JP2002181921A - Method and apparatus for generation of pulse data, method and apparatus for generation of shape data, as well as apparatus for generation of transmission pulse signal - Google Patents

Abstract

PROBLEM TO BE SOLVED: To generate a transmission signal whose side lobe is reduced sufficiently and which has an ideal frequency component. SOLUTION: Waveform data on the transmission signal are multiplied by window function data in a frequency region, waveform data on an ideal frequency region in which the frequency component as a main frequency component is acquired and which sets an unnecessary side lobe to a zero state is generated, and waveform data on a time region generated in such a way that the waveform data is inverse-Fourier-transformed is used as transmission data so as to be stored in a ROM 21. The transmission data is read out from the ROM 21 in a transmission period so as to be converted into an analog signal by a D/A converter 22, and the transmission signal whose side lobe is suppressed sufficiently can be obtained.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a transmission pulse signal generator for generating a transmission pulse signal to be transmitted to a wireless or wired transmission line.

[0002]

2. Description of the Related Art Conventionally, for example, in a radar apparatus, a transmission pulse signal is generated by one of the following three methods. (1) A method in which the output of a signal generation source is switched in a square step waveform, and turned on during the transmission period of the transmission pulse signal, and turned off during other periods. (2) A method in which the transmission level is changed during the transmission pulse signal emission period and other periods. (3) A method in which a transmission pulse signal is multiplied by a step-like waveform signal using a mixer to shape it.

The method (1) is a method of generating a transmission pulse signal by switching and outputting a signal, and is equivalent to a case where the output of a signal generation source is multiplied by a square step signal in a time domain. I can say. This method has a problem that a high level side lobe is generated over a wide range around a fundamental frequency component (main lobe) of a transmission signal due to a sudden change of a switch ON / OFF signal. I have.

In the method (2), during the signal output period,
This is a method of reducing side lobes by changing a transmission level during a signal transmission period instead of performing binary switching expressed by a square step waveform. In general, a method of multiplying waveform data of a transmission pulse by trapezoidal data by digital processing is adopted.
Compared with the switching by the method (1), the level of the side lobe far from the main lobe can be reduced, but the side lobe still exists at a higher level.

The method (3) is a method in which a transmission pulse is multiplied by a step-like waveform signal using a mixer. In this method, not only does the phase change drastically, but also overshoot and undershoot occur in the shape output. Will happen. For this reason, it also has extra time components and frequency components in the portions where overshoot and undershoot occur.

FIG. 49 is a waveform diagram in which a burst signal is cut out by a square step wave by the method (1) (FIG. 50 is FIG. 49).
FIG. 51 is a waveform diagram showing the burst component enlarged, and FIG. 51 is a waveform diagram showing a burst signal cut out by a trapezoidal step wave by the method (2) (FIG. 52 is a waveform diagram showing the burst component enlarged in FIG. 51). Show.

In the above example, the results of FFT processing of each time domain data and measurement of the frequency distribution are shown in FIG.
3 to FIG. 55. 53 is a time-domain waveform of a square step wave (solid line) and a trapezoidal step wave (dotted line), FIG. 54 is a frequency distribution corresponding to each step wave, and FIG.
FIG. 3 is a frequency characteristic diagram showing an enlarged portion near 0 [MHz]. As is clear from these figures, in any of the step waves, the frequency components are distributed over a wide range in addition to the main frequency, and the side lobe is not sufficiently suppressed.

[0008]

The above-described conventional transmission pulse signal generation devices in a radar device and the like have all assumed the concept of "creating a waveform from the time domain". As a result, none of the methods can suppress the side lobe and do not extract the frequency component of the main lobe well. This is true not only for the transmission pulse of the radar apparatus but also for all transmission pulses of data or the like in a wireless or wired transmission path.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems, and has a transmission pulse signal generating apparatus capable of generating a transmission pulse signal having an ideal frequency component with sufficiently reduced side lobes. It is an object of the present invention to provide a pulse data generation method / apparatus and a shape data generation method / apparatus used in this apparatus.

[0010]

In order to achieve the above object, a pulse data generation method, a shape data generation method, a pulse data generation device, a shape data generation device, and a transmission pulse signal generation device according to the present invention are as follows. Such a characteristic configuration is provided.

(1) In the pulse data generation method according to the present invention, a main frequency component is cut out and unnecessary side lobes are reduced to a zero state by multiplying the waveform data of the transmission pulse signal by window function data in the frequency domain. This is characterized in that waveform data in an ideal frequency domain to be generated is generated, and this is subjected to Fourier inverse transform to generate pulse data as waveform data in a time domain.

(2) More specifically, a first procedure for creating basic waveform data of a transmission pulse signal, and replacing the basic waveform data created in the first procedure with waveform data in the frequency domain by Fourier transform. The second procedure and this second
A third procedure of multiplying the Fourier-transformed waveform data obtained in the procedure by window function data having a curve meeting a predetermined condition, and applying a Fourier inverse to the waveform data obtained in the third procedure. And converting the waveform data into time domain waveform data.

(3) In (1) or (2), the window function is one of a rectangular window, a Hamming window, a Hanning window, a Gaussian window, and a Blackman-Harris window.

(4) In the shape data generation method according to the present invention, square step wave data corresponding to the transmission width of a transmission pulse signal is converted into frequency domain waveform data by Fourier transform, and window function data having the top as a vertex is converted. It is characterized by generating shape data in the ideal frequency domain by multiplication and performing inverse Fourier transform on this to generate shape data in the time domain.

(5) More specifically, a first procedure for creating square step wave data corresponding to the transmission width of the transmission pulse signal, and a Fourier transform of the waveform data created in the first procedure, and a frequency domain And a second step of multiplying the waveform data obtained in the second procedure by a window function having a curve having a top as a vertex, which satisfies a predetermined condition. And a fourth procedure of performing inverse Fourier transform on the frequency domain waveform data obtained in the third procedure and replacing it with time domain waveform data to generate shape data. And

(6) In (4) or (5), the window function is one of a square window, a Hamming window, a Hanning window, a Gaussian window, and a Blackman-Harris window.

