JP2002311126A - Radar signal processor and radar system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To suppress the increase of the noise resulting from an inverse filtering process of radar signals, thereby improving the resolution even for signals of a low S/N ratio. SOLUTION: For radar received signals a plurality of Wiener inverse filters 12-1 to 12-n are provided in parallel, and Fourier transformation coefficients αof transfer functions thereof are set to different values ranging from 0 to 1. After filtering the signals, a geometric mean processor 13 takes a geometric mean value of the filtered signals, this indicating the synergistic effect of the resolution improving effect of smaller α with that of larger α. Thus, it makes possible to raise the resolution for near-range targets at a low noise for radar received signals of low S/N ratio.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

The present invention belongs to the technical field of signal processing for improving a target resolution of a signal received by a radar device.

[0002]

2. Description of the Related Art In general, the range resolution of a radar apparatus is determined by a transmission pulse width and a reception bandwidth of a receiver, and the angular resolution is determined by a beam width determined from an antenna aperture length. Therefore, conventionally, in order to improve the resolution, it has been practiced to reduce the transmission pulse width, widen the reception bandwidth, and increase the antenna aperture length.

Conventionally, as a signal recovery technique, a method of increasing the resolution by solving an inverse problem of an observation system such as deconvolution or an inverse filter has been known. Further, a method using a single Wiener inverse filter as described below has been known. Now, in the radar apparatus, when the observation signal is y, the true target distribution is x, the transfer function of the apparatus (determined by the antenna beam pattern, the transmission pulse signal, the reception bandwidth, and the like) is h, and the observation noise is n, the observation signal y Is represented by Equation 1. Here, * indicates convolution integration.

[0004]

## EQU1 ## y = h ^* x + n

If it is assumed that the observation noise n is sufficiently small and the observation signal y is Fourier transformed to Y, the convolution integral in the time domain becomes a product in the frequency domain.
Holds.

[0006]

## EQU2 ## Y = HX

H and X represent Fourier transform pairs of h and x, respectively. From now on, solving for X gives Equation 3.

[0008]

X = Y / H

By performing the inverse Fourier transform, a true target distribution x can be obtained. However, when the observation noise n cannot be neglected, an increase in noise is observed due to this effect, and a desired result cannot be obtained. As a solution to this, it is known to use a Wiener inverse filter incorporating the concept of a Wiener filter that maximizes the signal power to noise power ratio (S / N ratio) while trying to increase the signal resolution. ing. The Fourier transform F of the transfer function of this filter is represented by Expression 4. α is a coefficient.

[0010]

(Equation 4)

[0011]

However, in order to improve the resolution, narrowing the transmission pulse width to increase the reception bandwidth or increasing the antenna aperture length are naturally limited by electrical and physical restrictions. There is. Further, a method of increasing the resolution by solving an inverse problem of an observation system such as deconvolution or an inverse filter, when applied to a radar device, has a low S / N ratio of a received signal, and not only improves the resolution but also makes the inverse problem inappropriate. Due to the increase in noise (generation of false images) and the need for real-time performance due to the problem, complicated processing is difficult and practically difficult.

In the method using the Wiener inverse filter, in the Fourier transform shown in Expression 4, α =
The case of 0 is an inverse filter, and the case of α = 1 is a matched filter. Although the resolution can be improved with the inverse filter, the S / N ratio is deteriorated, while the S / N ratio is improved with the matched filter. However, it has a characteristic that the resolution is deteriorated. When α is between 0 and 1, it has intermediate characteristics.

Therefore, in the case of a radar reception signal having a very low S / N ratio, the value of α must be increased to improve the S / N ratio, and as a result, it becomes difficult to improve the resolution. If the value of α is approached to 0 to improve the resolution, noise is emphasized. As a result, there is a problem that a practical result cannot be obtained with a single Wiener inverse filter.

An object of the present invention is to provide a radar receiving signal having a low S / N ratio by using a plurality of Wiener inverse filters in view of the above-mentioned problems of the prior art.
It is an object of the present invention to provide a radar signal processing device capable of suppressing an increase in noise caused by inverse filtering and improving resolution.

[0015]

In order to achieve the above object, a first configuration of the radar signal processing apparatus according to the present invention is characterized by comprising the following means. (A) First scan memory for storing digitized radar video signals (b) Different coefficients α for performing inverse filtering on data in the distance direction or angle direction read from the scan memory Multiple winner inverses
A filter (c) a geometric meander for calculating a geometric mean of the outputs of the plurality of Wiener inverse filters (d) a second scan memory for storing the outputs of the geometric meander

According to a second configuration of the radar signal processing apparatus of the present invention, in addition to the first configuration, a threshold value for performing threshold processing on the output of each Wiener inverse filter before sending it to the geometric mean. A processor is provided.

