JP3007914B2 - Mode eigenvalue measurement method - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention is used in a sound wave propagation simulation method for calculating sound wave propagation characteristics in the ocean, and estimates a sound velocity distribution in a seabed sediment, which is an environmental condition of an input signal to a sound wave propagation characteristic calculation device. The method relates to a mode eigenvalue measuring method for measuring a normal mode eigenvalue necessary for the above.

[0002]

2. Description of the Related Art Conventionally, techniques in such a field include:
For example, there is one described in the following literature. Reference 1: J. Acoust. Soc. Am., 82 (3) (1987), Acoustical S
ociety of America, (US), SDRajan, et al., "Pertu
rbative inversion methods for obtaining bottom geo
acoustic parameters in shallowwater ", P.998-1017 Reference 2; J.Acoust.Soc.Am., 93 (1) (1993), Acoustical S
ociety of America, (US), SDRajanandS.D.Bhatt
a, "Evaluation on high-resolution frequency method
s for determining frequencies ofeigenmodes in shal
low water acoustic field ", P.378-389 Sound waves in the ocean repeatedly travel in the distance, repeating refraction in the seawater and reflections on the sea surface and the seabed. In particular, shallow oceans (hereinafter shallow sea area) In this case, sound waves that repeatedly reflect on the sea surface and the sea floor become dominant, and in order to accurately grasp this sound field, it is indispensable to grasp the speed of sound in the sea bottom sedimentary layer, which affects the sea bottom reflection loss. Return loss has traditionally been estimated by radiating pulses from a sound source and measuring the level of temporally separated submarine reflected waves, but it is difficult to separate submarine reflected waves temporally in shallow waters. Therefore, Rajan et al. Proposed another estimation method using continuous waves described in Reference 1.

[0003] This is not a method of directly measuring the seafloor reflection loss, but rather measuring a normal mode eigenvalue (hereinafter referred to as a mode eigenvalue) using the Hankel transform from a horizontal sound wave propagation characteristic, and determining a sound velocity distribution of the sound velocity distribution assuming this mode eigenvalue. This is a method of estimating the sound velocity in the seabed sedimentary layer required for calculation of seafloor reflection loss using the inverse problem solution method from the difference from the mode eigenvalue calculated from the initial value. Medium medium characteristics such as temperature and water depth is not changed in the horizontal direction (i.e., stratified medium) wave propagation characteristic P in (r, z _r, z _s) and the wave number spectrum G (k
_r, z _r, z _s), the following equation (1), given respectively Hankel transform pair shown in (2).

[0004]

(Equation 1) Now, as measured in the horizontal direction of the sound wave propagation characteristic _{P (r, z r, z} s), wave number spectrum G (k _r, z _r, z _s) is calculated by equation (2), the wave number spectrum G (k
_r, z _r, z _s) horizontal wavenumber k _r corresponds to the mode eigenvalue having a peak. In addition, the method described in the above reference 2 is
MU instead of Hankel transform in equations (1) and (2)
This is an example in which a high-resolution method such as SIC or ESPRIT is applied.

Here, the principle of a mode eigenvalue measuring method using MUSIC will be described. Sound waves emitted from a sound source moving in seawater are received by a fixed receiver. In this case, the sound source may be fixed and the receiver may be moved. A discrete Fourier transform (DFT) is sequentially performed on the received signal, and a signal of a frequency component in which the Doppler shift is considered in the frequency of the sound source signal is extracted. From the data between the source and the receivers of The extracted signal and position changes every moment, the sound field of distance characteristics (i.e., wave propagation characteristics in the horizontal _{direction) P (r, z r,} z s) is measured Is done. At this time, since the measurement is performed while either the sound source or the receiver moves, the data at different distances also have different reception times, so the phase is corrected using the synthetic aperture method so that the reception signals are at the same time. I do. Then, the distance _{r n (n = 0,1, ...} , N-1) sound for field y (n) = P from _{(r n, z r, z} s) ( sound wave propagation characteristics in the horizontal direction), the covariance Calculate the matrix R _yy . Each component of the covariance matrix R _yy is calculated by the following equation (3).