(7) The pulse data generating apparatus according to the present invention cuts out the main frequency component and reduces unnecessary side lobes to a zero state by multiplying the waveform data of the transmission pulse signal by the window function data in the frequency domain. This is characterized in that waveform data in an ideal frequency domain to be generated is generated, and this is subjected to Fourier inverse transform to generate pulse data as waveform data in a time domain.

(8) More specifically, a basic waveform data generating section for generating basic waveform data of a transmission pulse signal, and a frequency domain waveform data obtained by performing a Fourier transform on the basic waveform data generated by the basic waveform data generating section. , A window function multiplier for multiplying the waveform data obtained by the Fourier transformer by window function data having a curve that meets a predetermined condition, and a window function multiplier obtained by the window function multiplier. An inverse Fourier transform unit that performs inverse Fourier transform of the waveform data and replaces the waveform data with time-domain waveform data.

(9) In (7) or (8), the window function is one of a square window, a Hamming window, a Hanning window, a Gaussian window, and a Blackman-Harris window.

(10) The shape data generation device according to the present invention converts square step wave data equivalent to the transmission width of a transmission pulse signal into waveform data in the frequency domain by Fourier transform, and converts window function data having the top as a vertex. The waveform data is multiplied to generate waveform data in an ideal frequency domain, which is inversely Fourier transformed to waveform data in a time domain to generate shape data.

(11) Specifically, a square step wave data creating section for creating square step wave data corresponding to the transmission width of the transmission pulse signal, and the Fourier transform of the waveform data created by the square step wave data creating section Then, a frequency component is multiplied by a Fourier transform unit that replaces the waveform data in the frequency domain with a window function having a curve having a top apex that matches a predetermined condition with the waveform data obtained by the Fourier transform unit. A window function multiplying unit to be cut out, and a Fourier transform unit that generates shape data by performing Fourier inverse transform on the frequency domain waveform data obtained by the window function multiplying unit and replacing it with time domain waveform data. Features.

(12) In (10) or (11), the window function is a square window, a Hamming window, a Hanning window,
It is one of a Gaussian window and a Blackman Harris window.

(13) The transmission pulse signal generation device according to the present invention cuts out the main frequency component and sets unnecessary side lobes to the zero state by multiplying the waveform data of the transmission pulse signal by the window function data in the frequency domain. A pulse data storage unit that generates waveform data in the ideal frequency domain and stores the waveform data in the time domain generated by inverse Fourier transform as pulse data, and a pulse transmission period from the pulse data storage unit. A digital-to-analog converter that reads out the pulse data and converts it into an analog signal to generate a transmission pulse signal.

(14) Alternatively, square step wave data equivalent to the transmission width of the transmission pulse signal is converted into waveform data in the frequency domain by Fourier transform, and multiplied by window function data having the top as the top, to form the shape data in the ideal frequency domain. And a shape data storage unit that stores time-domain shape data generated by performing an inverse Fourier transform of the shape data.The shape data stored in the shape data storage unit is converted into pulse data generated in the time domain. A multiplication unit that reads the pulse data in accordance with the pulse width and multiplies the pulse data by using the read shape data as a coefficient; and generates a transmission pulse signal by converting the pulse data obtained by the multiplication unit into an analog signal. A digital-to-analog converter.

As described above, the generation of pulse data used in the conventional transmission pulse signal device has been based on the premise that "a waveform is formed from the time domain". This was considered on the premise of "creating a waveform from the frequency domain", and the "method / apparatus for extracting frequency domain data with a window function" of the present invention. With this method / apparatus, a transmission pulse signal having an ideal frequency component could be created by generating transmission pulse waveform data from the frequency domain.

More specifically, the main frequency component is cut out in the frequency domain, and data in the ideal frequency domain is created so that the far side lobes are in the zero state. This is a method / apparatus for generating a pulse. The completed data in the time domain is data having frequency components as drawn in the data in the ideal frequency domain. In other words, only the main lobe is emphasized, and data is created so that the side lobe can be made zero. In addition, since there was no sharp change in phase as in the case of using a mixer, a clear waveform without overshoot and undershoot could be created.

Thus, the method / apparatus of the present invention can be said to be a method / apparatus capable of producing the most ideal data.

The reduction / elimination of side lobes is very beneficial. Since frequency components exist only in the main frequency band in the frequency domain data, not only can the frequency components of the transmitted pulse be reliably extracted, but they are not affected by pulses other than the transmitted pulse or surrounding radio waves.
It also has the feature of not affecting. Based on these characteristics, the method of the present invention can produce an ideal pulse, and can be said to be a simple, reliable, and efficient method.

[0029]

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

First, the background of the idea of the present invention will be described.

The simulation of “when a signal is transmitted only for a short time in a sufficiently long time (pulsed)” is performed by comparing the transmitted signal with the frequency component of the signal. Is analyzed to verify what frequency components are included in the transmitted signal, and it can be said that the simulation of the radar technology itself has been performed.

The present invention has been obtained by analyzing radar pulses based on this simulation and conducting research for creating a new ideal radar pulse.

FIG. 1 shows the configuration of a transmission device in a conventional radar device. In FIG. 1, an IF signal generated by an IF signal oscillator 11 is sent to a modulator 12 and converted into a pulse signal by a modulation signal. This pulse signal is IF /
After being converted into an RF signal by the RF frequency conversion unit 13,
After being amplified by the RF signal amplifying unit 14 and unnecessary frequency components are removed by a BPF (band-pass filter) 15, the power is amplified by a power amplifier 16 and transmitted from an unillustrated antenna to space. It should be noted that the dashed line device in FIG. 1 may not be present, as in the case of a wire or the like.

