JP3156815B2 - FM-CW radar signal processing device - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for sampling a beat signal obtained by mixing a transmission signal and a reception signal, converting the signal into corresponding digital signal data, and subjecting the signal data to digital signal processing such as fast Fourier transform. The present invention relates to a signal processing device of an FM-CW radar that analyzes the frequency of a beat signal by performing the above-described processing and measures the distance from the frequency of the beat signal to an object, and in particular, effectively removes sudden noise such as spike noise. TECHNICAL FIELD The present invention relates to a signal processing device for an FM-CW radar which performs digital signal processing in a form to prevent a decrease in distance measurement accuracy due to sudden noise.

[0002]

2. Description of the Related Art There is known a technique in which a noise component is removed by using a low-pass filter, a high-pass filter, a band-pass filter, etc., focusing on a difference between a desired signal and a frequency component of noise. However, when the frequency component of the signal and the frequency component of the noise substantially overlap, the filter cannot effectively remove the noise. Therefore, the applicant of the present invention disclosed in Japanese Patent Application No. 4-90403 that the spectrum data of the stationary noise during the non-transmission period is stored in the memory, and the spectrum data of the beat signal is corrected based on the spectrum data of the stationary noise. A technique for reducing the influence of stationary noise is proposed.

[0003]

The above-described correction of the spectrum data can reduce the influence of the stationary noise, but cannot eliminate sudden noise such as spike noise. When sudden noise is mixed, the distance measurement accuracy may be reduced, which is not preferable.

The present invention has been made to solve such a problem, and provides an FM-CW radar signal processing apparatus capable of reducing the influence of sudden noise and performing distance measurement more reliably. The purpose is to:

[0005]

According to a first aspect of the present invention, there is provided a signal processing apparatus for an FM-CW radar, wherein data of a digital beat signal is temporarily stored in a memory means, and the amplitude of the digital beat signal is changed to another level. Noise suppression processing means is provided for changing data so that the amplitude of a section larger than that of the section becomes zero or a small amplitude close to zero.

According to a second aspect of the present invention, there is provided a signal processing apparatus for an FM-CW radar, wherein the noise suppression processing means includes an amplitude density calculating means for obtaining an amplitude density of a digital beat signal within a period to be analyzed, and Noise determination threshold value setting means for setting a noise determination threshold value, and changing data so that the amplitude of the section having an amplitude exceeding the noise determination threshold value is zero or a small amplitude close to zero. It is characterized by comprising noise amplitude suppression means.

In the noise suppression processing means, an attenuation section is set on both sides of the section in which the amplitude is suppressed, and a signal for correcting data of the digital signal in the attenuation section so that the non-suppression signal and the suppression signal are smoothly connected. A configuration in which a waveform correction unit is provided may be used.

[0008]

According to a first aspect of the present invention, there is provided a signal processing apparatus for an FM-CW radar, wherein a noise for changing data so that the amplitude of the section becomes zero or a small amplitude close to zero for a section whose amplitude is larger than another level. The provision of the suppression processing means makes it possible to remove or suppress sudden noise. Therefore, the digital beat signal from which sudden noise has been removed or suppressed is subjected to digital arithmetic processing such as fast Fourier transform, and the distance is obtained from the frequency analysis result. Becomes possible.

The noise suppression processing means of the signal processing device for an FM-CW radar according to claim 2 first obtains the amplitude density of the digital beat signal within the analysis target period by the amplitude density calculation means, and then obtains a noise determination threshold value. The setting means determines that the portion having a high amplitude density is a beat signal, and sets an amplitude value higher than the beat signal by a predetermined value as a noise determination threshold value.
The noise amplitude suppressing means changes the amplitude of the section in the section exceeding the noise determination threshold to data of a small amplitude close to zero or near zero. Thus, the noise determination level can be varied according to the amplitude of the digital beat signal, and sudden noise can be effectively removed or attenuated regardless of the level of the beat signal. Since a signal-to-noise ratio equal to or higher than a predetermined value can be ensured, more reliable distance measurement can be performed.