[0006]

(Equation 2) If the covariance matrix R _yy is M rows and M columns, M eigenvectors [v ₁ , v ₂ ,..., V _M ] are generated by eigenvalue decomposition or singular value decomposition. Therefore, assuming that the number of propagating mode waves is L (L <M), the noise matrix E is defined as E = [v _{L + 1} ,..., V _M ].

For a mode wave having a wave number k, the following equation (4)
Steering vector _{e e = [exp (jkr 0} ), exp (jkr 1), ..., exp (jkr M-1))] shown in ^T ··· (4) where, T; defines the transpose. Each element of this steering vector e is
This represents the phase change of the traveling wave having the wave number k. When the wave number k matches the wave number of mode wave, a correlation value of a noise matrix E and the steering vector ^{e (e H EE H e)} (H; conjugate transpose) takes a small value. Therefore, the reciprocal (e ^H EE ^H e) ^-1 is a large value of the correlation values (e ^H EE ^H e), inverse of the correlation value of the wave number k (e ^H EE ^H e) mode eigenvalue from the maximum point of ^-1 Can be estimated.

[0008]

SUMMARY OF THE INVENTION However, MUSIC
The conventional mode eigenvalue measurement method using the method has the following problems. That is, since MUSIC extracts a signal having a large peak in the wave number space, there is a drawback that if there are many noises having a peak larger than the signal, the signal cannot be detected. Therefore, the conventional method has a problem that sufficient performance cannot be exhibited in an environment including noise. In order to solve this problem, the present invention provides a mode eigenvalue measuring method in which the effect of a noise component having a wave number component different from a signal component is reduced by applying MUSIC to the result of Hankel transform.

[0009]

According to a first aspect of the present invention, there is provided a method for measuring a mode eigenvalue, comprising: a transmitter having one fixed and the other having a uniform linear motion. And a receiver provided in seawater, and a discrete Fourier transform of a received signal generated by receiving a sound wave of a constant frequency output from the transmitter by the receiver was obtained. A sound wave propagation characteristic calculation process for performing a phase compensation for a moving time on a frequency spectrum of a plurality of frequencies to obtain a sound wave propagation characteristic when the transmitter and the receiver are at an equal distance interval, Discrete Fourier transform of the sound wave propagation characteristics to obtain a wave number spectrum representing sound wave propagation characteristics for each horizontal wave number component; and covariance between each wave number data adjacent to each wave number data of the wave number spectrum. Matrix Covariance matrix generation processing, eigenvalue decomposition or singular value decomposition is performed on the covariance matrix to obtain a plurality of eigenvectors, the plurality of eigenvectors are divided into a signal vector and a noise vector, and the transmitter and the High-resolution processing for calculating the reciprocal of the inner product of the noise vector and the steering vector of the signal component obtained from the horizontal distance between the receiver and the wave number of the received signal, and the reciprocal of the inner product has a peak. And a peak detection process in which the wave number having the peak is set as the normal mode eigenvalue.

According to the first aspect, since the mode eigenvalue measuring method is configured as described above, the input signal of the high resolution processing is obtained from the wave number spectrum limited by the narrow wave number band including the wave number component of the mode wave. This is the generated covariance matrix. According to a second aspect, in the first aspect, the discrete Fourier transform processing is performed after dividing the sound wave propagation characteristic obtained by the sound wave propagation characteristic calculation processing into a plurality of distance sections, and the covariance matrix generation processing is After generating a covariance matrix for each distance section generated from each wavenumber spectrum obtained by Fourier transform, the covariance matrix is calculated by performing addition. According to the second aspect, a covariance matrix divided into a plurality of distance sections is generated and then added.