Here, when the pulse signal is generated by the modulator 12, as described above, the side lobe cannot be sufficiently suppressed, and the waveform distortion of the transmission pulse makes it difficult to analyze the frequency of the reception pulse.

On the other hand, the present invention is characterized in that a transmission pulse having an ideal frequency characteristic is created by digital processing. Specifically, as shown in FIG. 2, the ROM (or FIR filter) 21 previously stores waveform data matching the transmission pulse, reads out the waveform data from the ROM 21 based on the modulation signal,
The signal is converted into an analog signal by the A converter 22 to generate an IF band transmission pulse. After the aliasing component of the sampling frequency is removed by the LPF (Low Pass Filter) or the BPF 23, the transmission pulse is transmitted to the IF as in FIG.
/ RF frequency converter 24 converts the RF signal
The signal is amplified by the F signal amplifying unit 25, unnecessary frequency components are removed by the BPF 26, the power is amplified by the power amplifier 27, and the amplified signal is transmitted from the antenna (not shown) to the space.

According to this configuration, since the waveform data is set in advance so as to have an ideal frequency characteristic, it is possible to sufficiently suppress the side lobe, and almost no waveform distortion of the transmission pulse occurs. Thus, the frequency of the received pulse can be analyzed with high accuracy.

By the way, the transmission pulse waveform data is
Conventionally, storing in a waveform memory such as M and reading and outputting it in synchronization with the transmission timing has been performed as one method of digital processing of signal generation. However, in the conventional technique for improving the frequency characteristics as described above, as shown in FIG. 3, pulse data is generated by multiplying time-domain data by a trapezoidal or rectangular coefficient. At this time, after converting the pulse data into data in the frequency domain by Fourier transform, the frequency components (sidelobes and the like) are analyzed in the frequency domain, and the coefficients are adjusted based on the analysis result, whereby the pulse data of the pulse data is analyzed. The frequency characteristics were improved.

However, in the pulse data created by multiplying the data in the time domain by the trapezoidal or rectangular coefficients, as described above, a high-level side lobe is generated over a wide range from the main lobe.

Therefore, in the present invention, as shown in FIG.
Pulse data is created by creating data in the frequency domain excluding the frequency band of the main lobe, multiplying it by a specific coefficient and pulsing it, and then converting it to data in the time domain by inverse Fourier transform. . In this case, since frequency components other than the frequency band of the main lobe are removed in the frequency domain, almost no side lobe remains in the data converted to the time domain by the inverse Fourier transform.

On the other hand, as a technique for reducing side lobes, a method of multiplying data in the time domain by a function called a “window function” is often used. In this method, the pulse data in the time domain is multiplied by the window function according to the input pulse width to gradually change the rising and falling edges of the pulse, and the pulse data is Fourier-transformed and replaced with the data in the frequency domain. This is a method of analyzing a signal component near the center of a pulse.

The method of multiplying the signal in the time domain by the window function as described above is often used in measuring instruments and the like. FIG. 5 shows an example of the configuration. In FIG. 5, a signal to be measured is multiplied by window function data, cut out, converted from time-domain data to frequency-domain data by Fourier transform, and subjected to frequency analysis.

FIGS. 6 and 7 show the flow of processing in this case. In FIG. 6, the signal to be measured (infinitely continuous signal)
In the case of (a), when a certain section is specified and analyzed, a continuous signal is generated by repeating the waveform signal (b) obtained by cutting the section (c), and the analysis is performed as if it continues infinitely. However, when the cut waveform (b) is repeated, it becomes discontinuous at the joining portion, and an error occurs. Therefore, as shown in FIG. 7, the cut waveform signal (b) is multiplied by the window function (d) matching the section to generate a signal (f) having the same level at the beginning and end, and this signal is generated. By repeating (f), an apparently continuous signal (g) is generated and analyzed.

It is considered that a transmission pulse signal is generated using the above method. In this case, as shown in FIG.
The pulse data is cut out by multiplying the data in the time domain with the window function data. At this time, the cut-out pulse data is replaced by frequency domain data by Fourier transform to check the occurrence of side lobes, and the window function is corrected if necessary.

However, according to this method, the rise and fall times of the signal are too long, and only the signal components near the center are utilized, so that it cannot be said that the transmission pulse is efficient.

The method of the present invention uses this window function for a pulse signal in the frequency domain. In other words, “extraction of frequency domain data by window function” according to the present invention
Is a method of reducing the side lobes using the curve of the window function, emphasizing the main lobe, and performing Fourier transform to generate a pulse signal from which the basic frequency component is efficiently extracted. .

FIG. 9 shows an example of the configuration. In FIG.
The data in the frequency domain is cut out by multiplying it by window function data corresponding to the frequency of the main lobe, and is returned to the data in the time domain by inverse Fourier transform. In this case, frequency domain data is cut out using a window function, so that frequency domain data including ideal frequency components is created, and when this is converted back to time domain data by Fourier inverse transform, frequencies other than the main lobe frequency are suppressed. The obtained pulse data is obtained.

Hereinafter, the present invention will be described with reference to specific examples.

There are several types of window functions, such as a rectangular window, a Hamming window, a Hanning window, and a Gaussian window. Any window function is effective for sidelobe reduction. here,
A result of a simulation performed using a window function called a “Blackman Harris window” will be described. Note that the Blackman Harris window w (n) is defined as follows.

W (n) = 0.35875−0.48829 ^* cos (2πn / N)
+0.14128 ^* cos (4πn / N) -0.01168 ^* cos (6πn / N) where w (n) is the value of the window function n: 0,1,2, ..., (N-2), (N-1 ) N: Total number of window functions The procedure of the method of “cutting out a frequency domain pulse signal by a window function” is as follows.