A signal waveform correcting means is provided to correct the signal waveforms on both sides of the suppression section so that the non-suppression signal and the suppression signal are smoothly connected, so that the amplitude of the sudden noise at the beginning and at the tail can be reduced. Even if there are small parts,
It can be suppressed. Even if the amplitude of the beat signal fluctuates, the noise before and after the noise suppression section is attenuated even if the noise judgment threshold value is set to be rather high so as not to suppress the peak portion of the beat signal. Therefore, the noise component can be effectively suppressed.

[0011]

Embodiments of the present invention will be described below with reference to the accompanying drawings. FIG. 1 is a functional block diagram of a signal processing device according to the present invention, and FIG. 2 is an FM-C equipped with the signal processing device.
FIG. 2 is an overall block configuration diagram of a W radar device.

An FM-CW radar device 1 shown in FIG. 2 generates an FM signal 2a which sweeps a fundamental frequency, generates an FM signal 2a, and transmits the FM signal 2a to a transmission signal TS and a local signal L.
S, a power divider 3 for distributing the received signal R to the antenna 4 and a transmission signal TS to the antenna 4 and a reception signal R received by the antenna 4
Circulator 5 for efficiently extracting S, and receiving signal R
It comprises a mixer 6 for mixing S and the local signal LS to generate a beat signal BS, a signal processing device 7, and a display device 8 for displaying information on the distance to the object. The FM signal 2a is swept based on a sweep command signal 7a supplied from the signal processing device 7.

As shown in FIG. 1, a signal processing device 7 includes an A / D converter 11 that samples and quantizes a beat signal BS and converts it into a corresponding digital beat signal DS, and converts the data of the digital beat signal DS in time series. And a memory means 12 such as a RAM for storing the digital beat signal BS. The digital beat signal BS is subjected to digital arithmetic processing such as fast Fourier transform (FFT) and frequency analysis is performed. Digital operation means 13 for obtaining the distance to the object, and noise suppression processing means 14 according to the present invention
And

The noise suppression processing means 14 includes an overall operation control means 21, an amplitude density calculation means 22, a noise determination threshold value setting means 23, a noise amplitude suppression means 24, and a signal waveform correction means 25. Each means 21 to 25 is configured to operate based on a control program created in advance using a microcomputer system.

The overall operation control means 21 is provided with a sweep command signal 7
a, the sweep of the FM signal generation circuit 2 and the stop of the sweep are controlled, and during the sweep, the A / D conversion command 21a
To put the A / D converter 11 into an operating state. Also,
The address information 21b and the write command 21c are output to write the data of the digital beat signal DS sequentially output from the A / D converter 11 to the memory means 12, and the amplitude density calculation command 21e is output to output the amplitude density calculation command 21e. The means 22 is set in the operating state, and the calculation of the amplitude density of the data to be written to the memory means 12 is performed.

The amplitude density calculating means 22 includes an A / D converter 1
The amplitude of the beat signal BS is determined based on the sampling data DS sequentially output from 1 and the amplitude density indicating the relationship between the amplitude and the number of occurrences is determined. The amplitude LMAX having the maximum amplitude density and the spread of the density ( The data of (variance) δ is supplied to the noise determination threshold value setting means 23.

FIG. 3 is an explanatory diagram showing an example of a method of obtaining the amplitude. The reference level K is set to a level that is the center of the amplitude of the original beat signal BS. The amplitude density calculating means 22
The data DS output from the A / D converter 11 is compared with the reference level K, and half the time from when the data DS crosses the reference level K (including when the data DS coincides with the reference level K) to when the data DS crosses next is half. The maximum or minimum data UM1, UL1, UM2, UL2,... Are determined for each cycle section, and the noise information suppressing means is provided by associating the address information 21b at the start of each half cycle with the maximum or minimum data in the half cycle period. 24 is configured to be written into a memory 24M provided therein. Further, the amplitude density calculating means 2
2 is configured to count the number of occurrences for each preset amplitude section.