In the third invention, a transmitter and a receiver, one of which is fixed and the other of which is linearly moved at a constant speed, are provided in seawater,
Moving time for a plurality of frequency spectra obtained by performing a discrete Fourier transform on a received signal generated by receiving a sound wave of a constant frequency output from the transmitter by the receiver. A sound wave propagation characteristic calculation process for obtaining sound wave propagation characteristics when the transmitter and the receiver are equidistantly spaced by performing phase compensation for the same, and performing Hankel transform on the sound wave propagation characteristics By limiting the wavenumber band to the band containing the mode wave component and performing the inverse Hankel transform,
A band limiting process for obtaining a band-limited sound propagation characteristic,
A covariance matrix generation process for obtaining a covariance matrix between the data and the high resolution process and the peak detection process of the first invention are performed on the band-limited sound propagation characteristic data. According to the third aspect, the input signal of the high-resolution processing is a covariance matrix generated from complex amplitude data limited in a narrow wavenumber band including a wavenumber component of a mode wave.

In a fourth aspect based on the third aspect, the covariance matrix generation processing is performed after dividing the band-limited sound wave propagation characteristic obtained by the band limitation processing into a plurality of distance sections, and thereafter, accumulating. To calculate the covariance matrix. According to the fourth aspect, a covariance matrix divided into a plurality of distance sections is generated and then added. Fifth
In the third invention, in the third invention, the sound wave propagation characteristic calculation processing includes dividing the received signal sequence into a plurality of distance sections, and then performing discrete Fourier transform to obtain sound wave propagation characteristics for each of the distance sections. The processing is performed on the sound wave propagation characteristics for each of the distance sections, and the covariance matrix generation processing includes:
The covariance matrix is calculated by accumulating the covariance matrix for each distance section. According to the fifth aspect, the discrete Fourier transform of the received signal is performed by dividing into a plurality of distance sections, and a covariance matrix divided into a plurality of distance sections is finally generated and added. Therefore, the above problem can be solved.

[0013]

FIG. 2 is a diagram for explaining the principle of the present invention. In this figure, an example is shown in which modes 1 and 2 cannot be separated by Hankel transform. Since high-level noise exists outside the range of wave numbers s indicated by hatching in the figure, when MUSIC is applied, this noise is regarded as a mode wave of the sound source signal, and both or one of the modes 1 and 2 is generated. It can be hidden. Therefore, the following (a),
By the method of (b) or the like, the effect of a large noise component outside the wave number range s is removed, and the performance of MUSIC is brought out. (A) Apply MUSIC to the covariance matrix of the wavenumber spectrum from which the hatched portion (wavenumber range s) is extracted. In this case, a steering vector obtained by performing a Fourier transform on the steering vector e in Expression (4) is used. (B) MUSIC is applied to the covariance matrix of the signal (corresponding to the sound wave propagation characteristic of the sound field in which the wave number band is limited) subjected to the inverse Hankel transform after extracting the shaded portion (wave number range s). In this case, the same steering vector as the steering vector e in Equation (4) is used.

First Embodiment FIG. 1 is a block diagram showing a configuration of a mode eigenvalue measuring apparatus for implementing a mode eigenvalue measuring method according to a first embodiment of the present invention. The mode eigenvalue measuring device includes a propagation characteristic measuring device 1, a first discrete Fourier transform processing unit 2, a synthetic aperture processing unit 3, a second discrete Fourier transform processing unit 4,
It includes a covariance matrix generation unit 5, a high resolution processing unit 6, and a peak detection processing unit 7. The propagation characteristic measuring device 1 includes a receiver fixed in water and a sound source (transmitter) moving in water.
And has a function of generating the received signal S1. The first discrete Fourier transform processing unit 2 receives the received signal S1
Are sequentially input, a discrete Fourier transform is performed, and a discrete Fourier transform output signal S2 is output. The synthetic aperture processing unit 3 has a function of generating complex amplitude data S3 at an equal distance at the same time from the discrete Fourier transform output signal S2. The second discrete Fourier transform processing unit 4 has a function of inputting the complex amplitude data S3 and outputting a wave number spectrum S4. The covariance matrix generation unit 5 has a function of generating a covariance matrix S5 between each wave number data adjacent to each wave number data of the wave number spectrum S4. The high-resolution processing unit 6 performs MUSI on the covariance matrix S5.
C and has a function of generating a processing result S6.
The peak detection processing section 7 has a function of inputting the processing result S6 and outputting a wave number (mode specific value) S7 having a peak. Next, each process (A) to (E) of the mode eigenvalue measuring method of the present embodiment will be described with reference to FIG.