(Procedure for Creating Sidelobe Reduction Signal) Basic waveform data (hereinafter, pulse data) of a pulse signal is created.

By Fourier transform, the pulse data is replaced with frequency domain data.

The pulse data after the Fourier transform is subjected to a complex multiplication by applying a window function curve that meets the conditions. At this time, the same window function is also multiplied on the side turned at half the sampling frequency.

The pulse data generated in step (1) is returned to pulse data in the time domain by inverse Fourier transform.

The pulse data obtained in is converted into an analog signal, which is used as a transmission pulse.

Next, a method of generating two patterns of (1) pulse data having only a specific frequency component and (2) shape data used as a shape coefficient based on the above method will be described.

(1) Method of Generating Pulse Data Having Specific Frequency Components This method is useful, for example, when designating a specific frequency band and creating pulse data to be transmitted. By storing the pulse data generated by this method in a ROM or the like and reading out and outputting the data sequentially at transmission timing, a pulse having a desired frequency component can be emitted. FIG. 10 shows a configuration example of the pulse data generation device.

In FIG. 10, reference numeral 31 denotes a signal source for generating burst data corresponding to the pulse width of the main frequency. The burst data output from this signal source 31 is converted by the Fourier transformer 32 from data in the time domain to data in the frequency domain. The data is converted into data, and a window function generator 34
Are multiplied by the window function data in the frequency domain generated by the above and cut out by the window function.
5 returns to the time domain data, whereby pulse data with reduced side lobes is obtained. This pulse data retains the original pulse data, and its rising and falling change in a smooth curve.

The window function data generated by the window function generator 34 has a main lobe band (target band: center frequency ft) with respect to the sampling frequency fs.
And a Blackman-Harris window in which the peak is located in a band (center frequency fs-ft) obtained by folding this target band at fs / 2. Let N be the number of frequency band points in this window.

The following pulse is actually considered.

[Example 1] A signal of 30 [MHz] is set to 1
Pulse data sampled at 50 [MHz] for 2 [μs] In [Example 1], the state of the pulse data in the time domain (burst wave) and the state of the pulse data subjected to Fourier transform and replaced with the frequency domain are shown. FIG. 14 respectively
(FIG. 15 is an enlarged view of the burst portion in FIG. 14), FIG.
(FIG. 17 is an enlarged view around 30 [MHz] of FIG. 16). In addition, 2 [μ] at a sampling frequency of 150 [MHz]
s], the number of data becomes 300 points.

The signal of [Example 1] is in the same state as the simulation in which the signal was transmitted by switching for a sufficiently long time. Referring to FIG.
It can be seen that the main lobe is a frequency component of [MHz] and a portion of 120 [MHz] obtained by folding it at half the sampling frequency. Also, it can be seen that countless side lobes are generated throughout. 30
When the vicinity of the [MHz] component is enlarged, it can be seen that the level of the side lobe near the main lobe (near side lobe) is also large.

As described above, in the pulse data in which the signal is suddenly changed from the state where there is no output, the level of the side lobe is high, and the side lobes are innumerably spread over the entire frequency components.

In order to greatly reduce the side lobes and generate pulse data that emphasizes the main lobe, a signal with reduced side lobes is generated by the method of the present invention under the following conditions.

(Conditions) The frequency component from the fundamental frequency component is 23 at ± 1 [MHz] away from the fundamental frequency (30 [MHz]).
Dropping by [dB] or more.

± 5 from the fundamental frequency (30 [MHz])
A frequency component from the fundamental frequency component is reduced by 60 dB or more at a portion separated by [MHz].

When converted to the time domain,
Pulse data that has a moderate fall and a flat part.

In order to create pulse data in the time domain satisfying the above conditions, data is created in the following procedure.

(Procedure for signal side lobe reduction: FIG. 11)
See) 30 [MH] at sampling frequency 150 [MHz]
z] is generated by sampling 2 [μs], and Fourier transform is performed.

The data (including the complex number) subjected to the Fourier transform at ± 1 [MH] from the fundamental frequency 30 [MHz]
The frequency components at 29 [MHz] and 31 [MHz] of z] are reduced by about 18 [dB]. From here, another 5 [d
B] Adjust the value of the total number N of window functions so as to obtain falling window function data and perform complex multiplication (here, N = 4
0). Similarly, 25 from 30 [MHz] to ± 5 [MHz]
The frequency components of [MHz] and 35 [MHz] are 60 [d].
B] Also confirm that it has dropped.

7 which is の of the sampling frequency
Frequency component folded from 5 [MHz] (120 [MH]
z]) is also multiplied by a similar window function.

The portion not covered by the window function is set to have zero frequency component (in a state where zero is complexly multiplied).

The pulse data in the frequency domain (data in the ideal frequency domain) obtained by the above is inversely Fourier-transformed and converted into pulse data in the time domain.

In the time-domain pulse data generated in step (1), the small signals before and after the burst signal are replaced with zero, and a Fourier transform is performed to check whether the conditions are satisfied.

N = 40 used to satisfy this condition
FIG. 18 (FIG. 19 is an enlarged view around 30 [MHz] in FIG. 18).
(Procedure). The pulse data of the ideal frequency region cut out by multiplying by the window function is shown in FIG.
(Enlarged view around [MHz]) (procedure). By returning the pulse data in the ideal frequency domain extracted by this window function to the time domain by Fourier inverse transform, pulse data in the time domain having frequency components extracted in the frequency domain is completed. The pulse data after Fourier inverse transform is shown in FIG. 22 (FIG. 23 is an enlarged view of the burst portion in FIG. 22) (procedure).