For example, when the data DS (t1) of the digital beat signal crosses the reference level L from bottom to top at time t1, the amplitude density calculating means 22 outputs the address information 21b at which the data DS at that time is written to the memory means 12. Is supplied to a data input terminal (not shown) of the memory 24M in the noise amplitude suppression means 24 via the data signal line (data bus) 22a, and a write command 22b is output to output the address information 21b of the memory 24M, for example. Write to address 0. Noise amplitude suppression means 24
Is provided with an address counter means (not shown).
Each time the writing is completed, the write address of the memory 24M is incremented.

The amplitude density calculating means 22 determines that the data DS has entered the upper half cycle period based on the fact that the data DS crosses the reference level K from below to above, and performs processing for detecting the maximum value of the data. Do. For example, while temporarily storing the data when the reference level K is crossed, the next data output from the A / D converter 11 is compared with the temporarily stored data, and the data having the larger value is temporarily stored. . This is repeated until the data DS crosses the reference level K next time, and the data DS temporarily stored at the time when the reference level K crosses (time t2) is the maximum data U of the half cycle period.
M1 is supplied to the memory 24M via the data signal line 22a, and a write command 22b is output to write the maximum data UM1 to, for example, address 1 of the memory 24M.

Further, the amplitude density calculating means 22 calculates difference data between the maximum data UM1 and the reference level K, sets the obtained difference data as the maximum amplitude data KU1 within the half cycle period, and sets the data KU1 in advance. It is determined which of the amplitude sections belongs to, and the number of occurrences of the belonging amplitude section is set to one. After writing the maximum data UM1 to the memory 24M, the address information 21b of the memory means 12 at time t2 is transferred to the data signal line 22.
a to the memory 24M, and generates a write command 22b to write the address 21b of the memory means 12, which is the start point of the next half cycle (the current half cycle end point), to, for example, address 3 of the memory 24M. .

Based on the fact that the data DS crosses the reference level K from top to bottom at time t2, the amplitude density calculating means 2
2 judges that the lower half cycle period has been entered, and A /
The minimum data among the data DS sequentially output from the D converter 11 is temporarily stored, and at the time when the reference level K is crossed (time t3), the minimum level data LM1 is transferred to, for example, address 4 of the memory 24M. Write. Also,
The amplitude density calculating means 22 determines which of the amplitude sections the difference data (lower maximum amplitude LM1) between the minimum level data LM1 and the reference level K belongs to, and increases the number of occurrences of the amplitude section to which it belongs by +1 ( Increment).

By repeating such an operation for all the amplitudes within the analysis target period, the address information 21b of the memory means 12 corresponding to the start point of each half cycle is stored at, for example, an odd address of the address of the memory 24M. The address contains the maximum data U for the half cycle period.
M1, UM2,... Or minimum data UL1, UL2,
... are stored. In addition, a predetermined distribution of occurrence for each amplitude section can be obtained.

The amplitude density calculating means 22 divides the upper half cycle and the lower half cycle into groups to obtain the amplitude densities, respectively.
The noise determination threshold above the reference level K and the noise determination threshold below the reference level K may be individually set.

A series of sweeps is completed and an amplitude density calculation command 21 is issued.
When the output of e is stopped, the amplitude density calculating means 22 obtains the data of the amplitude LMAX at which the amplitude density (the number of occurrences) becomes maximum and the spread (variance) δ of the density within the designated analysis target period, It is supplied to the noise determination threshold value setting means 23. The noise determination threshold value setting means 23 sets a noise determination threshold value (amplitude data determined as noise) LTH, and supplies the set amplitude data LTH to the noise amplitude suppression means 24. When the amplitude density is calculated for each of the upper and lower half cycles, the noise determination thresholds LTH (+) and LTH (-) are set for each.
The setting of the noise determination threshold LTH may be a constant multiple (for example, 1.2 times) of the amplitude LMAX at which the amplitude density is maximum, or may be spread (variance) with respect to the amplitude LMAX at which the amplitude density is maximum. ) A constant multiple of δ (constant may be 1)
May be added.