(A) Sound Wave Propagation Characteristic Calculation Process In the propagation characteristic measuring apparatus 1, a sound wave having a constant frequency output from a transmitter is received by a receiver to generate a received signal S1. The received signal S1 is sent to the first discrete Fourier transform processing unit 2. The first discrete Fourier transform processing unit 2, wave propagation characteristics P of each frequency component by performing discrete Fourier transform on the received signals S1, it generates a ^{_{(r, z r, z s}} ). This sound wave propagation characteristic is corrected by the Doppler shift, and sent to the synthetic aperture processing unit 3 as a sound wave propagation characteristic S2 corresponding to a frequency equal to the sound source frequency.
The synthetic aperture processing unit 3 corrects the phase so that the sound wave propagation characteristic S2 becomes the propagation characteristic of the received signal at the same time, and the complex amplitude data S3 in which the distance between the sound source and the receiver becomes an equal distance by interpolation. _{S3 = P (r, z r} , z s) to produce a. The complex amplitude data S3 is sent to the second discrete Fourier transform processing unit 4.

(B) Discrete Fourier Transform Processing The second discrete Fourier transform processing unit 4 generates the complex amplitude data S
Wavenumber spectrum S4 S4 performs discrete Fourier transform to _{3 = G (k r, z} r, z s) and outputs a. The wave number spectrum S4 is sent to the covariance matrix generation unit 5. (C) Covariance Matrix Generation Processing The covariance matrix generation unit 5 extracts a band component including a mode wave component from the wavenumber spectrum S4, and performs a sound propagation characteristic y (n) y (n) = G (k _rn , Z _r , z _s ) to calculate a covariance matrix S5S5 = R _yy shown in Expression (3). The covariance matrix S5 is sent to the high-resolution processing unit 6.

(D) High Resolution Processing The high resolution processing unit 6 performs MUSIC on the covariance matrix S5. That is, eigenvalue decomposition or singular value decomposition is performed on the covariance matrix S5 to generate a noise matrix E, and a steering vector in a wave number space that is a vector of signal components (that is, a steering vector e shown in Expression (4) is Fourier-transformed) Is calculated, and the reciprocal of the correlation value between them (e ^H EE)
^He e) ^-1 is calculated to generate a processing result S6. The processing result S6 is sent to the peak detection processing section 7. (E) Peak detection processing The peak detection processing unit 7 detects the peak of the processing result S6,
This is output as the mode unique value S7. As described above, in the first embodiment, a covariance matrix generated from a wavenumber spectrum limited in a narrow wavenumber band including a wavenumber component of a mode wave is used as an input signal of the high-resolution processing unit 6 for performing MUSIC. As a result, the effects of noise outside the wavenumber band can be eliminated, and the performance of MUSIC can be improved.

Second Embodiment FIG. 3 is a block diagram showing a configuration of a mode eigenvalue measuring apparatus for implementing a mode eigenvalue measuring method according to a second embodiment of the present invention. Are denoted by the same reference numerals. In this mode eigenvalue measurement device, FIG.
, A distance section dividing unit 8 and an adder 9 are added to the mode eigenvalue measuring device, and N discrete Fourier transform processing units 4-1 to 4-N instead of the second discrete Fourier transform processing unit 4, and covariance N covariance matrix generators 5 instead of matrix generator 5
-1 to 5-N are provided. The distance section dividing unit 8
It has a function of dividing the complex amplitude data S3 into N distance sections to obtain complex amplitude data S8-1 to S8-N at equidistant intervals. Discrete Fourier transform processing units 4-1 to 4-N
Input the complex amplitude data S8-1 to S8-N at equidistant intervals and input wavenumber spectra S4-1 to S4-N.
Are output. The covariance matrix generation units 5-1 to 5-N generate wave number spectra S4-1 to S4.
It has a function of generating covariance matrices S5-1 to S5-N between adjacent wavenumber data for each -N wavenumber data. The adder 9 has N covariance matrices S5-1.
To S5-N. Others are shown in Figure 1
This is the same configuration as.