From FIGS. 22 and 23, it can be seen that the pulse data after the Fourier inverse transform is in such a state that the waveform rises and falls with gentle steps. In addition, it can be seen that it has a flat portion and efficiently utilizes data. However, it can be seen that the waveform is slightly different from the trapezoidal waveform. However, compared to the input data, the data is slightly longer in time (about 3 [μs]).

For confirmation, the newly created pulse data in the time domain is returned to pulse data in the frequency domain by Fourier transform. Considering the principle of Fourier transform / inverse Fourier transform, pulse data in an ideal frequency region extracted by a window function should return. However, in the pulse data in the time domain created by the procedure, since there is a very small output before and after the bursting signal, the Fourier transform is performed after setting the small signal to zero in the procedure,
Checking conditions. Fig. 2 shows the result after Fourier transform.
4 (FIG. 25 is an enlarged view around 30 [MHz] of FIG. 24).

As is clear from FIGS. 24 and 25,
Neighboring side lobes generated near the fundamental frequency of 30 [MHz] are greatly reduced. In particular, the far side lobe drops by 110 dB or more from the main lobe, and is reduced to a level that can be regarded as almost zero.

Therefore, according to the above method, it is possible to generate pulse data in which the main lobe in the frequency domain is emphasized and the side lobe is greatly reduced. Also,
Since pulse data having the same frequency component as the pulse data in the ideal frequency domain can be generated, it is possible to obtain pulse data in the time domain having pulse data in the frequency domain without wasteful constraints, and to generate pulses in the ideal frequency domain. A frequency component equivalent to the data is obtained. Further, the pulse data in the time domain after the conversion does not significantly impair the shape of the original pulse data, and can be said to be an excellent method in terms of efficiency.

The pulse data generated as described above is stored in a ROM or the like, and read out at the transmission timing to produce pulse data aiming at a very narrow band even in the entire frequency domain. Can be sent. In addition, there is also an advantage that interference with surrounding radio waves does not occur because there is no side lobe.

The following is a summary of the method of “cutting out pulse data in the frequency domain using the window function”.

By adjusting the curve of the window function (the number of window functions), it is possible to easily generate a signal containing a frequency component that meets the conditions, and it is efficient.

Pulse data having frequency components equivalent to pulse data in an ideal frequency region can be generated.

The pulse data in the time domain that does not significantly impair the shape of the original waveform can be generated.

Pulse data that can be significantly reduced to a level at which distant side lobes can be regarded as substantially zero can be generated.

It is possible to generate pulse data that does not interfere with other radio waves.

(2) Method for Generating Shape Data Used as Shape Coefficients Using the same method as (1), it is also possible to develop coefficients in the time domain and create “shape data” using an FIR filter. is there. Here, a case where data of a shape coefficient corresponding to a transmission width is generated is considered.

The data created in the above [Embodiment 1] shows that an infinite sine wave of 30 [MHz] is converted to 2 [μs]
It can be seen that the state is the same as the state cut out by the square step wave. This means, "In some long enough time,
This is a simulation of "transmitting a signal for a short time", which is the same as the result of analyzing the frequency component at that time. From the analysis results, it was found that in the past, countless frequency components (side lobes) different from the transmitted frequency components were generated. It was also found that the method of the present invention can significantly reduce the side lobe by extracting pulse data in the time domain with a step wave that gradually changes at the rise and fall.

However, in the method described above, the pulse data has a specific frequency component and is emitted.
Now, consider how to create data that can reduce side lobes regardless of the frequency components of the original data. At this time, it is considered that by creating the time data of the step wave that gradually changes at the rising and falling times, a shape coefficient that can reduce the side lobe of any signal is created.

The waveform data of the step wave of the shape coefficient (hereinafter, shape data) can be created by the same procedure as the “method of cutting out by a window function in the frequency domain”. FIG. 12 shows a configuration example of this shape data generation device.

In FIG. 12, reference numeral 41 denotes a signal source for generating data (square step data) corresponding to a pulse width. Data output from this signal source 41 is a Fourier transformer 42.
The data is converted to frequency domain data by a multiplier 43, multiplied by the frequency domain window function data generated by the window function generator 44, cut out by the window function, and timed by the inverse Fourier transformer 45. Returned to area data. Further, the width is adjusted so that the original pulse width is secured, and is written in the shape coefficient memory 47 as shape coefficient data for reducing side lobes. In operation, pulse data in the time domain from the transmission pulse signal source 48 is supplied to the multiplier 49 and the shape coefficient memory 4
7 is multiplied by the shape coefficient data read from. As a result, pulse data with reduced side lobes can be obtained.

The procedure for creating the shape coefficient will be described below.

(Step Wave Processing) Waveform data of a square step wave corresponding to the transmission width (here, 300 points) is created.

The Fourier transform is performed on the waveform data of to obtain waveform data in the frequency domain.

The waveform data is complex-multiplied by window function window data in which the top W (N / 2) of the Blackman-Harris window and the last W ((N / 2) −1) are arranged, and the frequency component is cut.

The inverse Fourier transform of the frequency domain waveform data is performed to convert it into time domain waveform data.

For the waveform data in the time domain obtained in step, the small signal before and after the step is cut and the width is adjusted. Thereby, shape coefficient data is obtained.

An example of creating step data (shape data) for reducing side lobes in the above procedure will be described.

[Example 2] Data for reducing side lobes with a pulse width of 300 points In this [Example 2], a sine wave is multiplied by a square step wave and shape data in the time domain and compared and verified. . It is assumed that the shape data has moderate rising and falling and has a flat portion.

FIGS. 26 and 27 (FIG. 28 shows the results of performing a Fourier transform with a square step wave of 300 points.
7 (enlarged view of the first frequency component). From the figure, it can be seen that the top component of the Fourier transform result, that is, the DC component, of the square step wave is the main lobe, and the side lobes are spread over the whole. By suppressing this spread and generating data with reduced side lobes, it should be possible to create shaped data.

