JPH0342433B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Technical Field) The present invention relates to a logarithmic sensor array whose main pole width is constant within a reception band by weighting sensor output signals.

(Background Art) In sonar, radar, etc., a sensor array consisting of a large number of sensors is used to form a receiving beam with sharp directivity to detect, position, and classify targets. In order to improve the accuracy of target detection, positioning, and classification, it is important to sharpen the directivity of the receiving beam and at the same time keep the main pole width constant within the receiving band. "Generally, the transmitting and receiving sensitivity with respect to the direction of the sensor array is expressed as a directivity pattern. In this pattern, the pole (beam) including the beam axis is the main pole.
The poles attached to them are called side poles.
lobes), and they are placed in order from the one closest to the main pole,
The second... is called the sub-pole. Generally, the beam used by sonar and radar is the main pole. The main pole width is the width of the main pole, also called beam width, and is the angle between the directions at which the sensitivity is -3 or -6 dB of the beam axis. The width of the main pole is inversely proportional to the aperture length of the sensor array and has a characteristic that it is proportional to the wavelength.

Next, the advantage of keeping the main pole width constant will be explained.

When an array is used to receive signals from a target and detect a target, the azimuth accuracy of detection (accuracy in knowing in which direction the target is located) depends on the external environment such as the signal-to-noise ratio of the signal from the target. There is also an influence on the sensor array, but the width of the main pole and the level of the sub-pole have an effect on the sensor array.

Equidistant sensor arrays, which are commonly used as arrays, use a window function such as the Kaiser-Betzel function to lower the sub-pole level and improve the azimuth accuracy of detection. That is, the aperture length can be changed by changing the window function. Therefore, the window function is changed (normalized) so that the product of the aperture length and the wavelength is constant for the wavelength.
This makes the main pole width constant. Efforts are also being made to increase the length of the array in order to reduce the width of the main pole and improve the azimuth accuracy of detection.

However, the equally spaced sensor array has the property that the main pole width changes depending on the frequency (the lower the frequency, the larger the main pole width, and the higher the frequency, the smaller the main pole width). Therefore, a uniformly spaced sensor array will depend on the frequency of the signal from the target.
It has the disadvantage that the detection accuracy changes (the lower the frequency, the lower the accuracy).

On the other hand, if a pseudo-logarithmic sensor array or a logarithmic sensor array is used, the main pole width can be made constant within the reception band, and the azimuth accuracy of detection can be made constant within the reception band. Pseudo-logarithmic sensor arrays have been devised for this purpose;
This array is a combination of equally spaced sensor arrays, and the sum of squares of the weights of each sensor does not change smoothly with frequency, but has irregularities. Therefore, the gain for isotropic noise has irregularities with respect to frequency, which is an obstacle to target detection and classification. As mentioned above, it is possible to keep the main pole width constant by changing the window function, but for the reasons described below, the window function is usually not changed, and the higher the frequency, the smaller the main pole width becomes. Become. If the sensor array's sensor spacing exceeds 1/2 wavelength, a grating globe will occur, so the spacing should be 1/2.
Must be 2 wavelengths or less. If the main pole width of an equally spaced sensor array of N sensors with a spacing of λ ₀ /2 at frequency 0 is θ ₀ , then 2N sensors are required for the main pole width to be θ ₀ at frequency ₀ / ₂ . Become. In other words, 2N sensors are required to keep the main pole width constant from frequency _0/2 to ₀ . On the other hand, if a pseudo-logarithmic sensor array or a logarithmic sensor array is used, similar characteristics can be obtained with about 1.5N sensors. Therefore, when the main pole width is made constant by changing the window function, the equally spaced sensor array is uneconomical and is not used.

A typical example of the configuration of a conventional pseudo-logarithmic sensor array is shown in FIG. 1 _-k ~1 _k is the sensor, 2 _-k ~
2 _k is a preamplifier, 3 _-k to 3 _k is an equalizer, 4 _-k to 4 _k
is the output terminal of the sensor array. Note that K is 1/2 of the number of sensors. In an example of a conventional pseudo-logarithmic sensor array arrangement, the same number of sensors are arranged symmetrically about a central axis. A pseudologarithmic sensor array is constructed by combining a plurality of equally spaced sensor arrays. For example, the first stage is an array of equally spaced sensors arranged at half-wavelength intervals for frequency ₀ (in this example, consists of 32 sensors), and the second stage is arranged at half-wavelength intervals for frequency ₂₀ . The third stage is an equally spaced sensor array (here, composed of 32 sensors), and the third stage is an equally spaced sensor array (here, composed of 32 sensors) arranged at half-wavelength intervals for the frequency 4 ₀ . ), the fourth stage has a reception band of
3 above with a ₀ to 4 ₀ pseudologarithmic sensor array
It consists of two equally spaced sensor arrays (in this case, it consists of 64 sensors). The equalizer characteristics of a conventional pseudo-logarithmic sensor array are, for example, one sensor for frequencies ₂₀ , some sensors for frequencies ₄₀ , and some sensors for frequencies ₀ to ₄₀ . have against.