The mode eigenvalue measuring method in the mode eigenvalue measuring apparatus differs from FIG. 1 in the following points. That is, in the discrete Fourier transform processing, the distance section dividing unit 8
Divides the complex amplitude data S3 into N sets of distance sections and sends them to the discrete Fourier transform processing units 4-1 to 4-N. In each of the discrete Fourier transform processing units 4-1 to 4-N, the same processing as that of the discrete Fourier transform processing unit 4 of the first embodiment is performed. In the covariance matrix generation processing, each of the covariance matrix generation units 5-1 to 5-N performs the same processing as that of the covariance matrix generation unit 5 of the first embodiment for each distance section. Are obtained and sent to the adder 9. The adder 9 removes uncorrelated noise components by adding the N covariance matrices S5-1 to S5-N, and sends the result to the high-resolution processing unit 6 as an addition signal S9. Hereinafter, the same processing as in the first embodiment is performed.
As described above, in the second embodiment, in addition to the advantages of the first embodiment, the covariance matrices S5-1 to S5-N divided into N distance sections are generated and then added. Therefore, uncorrelated noise components are removed, and processing that is not affected by random noise can be expected.

Third Embodiment FIG. 4 is a block diagram showing a configuration of a mode eigenvalue measuring apparatus for implementing a mode eigenvalue measuring method according to a third embodiment of the present invention. Are denoted by the same reference numerals. In this mode eigenvalue measurement device, FIG.
A wavenumber bandpass filter 10 is provided instead of the discrete Fourier transform processing unit 4 in the middle. The wavenumber bandpass filter 10 has a function of receiving the complex amplitude data S3 and generating equidistantly spaced complex amplitude data S10 in a band-limited wavenumber band. Other configurations are the same as those in FIG. Next, each processing (A) to (E) of the mode eigenvalue measuring method of the present embodiment will be described with reference to FIG.

(A) Sound Wave Propagation Characteristics Calculation Processing This is the same as in the first embodiment. (B) band limitation processing frequency band filter 10 performs Hankel transform complex amplitude data S3 S3 = P (r, z r, z s) with respect to, limit the frequency band to band including the components of the mode wave by performing an inverse Hankel transform, the band-limited complex amplitude data S10 S10 = P (r, z r, z s) to produce a. The band-limited complex amplitude data S10 is
It is sent to the covariance matrix generator 5. (C) the covariance matrix generation process covariance matrix generation unit 5, by using the complex amplitude data S10 which has been band-limited, wave propagation characteristic y (n) y (n) = P (r n, z r, z s ) And calculate the covariance matrix S5 S5 = R _yy shown in equation (3). The covariance matrix S5 is sent to the high-resolution processing unit 6.

(D) High-resolution processing The high-resolution processing unit 6 performs MUSIC on the covariance matrix S5. That is, the noise matrix E is generated by performing the eigenvalue decomposition or the singular value decomposition of the covariance matrix S5, and the steering vector in the metric space that is the vector of the signal component (that is, the steering vector e shown in Expression (4)) is calculated. and, the reciprocal (e ^H EE ^H e) ^-1 of the correlation value of the two is calculated, to generate a processing result S6. The processing result S6 is sent to the peak detection processing section 7. (E) Peak detection processing The peak detection processing unit 7 detects the peak of the processing result S6,
This is output as the mode unique value S7. As described above, in the third embodiment, the covariance generated from the complex amplitude data S10 limited by the narrow wavenumber band including the wavenumber component of the mode wave is used as the input signal of the high-resolution processing unit 6 that performs MUSIC. Since the matrix S5 is used, the influence of noise outside the wavenumber band can be eliminated, and the performance of the MUSIC can be improved.