A procedure for actually generating shape data for 300 points will be described below. Here, the total number (value of N) of the window functions of the Blackman-Harris window to be taken out is created in two cases of 50 and 100.

(Procedure for creating 300 point shape data: see FIG. 13) Waveform data of a square step wave of 300 points is created and Fourier-transformed.

The result of is cut out by a window function. At this time, the top of the frequency domain waveform data after the Fourier transform is set to “1” which is the maximum value of the window function of the Blackman Harris window. To place the vertex of the window function at the top, n
= N / 2 to N-1. Also, at the end of the frequency domain waveform data, data of -1 from the vertex of the window function (n = 0 to 0)
N / 2-1).

The complex data is multiplied by the waveform data obtained by the above and the waveform data after the Fourier transform of the step wave.

In the portion where the window function is not applied, the frequency component is set to zero (a state where zero is complexly multiplied).

The waveform data in the frequency domain (waveform data in the ideal frequency domain) obtained by the above is inversely Fourier-transformed and converted into waveform data in the time domain.

The minute signal before and after the step in the time domain data generated in step is replaced with zero, the pulse width is adjusted, and the result is confirmed by performing Fourier transform.

The state of the window function data in the frequency domain used for clipping with the window function data (N = 50), the waveform data of the ideal frequency domain clipped by multiplying by the window function, and the shape of the time domain extracted by inverse Fourier transform Figure 2 shows the data
9, FIG. 30 (FIG. 31 is an enlarged view of the leading frequency component in FIG. 30), and FIG. Window function data (N = 10
FIG. 33 and FIG. 34 show the state of the window function data in the frequency domain used for extraction in (0), the waveform data in the ideal frequency domain extracted by multiplying by the window function, and the time domain shape data extracted by inverse Fourier transform. 35 is an enlarged view of the leading frequency component in FIG. 34) and FIG.

It should be noted here that there is a relationship between the total number of window functions (value of N) to be multiplied in the procedure and the shape data generated in the procedure. If the total number of window functions for cutting out data from the frequency domain is reduced (here, N = 50), the rise / fall time of the completed shape data will be long. Conversely, when the total number of window functions is increased (N = 100), a step wave having a short rise / fall time of the completed shape data is obtained.

Also, the shape data created by the procedure becomes more data than the original 300 points. Therefore, in order to adjust the number of points, the completed data was forcibly set to 300 points.

The completed shape data and a square step wave having a sampling frequency of 150 [MHz]
FIGS. 37 to 48 show a state when 0 [MHz] sin data is cut out and a state when a frequency component is analyzed by performing Fourier transform.

Here, FIG. 37 is a waveform diagram showing a burst signal cut out in a square step, FIG. 38 is a waveform diagram showing an enlarged burst portion of FIG. 37, and FIG. 39 is a diagram showing a result of Fourier transform of the burst signal of FIG. FIG. 40 shows frequency characteristics.
39 is an enlarged frequency characteristic diagram showing around 30 [MHz] in FIG. 39, FIG. 41 is a waveform diagram of a signal obtained by multiplying the shape data by 30 [MHz] sin data at a sampling frequency of 150 [MHz], and FIG. 41 is an enlarged waveform diagram showing the burst portion, FIG. 43 is a frequency characteristic diagram showing the result of Fourier transform in FIG. 41, and FIG. 44 is 30 [MH] in FIG.
FIG. 45 is a waveform diagram of a signal obtained by multiplying the shape data by 30 [MHz] sine data at a sampling frequency of 150 [MHz], and FIG. 46 is a diagram illustrating a burst portion in FIG. 47 is an enlarged waveform diagram, FIG. 47 is a frequency characteristic diagram showing the result of Fourier transform in FIG. 46, and FIG. 48 is an enlarged frequency characteristic diagram around 30 [MHz] in FIG.

Comparing FIG. 32 and FIG. 36, when the total number of window functions (value of N) is reduced, the rise / fall time of the shape data becomes longer. Conversely, when the total number of cut-out window functions is increased, the rise / fall time of the shape data becomes shorter.

Also, comparing the frequency component (FIG. 39) cut out by the square step wave with the frequency component (FIG. 43, FIG. 47) cut out by the shape data, the main lobe cut out by the shape data (30 [M
Hz] component) is slightly lowered. In addition, it can be seen that the main lobe has also spread slightly. However, the side lobes are significantly lower than those cut out by the square step wave, and it is understood that the distant side lobes are almost completely removed. Again, the total number of cut-out window functions has an effect, and the total number of window functions N = 50.
In FIG. 7, the side lobe is significantly reduced as compared with the case where N = 100.

When the frequency components are cut out from the frequency domain by using a window function to generate shape data, the rise / fall time when the time domain data is converted by the inverse Fourier transform and the tolerance of the included frequency components are allowed. It is necessary to consider both trade-offs (positions where the side lobe is zero). Taking this into consideration, by extracting frequency components from the frequency domain using a window function, it is possible to create shape data that eliminates side lobes. As is apparent from the result of the Fourier transform of the shape data, the shape data capable of reducing the side lobes other than the frequency components included in the signal to zero is completed.

From the above, it is understood that the window function is useful not only for waveform data in the time domain but also for waveform data in the frequency domain. The method of extracting data using a window function in the frequency domain not only can create data with excellent properties without sidelobes, but also finds coefficients for digital filters and coefficients for reducing sidelobes. Application is possible in the field.

[0116]

As described above, according to the present invention, a transmission pulse signal generating apparatus capable of generating a transmission pulse signal having an ideal frequency component with a sufficiently reduced side lobe, and an apparatus for use in this apparatus. Pulse data generation method /
An apparatus and a method / apparatus for generating shape data can be provided.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a transmission device in a conventional radar device.

FIG. 2 is a block diagram showing an embodiment of a transmission pulse signal generation device according to the present invention.