The pseudo-logarithmic sensor array is used in combination with a beamformer to form a receive beam. An example of the configuration of the beamformer is shown in FIG.
5 _-k to 5 _k are input terminals, 6 _-k to 6 _k are delay lines, 7 is an adder, and 8 is an output terminal. The output terminals 4 _-k to 4 _k shown in Fig. 1 are the input terminals 5 _-k to 5 _k shown in Fig. 2.
connected to. In FIG. 1, signals propagated through the air or water are input to sensors 1 _-k to 1 _k and converted into electrical signals. The electrical signal is amplified by preamplifiers 2 _-k to _2k and equalizers 3 _-k to 2k.
3 _k and output to output terminals 4 _-k to 4 _k . In Figure 2, the output terminal 4 in Figure 1
_-k ~ 4 The electrical signal output from _k is input terminal 5 _-
k to 5 _k , the propagation delay is compensated by delay lines 6 _-k to 6 _k , and the sum is captured by adder 7,
It is outputted from the output terminal 8 as a receiving beam.

When the sum of squares of the equalizer characteristics (weights) of the sensors constituting the pseudo-logarithmic sensor array is calculated, in the example shown above, the result is as shown in FIG. 5. Looking at the change in the sum of squares with respect to the change in frequency, the change is not smooth at frequencies ₀ and ₂₀ , and the characteristics as a whole are uneven. The gain for a single sensor in a sensor array is expressed as the sum of squares of weights for isotropic noise. Therefore, in a pseudo-logarithmic sensor array, the gain for isotropic noise does not change smoothly with frequency, but has irregularities.

Target detection and identification uses the target signal superimposed on isotropic noise, so pseudo-logarithmic sensor arrays may incorrectly detect a target when it is not present, or conversely detect a target at frequencies around ₀ and ₂₀ . However, there is a problem in that it cannot be detected because it is a characteristic of the sensor array.

(Problems to be solved by the invention) The present invention has been made to eliminate the above-mentioned drawbacks, and by effectively utilizing a well-known window function used for weighting a sensor array, the frequency of a received signal can be adjusted. The purpose of the present invention is to provide a logarithmic sensor array whose main pole width is constant regardless of the width of the main pole.

(Structure and operation of the invention) FIG. 3 is a block diagram showing an embodiment of the invention.
9 _-k to 9 _k are sensors, 2 _-k to 2 _k are preamplifiers,
10 _{-k to 10k} are equalizers that weight the output signals of the sensors 9 _-k to _9k to correct uneven spacing; 4 _-k
~ _4k is an output terminal. A _-k () to A _k () represent frequency characteristics of the equalizers 10 _-k to 10 _k . In the present invention, an equalizer 10 is used as a means for normalizing the window function.
_-k to _10k has frequency characteristics.

FIG. 4 shows equalizers 10 _-k to 10 _k shown in FIG.
FIG. 2 is a diagram for explaining a method of setting frequency characteristics A _-k () to _{A k} (). The horizontal axis is sensor 9 _-k ~9
The position of _k is shown, and the vertical axis represents the value of the window function.
B _-k ~B _k is the distance from the center of the logarithmic sensor array to each sensor, and B _j =B _-j =1-r|j|/1-rd-d/2 (j=1
~K). where d is the minimum spacing between sensors and r>1.

Let the length C _M be C _M =E×D/ _M for the reception frequency _M. Here, D is the propagation speed of the signal, and E is a constant related to the length of the array. At this time, it is assumed that sensors 9 _-n to 9 _n are included within a distance C _M from the center of the array.

The curve in Figure 4 has a length of 2× as an example of a window function.
The Kaiser-Betzell function for C _M is shown,
a ₁ , a _j , and a _n represent the values of the Kaiser-Betzel functions corresponding to the sensors 9 ₁ , 9 _j , and 9 _n , respectively.

