peak prediction value and transmission frequency information, the reference frequency determination unit 12 determines a different DFT reference frequency for each frequency band. The DFT unit 13 performs DFT on the input sample sequence with the determined reference frequency and outputs it.

  In the method of determining the length of zero using the frequency information at the time of transmission or determining the reference frequency, it is not necessary to determine the reference frequency based on an equation every time. In most cases, the frequency at the time of transmission is determined by the specification, so values corresponding to each frequency band are calculated in advance and stored in the memory, and the frequency information at the time of reception of which frequency band the received signal should be processed The value corresponding to the frequency band may be read from the memory.

FIG. 15 is a diagram for explaining the principle of the peak matching Fourier transform. Each stage in FIG. 15A shows a frame configuration when a received signal in a frequency band of center frequencies f ₁ , f ₂ , and f ₃ is Fourier transformed, for example. If the center frequencies at the time of transmission are different frequency bands, and these are subjected to Doppler shift, the Doppler frequencies differ by the ratio of the original center frequency. In the figure, the wavelengths are shown to be shorter in the order of f ₁ , f ₂ , and f ₃ . In this state, zeros having different lengths n _i (i = 0, 1,..., N _f ) are added to the received pulse train of each frequency band, that is, N _p baseband samples. Go.

Specifically, zero is added to the length of N _p + n _i so that P _i that is the wavelength of each frequency band is always a constant value. This can be written as shown in FIG. 15B when the Fourier transform frame length, that is, for convenience, 1 second is regarded as the same 1 second in any frequency band. That is the same length in which the frequency band is N p ₊ n _i, and the wavelength of the received signal is the same at all frequencies bands. As a result, the Doppler frequency peaks of all frequency bands become the same frequency after Fourier transform. Incidentally, the ratio of the Fourier transform length of length N _p which signal is present is different for each frequency band, which is no problem.

  Such an operation can be interpreted as shown in FIG. FIG. 16A schematically shows the spectrum of the received signal in each frequency band from 0 as the frequency. Since the transmission frequency is different for each frequency band, the spectrum peak of the received signal having the Doppler frequency as an intermediate frequency is also shifted little by little. The ratio of the center frequencies of these Doppler frequencies is determined by the ratio of transmission frequencies. Therefore, when the horizontal axis is compressed at a different ratio for each frequency band so that the transmission frequency becomes the same frequency in a pseudo manner, the result is as shown in FIG. 16B, and the spectrum peaks of the received signals in all frequency bands match. To do.

Therefore, the length of N _p + n _i , that is, the compression ratio is determined not by the Doppler frequency itself but by the frequency difference between the center frequency and the frequency band at the time of transmission, and does not depend on the target moving speed.

  Note that the zero addition in FIG. 15 is for convenience, and in the actual calculation, the portion where the signal is zero is zero regardless of what is multiplied, so no calculation is performed, and it is indicated by Np where the signal exists. It is sufficient to calculate only the area.

Next, a method of determining the specific value of n _i using Equation. Assume that the received signal sequence is defined as follows.

The original signal has an amplitude and a fixed phase component in addition to this, but is not related to the operation explanation. Therefore, it is omitted for the sake of simplicity. k is a pulse in the same frequency band, that is, a sample number. F _i is a Doppler frequency or a value equivalent thereto, and here, the original Doppler frequency f _{d, i} is multiplied by T ₂ which is a pulse repetition interval (PRI) and N _p. .

When this is Fourier-transformed with respect to the frequency m in a system in which N _p + n _i is 1 second, the Fourier transform coefficient C _{i, m} can be expressed by the following equation.

When the Fourier transform to be applied is DFT, m is not necessarily an integer and may be a real number. Since the purpose is to align the peaks of the spectrum of each frequency step after the Fourier transform, the portion F _i · (N _p + n _i ) / N _p corresponding to the frequency in equation (3) has the same value regardless of the change of i. it may be determined the value of n _i such that.

When F _i is expressed using the Doppler frequency F ₀ corresponding to the 0th frequency band, it can be expressed by the following equation.