FIG. 3 is a step diagram showing a conventional pulse data generation method for improving frequency characteristics.

FIG. 4 is a step diagram showing an outline of a pulse data generation method according to the present invention.

FIG. 5 is a step diagram for explaining a method of using a window function used in a conventional measuring instrument or the like.

FIG. 6 is a step diagram for explaining details of FIG. 5;

FIG. 7 is a step diagram for explaining details of FIG. 5;

FIG. 8 is a step diagram for explaining a pulse data generation method using the method of FIG. 5;

FIG. 9 is a step diagram showing a pulse data generation method according to the present invention.

FIG. 10 is a block diagram showing a configuration of a pulse data generation device according to the present invention.

FIG. 11 is a step diagram showing a procedure for reducing a signal side lobe in pulse data generation according to the present invention.

FIG. 12 is a block diagram showing a configuration of a shape data generation device according to the present invention.

FIG. 13 shows a diagram of a 30 in the shape data generation of the present invention
FIG. 4 is a step diagram showing a procedure for creating 0-point shape data.

FIG. 14 is a waveform chart showing a state (burst wave) of pulse data in a time domain in [Example 1].

FIG. 15 is an enlarged view of a burst part in FIG. 14;

FIG. 16 is a frequency characteristic diagram showing a state of pulse data that has been subjected to Fourier transform and replaced in the frequency domain in [Example 1].

FIG. 17 is an enlarged view around 30 [MHz] of FIG. 16;

FIG. 18 is a frequency characteristic diagram showing frequency domain window function data of a Blackman Harris window of N = 40 used to satisfy a predetermined condition in [Example 1].

FIG. 19 is an enlarged view around 30 [MHz] of FIG. 18;

FIG. 20 is a frequency characteristic diagram showing pulse data in an ideal frequency region cut out by multiplying by a window function in [Example 1].

FIG. 21 is an enlarged view around 30 [MHz] of FIG. 20;

FIG. 22 is a waveform chart showing pulse data after Fourier inverse transform in [Example 1].

FIG. 23 is an enlarged view of a burst part of FIG. 22;

FIG. 24 is a frequency characteristic diagram showing a result of Fourier transform of completed pulse data in [Example 1].

FIG. 25 is an enlarged view around 30 [MHz] of FIG. 24;

FIG. 26 is a waveform chart showing a 300-point square step wave in [Example 2].

FIG. 27 is a frequency characteristic diagram showing a result of Fourier transform of a 300-point square step wave in [Example 2].

FIG. 28 is an enlarged view of a leading frequency part in FIG. 27;

FIG. 29 is a table showing the window function data (N
= 50) is a frequency characteristic diagram showing the state of the window function data in the frequency domain used for the cutout.

30 is a frequency characteristic diagram showing waveform data in an ideal frequency region cut out by multiplying by the window function of FIG. 29 in [Example 2].

FIG. 31 is an enlarged view of a leading frequency component in FIG. 30;

FIG. 32 is a waveform diagram showing time-domain shape data obtained by performing an inverse Fourier transform on the output of FIG. 30 in [Example 2].

FIG. 33 is a graph showing the window function data (N
= 100), a frequency characteristic diagram showing the state of window function data in the frequency domain used for clipping.

FIG. 34 is a frequency characteristic diagram showing waveform data in an ideal frequency region cut out by multiplying by the window function of FIG. 33 in [Example 2].

FIG. 35 is an enlarged view of a leading frequency component in FIG. 34;

FIG. 36 is a waveform diagram showing time-domain shape data extracted by performing Fourier inverse transform on the output of FIG. 34 in [Example 2].

FIG. 37 is a waveform chart showing burst signals cut out in square steps in [Example 2].

FIG. 38 is an enlarged waveform diagram showing a burst portion in FIG. 37;

FIG. 39 is a frequency characteristic diagram showing a result of Fourier transform of the burst signal of FIG. 37;

FIG. 40 is an enlarged frequency characteristic diagram showing the vicinity of 30 [MHz] in FIG. 39;

FIG. 41 is a waveform diagram of a signal obtained by multiplying shape data by 30 [MHz] sine data at a sampling frequency of 150 [MHz] in [Example 2].

FIG. 42 is an enlarged waveform diagram showing a burst portion of FIG. 41;

FIG. 43 is a frequency characteristic diagram showing a result of the Fourier transform in FIG. 41.

FIG. 44 is an enlarged frequency characteristic diagram showing around 30 [MHz] in FIG. 43;

FIG. 45 is a waveform diagram of a signal obtained by multiplying shape data by 30 [MHz] sine data at a sampling frequency of 150 [MHz] in [Example 2].

FIG. 46 is an enlarged waveform diagram showing a burst portion in FIG. 45;

FIG. 47 is a frequency characteristic diagram showing a result of the Fourier transform in FIG. 46.

FIG. 48 is an enlarged frequency characteristic diagram showing the vicinity of 30 [MHz] in FIG. 47;

FIG. 49 is a waveform diagram in which a burst signal is cut out by a square step wave according to a conventional method.

FIG. 50 is an enlarged waveform diagram showing the burst component of FIG. 49;

FIG. 51 is a waveform diagram of a burst signal cut out by a trapezoidal step wave according to a conventional method.

FIG. 52 is an enlarged waveform chart showing the burst component of FIG. 51;

FIG. 53 is a waveform chart showing a time-domain waveform of a square step wave (solid line) and a trapezoidal step wave (dotted line).