Characteristics of the above equalizer 10 _-k to 10 _k at frequency _M
A _-k ( _M ) ~ A _k ( _M ) is A ₁ ( _M ) = A _-1 ( _M ) = 1/2 (B ₁ + B ₂ ) a ₁ × F ( _M ) A _j ( _M ) = A _-j ( _M ) -1/2 (B _j+1 -B _j-1 ) a _j F ( _M ) (j=2, 3,..., m-1), A _n ( _M ) A _-n ( _M ) = {C _M -1/2 (B _n-1 + B _n )} a _n × F ( _M ) A _k ( _M ) = A _{- k} ( _M ) = 0 (k = m + 1, ..., K) . Since the frequency _M is an arbitrary frequency within the reception band, the frequency characteristics A _-k () to A _k () of the equalizers 10 _-k to 10 _k are set according to the above formula. Here, F( _M ) is used to make _k 〓 ^j=1 A _i ()=constant. In the above formula, 1/2 (B ₁ + B ₂ ) × a ₁ is a rectangle whose length and width are a ₁ and 1/2 (B ₁ + B ₂ ), and 1/2 (B _j+1 −B _j-1 ) × a _j is a rectangle whose length and width are a _j and 1/2 (B _j+1 − B _j-1 ), {C _M −1/2 (B _n-1 + B _n )}×a _n whose length and width are a _n , { C _M -1/2(B _n-1 + B _n )}, so the Kaiser shown in Figure 4
The weight distribution expressed by the Betzel function is the above 2m
This means that it is approximated by rectangles. Since the sensor spacing is used with the rectangle as the horizontal side,
The Kaiser-Betzel function can be approximated regardless of the interval. Furthermore, since the length C _M is inversely proportional to the frequency _M , that is, proportional to the wavelength, if the horizontal axis in FIG. 4 is taken as a unit of wavelength, the same weight distribution will be obtained for the frequencies within the band. Therefore, the equalizer 10 according to the above formula
The characteristics A _-K () to A _K () of the characteristics of _−K to 10 _K are the weight values that correct the uneven spacing of the sensor,
The directivity is the same near the main pole as the equally spaced sensor array weighted according to the Kaiser-Betzel function, and the width of the main pole is constant with respect to changes in frequency in the logarithmic sensor array. In addition, when the equalizer has the above characteristics, the weight of each sensor changes smoothly with respect to frequency, so the sum of squares of the weights also changes smoothly, and the frequency characteristics of the above logarithmic sensor array with respect to isotropic noise. becomes smooth.

In other words, in the case of the equalizer characteristics as shown in Figure 3, the sum of squares of the weights of the logarithmic sensor array will be as shown in Figure 6, and will change smoothly in response to changes in frequency, with no unevenness in the characteristics. can not see. Therefore, the disturbances encountered with pseudo-logarithmic sensor arrays are eliminated.

In FIG. 3, signals propagated through the air or water are converted into electrical signals by sensors 9 _-K to _9K , amplified by preamplifiers 2 _{- K} to 2K,
The signals are weighted by equalizers 10 _-K to 10 _K and output to output terminals 4 _-K to 4 _K. The output signal is converted into a receiving beam by the beam former shown in FIG. Since the input of the beamformer is weighted by the equalizer, the main pole width of the receive beam is constant with respect to frequency. Furthermore, since the sum of squares of the weights changes smoothly with respect to frequency, the frequency characteristic of the logarithmic sensor array with respect to isotropic noise becomes smooth. Therefore, the disturbances that occur with the pseudo-logarithmic sensor array do not occur with the logarithmic sensor array of this embodiment.

As explained above, in the embodiment of the present invention, the Kaiser-Betzell function is normalized so that the length of the array with respect to the wavelength is constant, and the value obtained by multiplying by the sensor interval is used as the characteristic of the equalizer. By doing so, the width of the main pole remains constant despite changes in frequency, creating a logarithmic sensor array with smooth frequency characteristics against isotropic noise.

(Effects of the Invention) As explained above, according to the present invention, a well-known window function such as the Kaiser-Betzell function is used for weighting the logarithmic sensor array so that the length of the array with respect to the wavelength of the received signal is constant. Since it is standardized and corrected using the sensor spacing, the width of the main pole remains constant regardless of the frequency of the received signal, making it possible to realize a sensor array with smooth frequency characteristics against isotropic noise.

[Brief explanation of drawings]

FIG. 1 is a block diagram showing a conventional logarithmic sensor array, FIG. 2 is a block diagram showing a beamformer, FIG. 3 is a block diagram showing an embodiment of a logarithmic sensor array according to the present invention, and FIG. 4 is a block diagram showing the frequency of the embodiment. An explanatory diagram of a characteristic setting method, FIG. 5 is a diagram showing the sum of squares of weights of a conventional sensor, and FIG. 6 is a diagram showing a sum of squares of weights of a sensor according to the present invention. 1 _-K ~ 1 _K ...sensor, 2 _-K ~ 2 _K ...preamplifier, 3 _-K ~ 3 _K ...equalizer, 4 _{-K ~} 4 _K ...output terminal, 5 _{-K ~} 5 _K ...Input terminal, 6 _-K ~6 _K ...Delay line, 7...Adder, 8...Output terminal, 9 _{-K ~} 9
_K ...sensor, 10 _-K ~10 _K ...equalizer.

Claims (1)

[Claims] 1. Standardize multiple sensors arranged at logarithmic intervals and the window function used for weighting the sensor array so that the length of the sensor array is constant for the wavelength of the received signal, and then A plurality of equalizers are provided corresponding to each of the sensors, and the gain is a value corrected based on the distance between the sensors, and the output signal of each of the sensors is weighted by each of the equalizers. A logarithmic sensor array featuring