However, C _F is a value obtained by dividing the frequency step interval value f _st by the overall center frequency f ₀ . _{Here, (1 + iC F) ·} (N p + n i) / N p may be determined the value of _{n i} to be constant irrespective to i. First,

Put it. b may be any number as long as n _i is not a negative value such as negative. Calculated for ease of, here, placing b = 1 / a (N p _{C F),} and can solve n _i, it can be expressed by the following equation.

  When performing DFT, some suitable m are selected, and the DFT result with uniform peaks can be obtained by performing the Fourier transform of Equation (3) on the sample sequence of each frequency band. In the actual Fourier transform, the received sample sequence including amplitude and fixed phase components omitted in the process of deriving the equation is performed.

If the Doppler frequency is estimated in advance, the Doppler frequency corresponding to the zeroth frequency band is set to f _{d, 0} ,

It is sufficient to perform a Fourier transform on m in the vicinity of such a value.

However, at this time, depending on the relationship between the PRI value and the moving speed, the Doppler frequency spectrum may be folded.

In such a case, the appropriate m will exceed the value of N _p + n _i . In that case, in ST-FFT in which m can only be calculated from 0 to a value of N _p + n _i −1, the peaks cannot be aligned. Therefore, in such a case, it is preferable to calculate the surrounding area using DFT after registering m.

When performing this operation with ST-FFT is zero behind the different n _i pieces for each frequency band of the sample row addition to performing the operation. In that case it is necessary to select an appropriate value of b to be substantially an integer in the range n _i is the i. For example, the value b = 1 / (N _p C _F ) may deviate from an integer depending on the value of N _f , but a correction term b = 1 + 1 / (N _p C _F ) is added. As a result, a value close to an integer can be obtained.

  As a practical matter, if a Doppler frequency is aliased, it is difficult to calculate a Doppler frequency without ambiguity in the number of aliasing without prior knowledge. Therefore, it is desirable to know the Doppler frequency, that is, the range of the relative speed with respect to the target in some way in advance.

  An example of the DFT spectrum corresponding to each frequency band thus obtained is shown in FIG. Peak matching DFT was applied to the same signal as in FIG. It can be seen that the peak positions of all the curves match.

  Since the horizontal axis is m, the scale is different from FIG. In addition, since the calculation is performed by extracting only the vicinity of the expected Doppler frequency, the display position of the peak on the graph is different, but this is a problem only for the graph display.

It should be noted that, by so far spectrum display of how to take the T ₂ problem, all, the return of the Doppler frequency is generated. When DFT is used for peak matching and the horizontal axis is set to m, the same waveform is repeated at a constant period when m is taken from 0 to infinity instead of the occurrence of aliasing. Therefore, it is necessary to register in advance to some extent m.

(Embodiment 4)
Next, a fourth embodiment of the present invention will be described. As described above, when the peak matching is performed by DFT, it is better to have a certain estimated value of the moving speed or Doppler frequency in advance. Therefore, an embodiment in which the result of the Doppler frequency detection method without performing peak matching is used as a predicted value and detailed estimation of the vicinity of the predicted value is performed by the peak matching DFT will be described.

  FIG. 18 is a diagram showing a configuration of the frequency detector 1 according to the fourth embodiment of the present invention. In this embodiment, Doppler frequency detection without performing peak matching is performed, and this is used as an approximate estimated value, and peak matched DFT is performed around the estimated value.

  The received signal input to the frequency detector 1 is input to the baseband converter 2. The baseband converter 2 receives the frequency information at the time of transmission, and is converted into a baseband signal consisting of a plurality of samples in total by one sample per pulse by a local signal having a frequency at the time of transmission in each frequency band. The baseband signal is branched into two, and one is input to the Fourier transform unit 3. In the Fourier transform unit 3, Fourier transform is performed for each frequency band. The spectrum obtained by the Fourier transform is converted into an amplitude (or power) having no phase by the norm extraction unit 4. For the output, the adder 4 adds all the results of a plurality of frequency bands for each same frequency bin. The added spectrum is peak-detected by the peak detector 6 and an approximate value of the Doppler frequency is output.