When the noise determination threshold value LTH is given, the noise amplitude suppression means 24 calculates the allowable upper limit by adding the noise determination threshold value LTH to the reference level K, Determination threshold LT
H is subtracted to calculate an allowable lower limit. Note that the noise determination thresholds LTH (+), LTH
When (-) is given, an allowable upper limit and an allowable lower limit are set correspondingly. The maximum or minimum data UM1, UL1, UM2, UL2,... For each half cycle period stored in, for example, the address of the even address in the memory 24M in the noise amplitude suppression means 24, and the allowable upper limit value and the allowable lower limit value If there is a waveform having an amplitude exceeding the noise determination threshold LTH,
The data change request 24a is supplied to the overall operation control means 21, and the start point address 24b of the corresponding amplitude section is provided.
(The address before the address of the data exceeding the allowable upper limit value or the allowable lower limit value) and the end address (the start address of the next section, that is, the address of the data exceeding the allowable upper limit value or the allowable lower limit value) The address 24c after the first address) is supplied to the overall control means 21, and data 24d corresponding to the reference level K is supplied to the memory means 12.

The general control means 21 sequentially generates address information 21b from the designated start address 24b to the end address 24c, and generates a write signal 21c.
Is generated and the data of the corresponding section is updated to the data 24d corresponding to the reference level K. As a result, the amplitude of the data in the section determined to be noise is suppressed.

The signal waveform correcting means 25 recognizes the section length in which the amplitude is suppressed based on the start address 24b and the end address 24c of the section of the corresponding amplitude, and determines the attenuation section corresponding to the section length in which the amplitude is suppressed. Set before and after the suppression section. The length of the attenuation section is, for example, 1/3 of the suppression section.
〜. This is because even if a noise component is included before and after the suppression section, sudden noise has a sharp rise and fall in its signal waveform, so that it can be attenuated over a relatively short time range. That's because. When sudden noise having a long wave tail is to be removed, it is preferable to appropriately set the attenuation section according to the characteristics of the waveform. Then, the signal waveform correction means 25
While setting the address range corresponding to the attenuation section,
The information 25a relating to those address ranges is supplied to the overall control means 21 and the data 12b of the attenuation section is stored in the memory means 1.
2, and outputs data 25 b obtained by correcting the read data.
c is supplied to the overall control means 21 and the corrected data 25b
Is written into the memory means 12.

In the above-described attenuation correction, the connection between the original beat signal and the signal determined as noise and suppressed is shown in FIG.
As long as it is smooth as shown in the figure, for example,
As shown in FIG. 4B, a simple decrease (the front part of the suppression section) or a simple increase characteristic (the rear part of the suppression section) may be used, or an exponential decrease or increase characteristic shown in FIG. 4C may be used. .

When a series of data processing for noise removal is completed, the overall operation control means 21 sends a digital operation processing request 21f to the digital operation processing means 13 and, at the same time, sends a digital beat after data correction from the memory means 12. The signal data 12b is read. The digital operation processing means 13 outputs the corrected data 12b
, The frequency analysis and the calculation of the distance to the object are performed, and the measurement result 13a is displayed on the display device 8.

FIG. 5 is an explanatory diagram showing the operation of the noise suppression processing means. First, the beat signal is sampled in step S1, and the digital beat signal data DS is stored in the memory means 12. Then, in step S2,
The amplitude density is calculated by the amplitude density calculating means 22.
When the amplitude waveform of the beat signal BS includes the sudden noise N as shown in FIG.
(B). In FIG. 5B, the horizontal axis represents the amplitude level, and the vertical axis represents the amplitude density obtained by normalizing the number of occurrences of each amplitude by the number of occurrences of all amplitudes. In the case of the signal shown in FIG. 5A, the occurrence period of the sudden noise N is a very short period within the analysis target period, and therefore the amplitude level of the original beat signal becomes the peak of the density distribution. In FIG. 5B, the spread of the peaks of the density distribution indicates the spread of the density. In this case, since the noise N having a large amplitude is included, the spread of the larger amplitude is larger than the peak portion of the density. I have.