The radar signal processing apparatus of the present invention may be used alone or may be incorporated as part of a radar apparatus.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of the present invention uses a plurality of Wiener inverse filters as described above, so that even a radar reception signal having a low S / N ratio can be used.
It is intended to improve the resolution while suppressing an increase in noise caused by the inverse filter processing. As mentioned above,
The inverse filter is a Fourier transform shown in Expression 4, and the inverse filter with the coefficient α = 0 improves the resolution, but degrades the S / N ratio, and conversely, the coefficient α = 1
In the case of the matched filter, the S / N ratio is improved, but the resolution is deteriorated.

Therefore, in the first configuration of the present invention, a plurality of Wiener inverse filters having large and small values including the coefficients 0 and 1 within the range of 0 and 1 are provided. Resolution improvement of inverse filter and S of Wiener inverse filter with large α value
To achieve both the effect of improving the / N ratio,
The arithmetic mean of the output signal of the inverse filter is obtained.

In the second configuration, threshold processing is performed on the output of each Wiener inverse filter.
The noise signal is removed and the desired signal is enhanced. The radar signal processing device according to the present invention may be used separately from the radar device itself, or may be used as a part of the radar device.

[0021]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a radar signal processing apparatus according to the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of an embodiment of the first configuration of the present invention. The portion surrounded by the dotted line is the configuration of the embodiment of the radar signal processing device 2a of the present invention. The radar video signal sent from the radar sensor 1 is converted into a digital signal by the A / D converter 10 and stored in the scan memory 11. In this example,
Although the A / D converter 10 is included in the radar signal processing device 2a, the A / D converter 10 is not always necessary, and a digitized radar video signal may be input.

The data is read out from the scan memory 11 as data connected in the distance direction or data connected in the angular direction focused on the resolution, and sent to a plurality of Wiener inverse filters 12-1 to 12-n. The plurality of Wiener inverse filters are set so that the value of the coefficient α in Equation 4 is varied from 0 to 1.

That is, the values are set so as to be dispersed from a small value for improving the resolution to a large value for improving the S / N ratio. In this way, the output having the effect of improving the resolution and the output having the effect of improving the S / N ratio are combined with the geometric mean 13
, Where the geometric mean is determined. As a result, the output of the geometric mean averaging unit 13 has a large S / N ratio that suppresses an increase in noise due to the inverse filter processing while improving the resolution, even if the received signal has a low S / N ratio. A signal will be obtained. The output of the geometric mean averaging device 13 thus obtained is temporarily stored in the scan memory 14 and then provided for display on the display device 3.

FIG. 2 shows a configuration in which threshold processors 15-1 to 15-n are provided at the output of each Wiener inverse filter in addition to the configuration of FIG. This allows
Each Wiener inverse filter 12-1 to 12-n
Of the outputs, the output lower than the predetermined threshold level is cut off and output as if there was no signal output. Thereby, the effect of noise suppression is further improved. The outputs of the threshold value processors 15-1 to 15-n are stored in the scan memory 14 after the geometric mean is obtained by the geometric averager 13.

Next, in the configuration of FIG. 1, the number of Wiener inverse filters is set to four (n = 4), and the value of the coefficient α is changed from 0.02 close to the inverse filter to 0.1, 0.1,.
The simulation results when 0.5 and 1 (matched filter) are set are described below.

First, a true target distribution is assumed as shown in FIG. When the transfer function in the simulation is an antenna beam pattern as shown in FIG. 4 and noise is added to generate an observation signal, the result is as shown in FIG. FIG. 6 shows that α = 0.02
5 is a result obtained by filtering the observation signal of FIG. 5 with the Wiener inverse filter described above. Similarly, FIG. 7, FIG. 8, and FIG. 9 show the results of filtering by filters with α = 0.1, α = 0.5, and α = 1, respectively. FIG. 10 shows the result of geometric averaging of the outputs of these filters. The results show that the effects of improving the resolution and suppressing noise (improving the S / N ratio) of the target closer to the observation signal in FIG. I understand.