  The other of the two branched outputs of the baseband conversion unit 2 is input to the peak matching Fourier transform unit 28 in order to perform peak matching DFT. At the same time, an approximate value of the Doppler frequency, which is an output of the peak detection unit 6, or a value obtained by converting this into a moving speed is input to the peak matching type Fourier transform unit 28.

  The peak matching Fourier transform unit 28 determines a range of m for performing DFT based on the input approximate value of the Doppler frequency, and applies the peak matching DFT to the input baseband signal. Details are as described above. The output of the peak matching Fourier transform unit 28 is input to the norm extraction unit 17 and converted into amplitude (or power). The result is input to the adder 18 and the results of all frequency bands are added for the same m. The addition result is input to the peak detection unit 19, m for giving a peak is calculated, and the Doppler frequency is calculated from m and output.

  As described above, when there is frequency selective fading, the error does not increase with the method that does not perform peak matching, but the difference in Doppler frequency for each frequency band is not so large. Works well enough.

  In the peak matching type Fourier transform unit 28 that performs the peak matching type DFT, the range in which the peak may exist is determined in consideration of the possibility of the shift due to the frequency selective fading, and the range in which the DFT is performed is determined. Good.

  By doing so, even when there is frequency selective fading and there is a Doppler frequency aliasing due to the problem of PRI setting, highly accurate Doppler frequency detection can be performed.

  In the case where there is aliasing, it is necessary to register in advance the number of times of aliasing when detecting the Doppler frequency. Then, the respective offsets are added to the respective Doppler frequency detection results and output.

  FIG. 19 shows an example in which peak matching DFT is performed with m registered in advance. This is an example in which DFT is performed only in the vicinity of an estimated value of Doppler frequency determined from a previously input estimated value of Doppler frequency or an estimated value of relative movement speed. The horizontal axis is the bin number when the m subjected to DFT is numbered with an integer in order from 0. Although the power variation for each frequency band is lower than that in FIG. 17, the total SNR is as low as 0 dB. Therefore, the amplitude varies due to noise more than frequency selective fading, and the peak frequency also varies. However, it can be seen that the peak positions are roughly aligned.

  FIG. 20 shows the result obtained by adding the amplitudes, and the peak frequency is clear. Once the value of m giving the peak is determined, the Doppler frequency corresponding to the zeroth frequency band can be determined from Equation (7).

  The peak position estimation method from FIG. 20 is the same as described above, and may be performed by fitting peak waveforms, or may be peak search if DFT is performed at sufficiently fine intervals.

When the peak matching DFT is performed under the condition that the spectrum is folded due to the T ₂ setting problem, the peak matching characteristic, which is a feature of the present invention, is reduced in a repetitive waveform outside the range including m that gives a correct peak. . That is, in a repetitive spectrum outside the correct m range, since the compression rate on the horizontal axis is not correct, the spectrum is shifted horizontally for each frequency band. In this embodiment, after matching the peaks, addition is performed by amplitude (or power) and peak frequency detection is performed. At that time, if the range of m is not correct, the peak of the result of addition becomes low. .

  Therefore, for example, the number of spectrum folds can be estimated by the following method. The repetition period of the waveform is a value obtained by converting the folding period of the Doppler frequency into a dimension of m. What is necessary is just to compare the peak value after the addition which appears for every period, and to find the peak where this becomes the maximum. However, when the SNR is low, there is an influence of noise. Therefore, it is only necessary to sample the peak value every period, apply the smoothing to the value, and select the point that seems to be the maximum.

(Embodiment 5)
FIG. 21 is a diagram showing a configuration of the synthetic band radar 20 according to the fifth embodiment of the present invention.

  In the synthetic band radar 20, a pulse train having a step frequency is output from the transmission unit 25 using the output of the frequency synthesizer 26. This is radiated as radio waves through an appropriate RF process (not shown). The radio wave reflected upon hitting the target is input as a received signal to the synthetic band radar 20 as a received signal through an appropriate RF process (not shown).