When the amplitude density is obtained, in step S3, the noise determination threshold value setting means 23 calculates the threshold amplitude shown in FIG. 5C and sets the allowable upper limit value and the allowable lower limit value. In step S4, the noise amplitude suppression unit 24 changes data exceeding the allowable value range to data corresponding to zero amplitude. Further, the signal waveform correcting means 25 sets the section before and after the amplitude is changed to data equivalent to zero as an attenuation section, and the original beat signal and the part whose amplitude is changed to zero are smoothly connected to the data in this attenuation section. Correct the data as follows. As a result, the digital beat signal data corresponding to the signal waveform shown in FIG.
Will be stored in

Then, in step S5, the overall operation control means 21 activates the digital operation processing means 13 and
Since the digital beat signal data corresponding to the signal waveform shown in (d) is supplied to the digital arithmetic processing means 13, the digital signal such as the fast Fourier transform (FFT) is obtained based on the signal data from which the sudden noise N has been removed or reduced. The processing is performed, the distance to the object is obtained based on the frequency analysis result, and the distance is displayed on the display device.

In FIG. 1, the amplitude density calculating means 22 for obtaining the amplitude density of the digital beat signal within the period to be analyzed is shown.
And a noise determination threshold value setting means 23 for setting a noise determination threshold value based on the obtained amplitude density. The digital signal data of the section of the amplitude exceeding the noise determination threshold value has an amplitude of zero or close to zero. Although the noise suppression processing means 14 including the noise amplitude suppression means 24 for changing the data of the digital beat signal so as to have a small amplitude and the signal waveform correction means 25 is shown, the noise suppression processing means is shown in FIG.
As shown in (1), a configuration may be adopted in which the envelope of the signal waveform is obtained, and the signal in the section where the envelope exceeds the noise determination threshold value VTH is suppressed. Further, the average value of each amplitude waveform within the analysis target period is obtained without obtaining the amplitude density, and a process of reducing the amplitude to zero or suppressing the amplitude is performed for a section having an amplitude larger than the average value by a predetermined amount. Is also good. Further, an attenuation section may be provided before and after the suppression section.

Also, the configuration has been described in which the writing of the sampling signal DS to the memory means 12 and the calculation of the amplitude density are performed simultaneously. However, if there is a time margin before the distance measurement,
An arrangement may be made wherein the amplitude density is calculated for the data stored in the memory means 12. Noise amplitude suppression means 2
4 without providing the memory 24M, the noise determination threshold L
After the TH is set, data may be read again from the memory unit 12 and the section exceeding the threshold LTH may be detected while calculating the amplitude of each waveform again.

In this embodiment, all data that continues in time including the data in the suppression section and the attenuation section are supplied to the digital arithmetic processing unit 13, but the time for distance measurement is shorter than the measurement accuracy. If priority is given, a means for storing the address information of the suppression section and the attenuation section is provided in the overall operation control means 21, and the data excluding the suppression section or the data excluding the suppression section and the attenuation section are sent to the digital arithmetic processing means 13. You may make it supply. In the case of a configuration that removes data in such a noise mixed section,
Since there is no need to update the data in the memory means 12,
While the configuration of the noise suppression processing means 14 is simplified,
Since the data update process is not required, the time required for distance measurement can be reduced.

Next, the effects of the noise suppression processing means according to the present invention will be described with reference to specific examples. FIG. 7 is a signal waveform diagram in the case where noise having a large amplitude is added to a certain section. When a frequency spectrum is obtained from this signal waveform by fast Fourier transform (FFT), noise frequency components appear in addition to the original signal (frequency components proportional to the distance to the object) as shown in FIG. The accuracy of is reduced. Thus, the noise portion is suppressed using the noise suppression processing means 14 shown in FIG. 1 and attenuation processing is performed before and after the suppression section to obtain a signal waveform shown in FIG.
When the sampling data relating to the signal waveform is subjected to fast Fourier transform (FFT) by the digital arithmetic processing means 13, an original signal (a frequency component proportional to the distance to the object) appears remarkably, as shown in FIG. Therefore, highly accurate distance measurement becomes possible.

[0037]

As described above, the FM according to claim 1 is used.
-The signal processing device of the CW radar includes noise suppression processing means for changing data so that the amplitude of the section becomes zero or a small amplitude close to zero in the section where the amplitude is larger than the other levels. Sexual noise can be removed or suppressed. Therefore, the digital beat signal from which sudden noise has been removed or suppressed is subjected to digital arithmetic processing such as fast Fourier transform, and the distance is obtained from the frequency analysis result. Becomes possible.