Fourth Embodiment FIG. 5 is a block diagram showing a configuration of a mode eigenvalue measuring device for implementing a mode eigenvalue measuring method according to a fourth embodiment of the present invention. Common elements are denoted by common reference numerals. In this mode eigenvalue measuring device, a distance section dividing unit 8 and an adder 9 are added to the mode eigenvalue measuring device of FIG.
The covariance matrix generation units 5-1 to 5-N are provided. The distance section dividing unit 8 divides the complex amplitude data S10 into N distance sections and divides the complex amplitude data
-1 to S8-N. The covariance matrix generators 5-1 to 5-N generate complex amplitude data S8-1 to S8.
It has a function of generating the covariance matrices S5-1 to S5-N for -N. The adder 9 has a function of adding N covariance matrices S5-1 to S5-N. Other configurations are the same as those in FIG.

The mode eigenvalue measuring method in this mode eigenvalue measuring apparatus differs from FIG. 4 in the following point. That is, the distance section dividing unit 8, the complex amplitude data S10 S10 = P (r, z r, z s) and is divided into N distance intervals, the respective covariance matrix generation unit 5-1 to 5-N Send out. Covariance matrix generator 5-1
.About.5-N calculate the covariance matrices S5-1 to S5-N shown in equation (3) for each distance section and send them to the adder 9. The adder 9 accumulates the covariance matrices S5-1 to S5-N and sends it to the high resolution processing unit 6 as an addition signal S9.
Hereinafter, it is similar to the third embodiment. As described above, in the fourth embodiment, in addition to the advantages of the third embodiment,
Covariance matrices S5-1 to S5- divided into N distance sections
Since N is generated and then added, uncorrelated noise components are removed, and processing not affected by random noise can be performed.

Fifth Embodiment FIG. 6 is a block diagram showing a configuration of a mode eigenvalue measuring device for implementing a mode eigenvalue measuring method according to a fifth embodiment of the present invention. Common elements are denoted by common reference numerals. In this mode eigenvalue measuring device, the propagation characteristic measuring device 1 is connected to the distance section dividing unit 8. The distance section dividing unit 8 has a function of dividing the received signal S1 into N distance sections to obtain received signals S8-1 to S8-N. The output side of the distance section dividing unit 8 is connected to the discrete Fourier transform processing units 2-1 to 2-N, respectively. The discrete Fourier transform processing units 2-1 to 2-N sequentially input the divided received signals S8-1 to S8-N, perform discrete Fourier transform, and perform discrete Fourier transform signal S
It has a function of generating 2-1 to S2-N. Output sides of the discrete Fourier transform processing units 2-1 to 2-N are connected to the synthetic aperture processing units 3-1 to 3-N, respectively.

The synthetic aperture processing units 3-1 to 3-N output complex amplitude data S3-1 to S3-N at the same time and equidistant intervals from the discrete Fourier transform signals S2-1 to S2-N, respectively. have. Synthetic aperture processing units 3-1 to 3
-N output side is a wave number bandpass filter 10-1 to 10-N
Connected to each other. Wavenumber bandpass filter 10-1
10 to 10-N have a function of receiving the complex amplitude data S3-1 to S3-N and generating the complex amplitude data S10-1 to S10-N at equidistant intervals in a certain wavenumber band. Output sides of the wavenumber bandpass filters 10-1 to 10-N are connected to the covariance matrix generation units 5-1 to 5-N, respectively. After the covariance matrix generation units 5-1 to 5-N,
This is a configuration similar to that of FIG. The mode eigenvalue measuring method in this mode eigenvalue measuring device differs from the fourth embodiment in the following points.