  The received signal is input to the frequency detector 1 and the Doppler frequency is detected by the method described so far. When the Doppler frequency is folded, for example, the output of the accelerometer 30 mounted on the equipment including the synthetic band radar 20 is used to determine the number of turns, and then the Doppler frequency is estimated. I do. Further, the baseband signal for each frequency band is output from the frequency detector 1 to the representative value extraction unit 31.

  The calculated Doppler frequency is input to the speed calculator 21 and converted into a speed dimension.

  In the representative value extraction unit 31, the peak value of the spectrum of each frequency step is obtained, or calculated from the Doppler frequency of each frequency band output from the frequency detector 1 or the moving speed obtained by the velocity calculation unit 21. Spectral components corresponding to the Doppler frequency at each frequency step are extracted from the baseband signal and used as representative values for each frequency step. This is input to the phase correction unit 22.

  The phase correction unit 22 uses the speed calculated by the speed calculation unit 21 to correct the phase shift associated with the change in the relative distance during the detection period with respect to the target for the phase of the representative value of each frequency step.

The corrected representative value is input to the inverse Fourier transform unit 23, and is subjected to the inverse Fourier transform in the order of the frequency of the frequency step. From the result of the inverse Fourier transform, an impulse waveform having a peak at a position corresponding to the relative distance to the target is obtained, and the distance from the peak position to the target is detected and output.

  In the synthetic band radar, the speed detection error has a great influence on the distance detection error. Therefore, the frequency detector of the present invention can be used to detect a highly accurate Doppler frequency, that is, a moving speed, in a reliable and simple method. As a result, it is possible to improve the accuracy of the synthetic band radar itself.

(Embodiment 6)
FIG. 22 is a diagram showing a configuration of the flying object guiding apparatus 32 according to the sixth embodiment of the present invention. It is the structure which produces | generates the induction | guidance | derivation signal for guide | inducing a flying body from the frequency detector of this invention and the synthetic | combination zone | band radar containing this.

The output of the sensor 33 including an antenna and an accelerometer is input to the flying object guidance device 32. In the flying object guidance device 32, preprocessing is appropriately performed by the preprocessing unit 34 so as to be suitable for processing in the signal processing unit 35 in the next stage, such as amplification, frequency conversion, and analog-digital conversion. The signal processing unit 35 includes the synthetic band radar 20 of the present invention and N processing units 38-1 to 38 -N that detect other information necessary for generating the guidance signal. The other information necessary for generating the induction signal is, for example, the wave number and angle of the incoming wave. Information generated by the signal processing unit 35 is input to the guidance signal generation unit 36, and these signals are integrated to generate a signal necessary for flying object guidance. The generated guidance signal is output to the steering device 37, and the flying object is steered.

As described above, by using the synthetic band radar of the present invention for generating the guidance signal of the flying object, it is possible to generate a guidance signal with higher sensitivity and accuracy.

In FIG. 22, the signal and information proceed without returning from the sensor to the steering device, but information is exchanged between the blocks as necessary. For example, the output of the synthetic band radar 35 is passed to the processing unit 38-1, and processing based on the range information, for example, angle detection is performed.

  The frequency detector of the present invention is a method for detecting the frequency or Doppler frequency of a signal whose frequency changes in a stepwise manner, and the applicable range is not limited to a synthetic band radar, but also performs frequency hopping. It can also be applied to other systems. For example, in a system that performs frequency hopping, the specified center frequency and frequency interval between frequency bands and the hopping sequence are known, but there is a problem with the accuracy of the oscillator, and the exact center frequency of the transmitter is It can also be applied when you want to detect it. In this case, the receiver may convert the received signal into a baseband signal using its own oscillator output based on a known specification, and perform the same processing. This is not the detection of the Doppler frequency but the detection of the center frequency shift of the oscillator.

  Note that the present invention is not limited to the above-described embodiment, and can be embodied by modifying the components without departing from the scope of the invention in the implementation stage. Various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the above embodiments. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, the constituent elements over different embodiments may be appropriately combined.