Next, the simulation results when threshold processing is performed as in the configuration of FIG. 2 will be described. FIG. 11 shows the true target distribution. FIG. 12 shows a transfer function. FIG. 13 shows an observation signal generated by adding noise to the transfer function of FIG. FIG. 14, FIG. 15, FIG. 16, FIG.
7 is a diagram when threshold processing is performed on each filter output of α = 0.02, α = 0.1, α = 0.5, and α = 1, and the broken line is before threshold processing and the solid line is threshold processing. Later.
FIG. 18 shows the result of geometric mean processing of each output after threshold processing. As can be seen from FIG. 10, the noise is further reduced.

[0028]

As described above, the radar signal processing apparatus of the present invention uses a plurality of Wiener inverse filters in parallel to perform an inverse filter process on an input signal, and uses a transfer function of its transfer function. After setting the coefficient α of the Fourier transform equation to a value different from each other in the range of 0 to 1 and performing filtering, a geometric mean of them is obtained, so that the resolution improvement effect of the smaller α and the resolution improvement of the larger α There is an advantage that the noise suppression effect (S / N ratio improving effect) exhibits a synergistic effect, and a signal processing result with high resolution and high S / N ratio can be obtained for a target in proximity. In addition, there is an advantage that different noise suppression effects can be obtained by performing threshold processing on the filter output.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of a first configuration of a radar signal processing device of the present invention.

FIG. 2 is a block diagram of a second embodiment of the radar signal processing apparatus according to the present invention;

FIG. 3 is a diagram showing a true target distribution assumed in a simulation corresponding to the configuration of FIG. 1;

FIG. 4 is a diagram illustrating a transfer function of the radar apparatus set in a simulation corresponding to the configuration of FIG. 1;

FIG. 5 is a diagram showing an observation signal obtained by adding noise to the transfer function of FIG. 4;

FIG. 6 is a diagram illustrating the result of the observation of FIG.
It is a figure showing the result of having filtered with an inverse filter.

FIG. 7 is a diagram illustrating the results of the Wiener method in which the observed signal of FIG.
It is a figure showing the result of having filtered with an inverse filter.

FIG. 8 is a diagram illustrating the result of the observation of FIG.
It is a figure showing the result of having filtered with an inverse filter.

FIG. 9 is a diagram illustrating a result of filtering the observation signal of FIG. 5 by a Wiener inverse filter (matched filter) with α = 1.

FIG. 10 is a diagram showing a result of geometric mean processing of the filtering results of FIGS. 6, 7, 8 and 9;

11 is a diagram illustrating a true target assumed in a simulation corresponding to the configuration of FIG. 2 (when performing threshold processing);

FIG. 12 is a diagram illustrating a transfer function of the radar apparatus set in a simulation corresponding to the configuration of FIG. 2 (when performing threshold processing).

FIG. 13 is a diagram illustrating an observation signal obtained by adding noise to the transfer function of FIG. 12;

FIG. 14 is a diagram showing a result obtained by filtering the observation signal of FIG. 13 with a Wiener inverse filter with α = 0.02 and then performing threshold processing.

FIG. 15 is a diagram showing a result of filtering the observation signal of FIG. 13 with a Wiener inverse filter with α = 0.1 and then performing threshold processing.

FIG. 16 is a diagram showing a result obtained by subjecting the observation signal of FIG. 13 to a threshold value process after filtering by a Wiener inverse filter with α = 0.5.

FIG. 17 is a diagram illustrating a result of performing threshold processing after filtering the observation signal of FIG. 13 with a Wiener inverse filter (matched filter) in which α = 1.

FIG. 18 is a diagram showing the result of geometric mean processing of the results of FIGS. 14, 15, 16 and 17;

[Explanation of symbols]

 REFERENCE SIGNS LIST 1 radar sensor 2 a radar signal processing device 2 b radar signal processing device 3 display device 10 A / D converter 11 scan memory 12-1 to 12-n Wiener inverse filter 13 geometric meanr 14 scan memory 15-1-15- n threshold processor

Claims (3)

[Claims] 1. A radar signal processing apparatus comprising: (A) First scan memory for storing digitized radar video signals (b) Different coefficients α for performing inverse filtering on data in the distance direction or angle direction read from the scan memory Multiple winner inverses
A filter (c) a geometric meander for calculating a geometric mean of the outputs of the plurality of Wiener inverse filters (d) a second scan memory for storing the outputs of the geometric meander 2. In addition to the configuration of claim 1, each winner
A radar signal processing device comprising a threshold value processor for performing a threshold value process on an output of an inverse filter before sending it to a geometric mean. 3. A radar device comprising the radar signal processing device according to claim 1.