[0038] The noise suppression processing means of the signal processing device of the FM-CW radar according to claim 2 sets a noise determination threshold value based on the amplitude density, and sets a noise determination threshold value for a section exceeding the set noise determination threshold value. Because the amplitude of the section is changed to zero or data of small amplitude close to zero, the noise judgment level can be changed according to the amplitude of the beat signal, and suddenness can occur regardless of the level of the beat signal. Noise can be effectively removed or attenuated. Since a signal-to-noise ratio equal to or higher than a predetermined value can be ensured, more reliable distance measurement can be performed.

A signal waveform correcting means is provided to correct the signal waveforms on both sides of the suppression section so that the non-suppression signal and the suppression signal are smoothly connected to each other, so that the initial noise and the tail of the sudden noise have small amplitudes. Even if there is a part, it can be suppressed. Even if the amplitude of the beat signal fluctuates, the noise before and after the noise suppression section is attenuated even if the noise judgment threshold value is set to be rather high so as not to suppress the peak portion of the beat signal. Therefore, the noise component can be effectively suppressed.

[Brief description of the drawings]

FIG. 1 is a functional block configuration diagram of a signal processing device of an FM-CW radar according to the present invention.

FIG. 2 is an overall block configuration diagram of an FM-CW radar device including the signal processing device.

FIG. 3 is an explanatory diagram showing an example of a method of obtaining an amplitude.

FIG. 4 is an explanatory diagram showing a specific example of a signal waveform and an attenuation characteristic in an attenuation section.

FIG. 5 is an explanatory diagram showing an operation of a noise suppression processing unit.

FIG. 6 is an explanatory diagram showing the operation of another configuration example of the noise suppression processing means.

FIG. 7 is a signal waveform diagram when a large amplitude noise is added to a certain section.

8 is an explanatory diagram showing a frequency spectrum of the signal waveform shown in FIG. 7;

FIG. 9 is a signal waveform diagram in which noise is suppressed by noise suppression processing means and attenuation processing is performed on attenuation sections before and after the noise.

FIG. 10 is an explanatory diagram showing a frequency spectrum of the signal waveform shown in FIG. 9;

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 FM-CW radar device 7 signal processing device 11 A / D converter 12 memory means 13 digital calculation processing means 14 noise suppression processing means 21 overall operation control means 22 amplitude density calculation means 23 noise determination threshold setting means 24 noise amplitude Suppression means 25 Signal waveform correction means

Continuation of the front page (56) References JP-A-5-249233 (JP, A) JP-A-3-175389 (JP, A) JP-A-4-142486 (JP, A) JP-A-4-348292 (JP) (A) JP-A-4-265882 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01S 13/32-13/48 G01S 13/88-13/95

Claims (3)

(57) [Claims] 1. A digital signal processing device performs digital arithmetic processing on a digital beat signal obtained by sampling and quantizing a beat signal obtained by mixing a transmission signal and a reception signal, and performs frequency analysis based on the frequency analysis result. In a signal processing device of an FM-CW radar for obtaining a distance to an object, a memory means for temporarily storing the data of the digital beat signal, and the amplitude of the data stored in the memory means is higher than other levels. An FM comprising noise suppression processing means for changing data so that the amplitude of a large section becomes zero or a small amplitude close to zero.
A signal processing device for a CW radar. 2. The noise suppression processing means according to claim 1, wherein said noise suppression processing means calculates an amplitude density of the digital beat signal within a period to be analyzed, and a noise determination threshold for setting a noise determination threshold based on the determined amplitude density. Value setting means, and noise amplitude suppressing means for changing data of the digital signal in the section of the amplitude exceeding the noise determination threshold so that the amplitude becomes zero or a small amplitude close to zero. The signal processing device for an FM-CW radar according to claim 1. 3. The noise suppression processing means sets attenuation sections on both sides of a section where data is changed so that the amplitude becomes zero or a small amplitude close to zero, so that the non-suppression signal and the suppression signal are smoothly connected. 2. The signal processing device for an FM-CW radar according to claim 1, further comprising a signal waveform correction unit configured to correct data of the digital signal in the attenuation section.