DESCRIPTION OF SYMBOLS 1 Frequency detector 2 Baseband transformation part 3 Fourier transformation part 4 Norm extraction part 5 Addition part 6 Peak detection part 9 Zero length determination part 10 Signal adjustment part 11 ST-FFT part 12 Reference frequency determination part 13 DFT part 27 DTF conversion part 28 Peak matched Fourier transform

Claims (10)

A baseband converter that converts a received signal of frequency hopping composed of a plurality of input frequency bands into a baseband signal of a plurality of samples;
A Fourier transform unit that Fourier transforms the baseband signal transformed by the baseband transform unit into a plurality of frequency bins;
A norm transform unit that transforms the value of each frequency bin of the transform result transformed by the Fourier transform unit into a norm;
An adder for adding the norm for each frequency bin;
A peak frequency detection unit for detecting a peak frequency of the result added by the addition unit;
A frequency detector comprising: The frequency detector according to claim 1, wherein the peak frequency detection unit detects a Doppler frequency corresponding to a center frequency of an entire band including the plurality of frequency bands.   The frequency detector according to claim 1, wherein the Fourier transform unit performs a short time FFT by adding zero to the baseband signal.   The Fourier transform unit sets a reference frequency interval at the time of Fourier transform in a vicinity of a frequency at which a peak is expected to be set more finely than an interval determined by the number of the plurality of samples. The frequency detector according to 1.   The frequency detector according to claim 1, wherein the Fourier transform unit performs short-time FFT by adding zeros having different lengths to the baseband signal. The frequency detector according to claim 1, wherein the Fourier transform unit performs discrete Fourier transform by changing a reference frequency for each frequency band. The frequency detector according to claim 6, wherein the reference frequency that is different for each frequency band has a ratio corresponding to a transmission frequency of the frequency band. A second Fourier transform unit that Fourier transforms a frequency in the vicinity of the peak frequency detected by the peak frequency detection unit into a plurality of frequency bins as a Fourier transform reference frequency;
A second norm transform unit for transforming each frequency bin of the Fourier transform result by the second Fourier transform unit into a norm;
A second addition unit that adds the norm transformed by the second norm transformation unit for each same frequency bin;
A second peak detection unit for detecting a peak frequency of the addition result added by the second addition unit;
Further comprising
The frequency detector according to claim 2, wherein the reference frequency used in the second Fourier transform unit is different for each frequency band at a ratio corresponding to a transmission frequency in the frequency band. A baseband converter that converts a received signal of frequency hopping composed of a plurality of input frequency bands into a baseband signal of a plurality of samples;
A Fourier transform unit that Fourier transforms the baseband signal transformed by the baseband transform unit into a plurality of frequency bins;
A norm transform unit that transforms the value of each frequency bin of the transform result transformed by the Fourier transform unit into a norm;
An adder for adding the norm for each frequency bin;
A peak frequency detection unit for detecting a peak frequency of the result added by the addition unit;
A synthetic band radar, wherein a relative speed to a target is detected from a peak frequency detected by the peak frequency detection unit, and a distance from the received signal corrected using the relative speed to the target is detected. A pre-processing unit that converts a signal input from the sensor into a form suitable for processing, a signal processing unit that performs signal processing on the output of the pre-processing unit, and a steering signal that generates a guidance signal from the output of the signal processing unit An induction signal generation unit that outputs to the device,
The signal processing unit
A baseband converter that converts a received signal of frequency hopping composed of a plurality of input frequency bands into a baseband signal of a plurality of samples;
A Fourier transform unit that Fourier transforms the baseband signal transformed by the baseband transform unit into a plurality of frequency bins;
A norm transform unit that transforms the value of each frequency bin of the transform result transformed by the Fourier transform unit into a norm;
An adder for adding the norm for each frequency bin;
A peak frequency detection unit for detecting a peak frequency of the result added by the addition unit;
Including a synthetic band radar that detects a relative speed to a target from a peak frequency detected by the peak frequency detection unit, and detects a distance from the received signal corrected using the relative speed to the target. A flying object guidance apparatus characterized by that.