FIG. 54 is a frequency characteristic diagram showing a frequency distribution corresponding to each step wave of FIG. 53;

FIG. 55 is an enlarged frequency characteristic diagram showing the vicinity of 30 [MHz] in FIG. 54;

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 11 ... IF signal oscillator 12 ... Modulator 13 ... IF / RF frequency conversion part 14 ... RF signal amplification part 15 ... BPF 16 ... Power amplifier 21 ... ROM (or FIR filter) 22 ... D / A converter 23 ... LPF or BPF 24 IF / RF frequency converter 25 RF signal amplifier 26 BPF 27 power amplifier 31 burst data signal source 32 Fourier converter 33 multiplier 34 window function generator 35 Fourier inverse transformer 41 square Step data signal source 42 ... Fourier transformer 43 ... Multiplier 44 ... Window function generator 45 ... Inverse Fourier transformer 47 ... Shape coefficient memory 48 ... Transmission pulse signal source 49 ... Multiplier

Claims (14)

[Claims] 1. Multiplying waveform data of a transmission pulse signal in a frequency domain by window function data to cut out a main frequency component and generate waveform data in an ideal frequency domain in which unnecessary side lobes are in a zero state, A pulse data generation method, which performs a Fourier inverse transform to generate pulse data as time-domain waveform data to generate pulse data. 2. A first procedure for creating basic waveform data of a transmission pulse signal, a second procedure for replacing the basic waveform data created in the first procedure with frequency domain waveform data by Fourier transform, A third procedure of multiplying the Fourier-transformed waveform data obtained in the second procedure by window function data having a curve meeting a predetermined condition; and a waveform data obtained in the third procedure. A fourth procedure for performing inverse Fourier transform on the time domain data and converting the waveform data into waveform data in the time domain. 3. The pulse data generation method according to claim 1, wherein the window function is one of a square window, a Hamming window, a Hanning window, a Gaussian window, and a Blackman-Harris window. 4. Fourier transform transforms square step wave data corresponding to the transmission width of a transmission pulse signal into frequency domain waveform data, and multiplies the window function data with the top as a vertex to generate ideal frequency domain shape data. A shape data in a time domain by performing an inverse Fourier transform on the shape data. 5. A first procedure for creating square step wave data corresponding to the transmission width of a transmission pulse signal, and a fourth procedure for performing Fourier transform on the waveform data created in the first procedure and replacing the waveform data with frequency-domain waveform data. A second procedure, and a third procedure of multiplying the waveform data obtained in the second procedure by a window function having a curve having a top as a vertex and meeting a predetermined condition to cut out a frequency component; And a fourth step of performing inverse Fourier transform of the frequency domain waveform data obtained in the third procedure and replacing the frequency domain waveform data with time domain waveform data to generate shape data. . 6. The shape data generating method according to claim 4, wherein said window function is one of a square window, a Hamming window, a Hanning window, a Gaussian window, and a Blackman-Harris window. 7. An ideal frequency domain waveform data in which a main frequency component is cut out and unnecessary side lobes are set to a zero state by multiplying the waveform data of the transmission pulse signal by the window function data in the frequency domain, A pulse data generation device, which performs a Fourier inverse transform to generate pulse data in time domain waveform data to generate pulse data. 8. A basic waveform data generating section for generating basic waveform data of a transmission pulse signal, and a Fourier transform section for replacing the basic waveform data created by the basic waveform data creating section with frequency domain waveform data by Fourier transform. A window function multiplier for multiplying the waveform data obtained by the Fourier transformer by window function data having a curve that matches a predetermined condition; and a Fourier inverse transform of the waveform data obtained by the window function multiplier. A pulse data generating apparatus, comprising: an inverse Fourier transform unit for converting the data into time domain waveform data. 9. The pulse data generation device according to claim 7, wherein the window function is one of a square window, a Hamming window, a Hanning window, a Gaussian window, and a Blackman-Harris window. 10. The square step wave data corresponding to the transmission width of a transmission pulse signal is converted into frequency domain waveform data by Fourier transform, and multiplied by a window function data having the top as a vertex, to generate ideal frequency domain waveform data. A shape data generating apparatus for generating shape data by converting the data into time-domain waveform data by performing an inverse Fourier transform. 11. A square step wave data creating section for creating square step wave data corresponding to the transmission width of a transmission pulse signal, and a Fourier transform of the waveform data created by the square step wave data creating section to obtain a frequency domain waveform. A Fourier transform unit that replaces the data, a window function multiplying unit that cuts out frequency components by multiplying the waveform data obtained by the Fourier transform unit by a window function having a curve with the top as a vertex that meets a predetermined condition. A Fourier transform unit that performs inverse Fourier transform of the frequency domain waveform data obtained by the window function multiplying unit and generates shape data by replacing the frequency domain waveform data with time domain waveform data. apparatus. 12. The window function may be a square window, a Hamming window,
The shape data generation device according to claim 10, wherein the shape data generation device is one of a Hanning window, a Gaussian window, and a Blackman-Harris window. 13. Multiplying waveform data of a transmission pulse signal by window function data in a frequency domain to generate a waveform data in an ideal frequency domain in which a main frequency component is cut out and unnecessary side lobes are set to a zero state, A pulse data storage unit that stores, as pulse data, time-domain waveform data generated by performing a Fourier inverse transform on the pulse data. A pulse data is read from the pulse data storage unit during a pulse transmission period, and is converted into an analog signal. A transmission pulse signal generation device, comprising: a digital-to-analog converter that generates a transmission pulse signal. 14. A square step wave data corresponding to a transmission width of a transmission pulse signal is converted into frequency domain waveform data by Fourier transform, and multiplied by window function data having a top as an apex to generate ideal frequency domain shape data. A shape data storage unit for storing time-domain shape data generated by performing an inverse Fourier transform of the shape data, and converting the shape data stored in the shape data storage unit to a pulse width of pulse data generated in the time domain. A multiplying unit for reading together and multiplying the pulse data by using the read shape data as a coefficient; and a digital / analog generating a transmission pulse signal by converting the pulse data obtained by the multiplying unit into an analog signal. A transmission pulse signal generation device comprising: a conversion unit.