That is, the distance section dividing unit 8 receives the received signal S1
Is divided into N distance sections, and a discrete Fourier transform processing unit 2
-1 to 2-N. Discrete Fourier transform processing unit 2-
1-2-N, synthetic aperture processing units 3-1-3-N, wavenumber bandpass filters 10-1-10-N, and covariance matrix generation unit 5
−1 to 5-N are the discrete Fourier transform processing unit 2, the synthetic aperture processing unit 3, the covariance matrix generation unit 5 in the third embodiment.
Are performed, and N covariance matrices S5-
1 to S5-N are obtained and sent to the adder 9. After the adder 9, the same processing as in the fourth embodiment is performed. As described above, in the fifth embodiment, the discrete Fourier transform of the received signal S1 is performed by dividing it into certain distance sections, and finally, the covariance matrices S5-1 to S5-N are generated and added. As a result, uncorrelated noise components are removed, and processing free from the influence of random noise can be performed.

[0028]

As described above in detail, according to the first aspect, the covariance generated from the wavenumber spectrum limited in a narrow wavenumber band including the wavenumber component of the mode wave as an input signal for high resolution processing. Since I used a matrix,
The effect of noise outside the wavenumber band can be eliminated, and the performance of MUSIC can be improved. According to the second aspect, in addition to the effect of the first aspect, since a covariance matrix divided into a plurality of distance sections is generated and then added, uncorrelated noise components are removed, and random noise is removed. Process that is not affected by According to the third aspect, a covariance matrix generated from complex amplitude data limited in a narrow wavenumber band including a wavenumber component of a mode wave is used as an input signal for high-resolution processing. The effects of noise can be eliminated, and the performance of MUSIC can be improved.

According to the fourth aspect, in addition to the effect of the third aspect, since a covariance matrix divided into a plurality of distance sections is generated and then added, uncorrelated noise components are removed. Processing not affected by random noise can be performed.
According to the fifth aspect, the discrete Fourier transform of the received signal is performed by dividing it into a plurality of distance sections, and a covariance matrix is finally generated and added. Therefore, uncorrelated noise components are removed. In addition, processing that is not affected by random noise can be performed.

[Brief description of the drawings]

FIG. 1 is a configuration block diagram of a mode eigenvalue measurement device according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating the principle of the present invention.

FIG. 3 is a configuration block diagram of a mode eigenvalue measurement device according to a second embodiment of the present invention.

FIG. 4 is a configuration block diagram of a mode eigenvalue measurement device according to a third embodiment of the present invention.

FIG. 5 is a configuration block diagram of a mode eigenvalue measurement device according to a fourth embodiment of the present invention.

FIG. 6 is a configuration block diagram of a mode eigenvalue measuring device according to a fifth embodiment of the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 Propagation characteristic measuring apparatus 2 1st discrete Fourier transform processing part 3,3-1-3-N Synthetic aperture processing part 4,4-1-4-N 2nd discrete Fourier transform processing part 5,5-1-1 5-N covariance matrix generation unit 6 high-resolution processing unit 7 peak detection processing unit 8 distance section division unit 9 adder 10, 10-1 to 10-N wavenumber bandpass filter

──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Tsuneo Ishiwatari 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (72) Inventor Toru Morishita 1-7-112 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (72) Inventor Shunji Ozaki 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (56) References JP-A-4-744985 (JP, A) JP JP-A-9-33652 (JP, A) JP-A-11-23700 (JP, A) JP-A-9-274081 (JP, A) (58) Fields investigated (Int. Cl. ⁷ , DB name) G01H 3 / 00 G01H 5/00 G01S 15/89

Claims (5)

(57) [Claims] 1. A transmitter and a receiver, one of which is fixed and the other of which performs a constant-velocity linear motion, are provided in seawater, and a sound wave of a fixed frequency output from the transmitter is received by the receiver. The phase difference of the moving time is applied to the frequency spectrum of a plurality of frequencies obtained by performing the discrete Fourier transform on the received signal generated by receiving the A sound wave propagation characteristic calculating process for obtaining a sound wave propagation characteristic when the intervals are equal distance intervals; a discrete Fourier transform process for obtaining a wave number spectrum representing the sound wave propagation characteristic for each horizontal wave number component by performing a discrete Fourier transform on the sound wave propagation characteristic; A covariance matrix generation process for obtaining a covariance matrix between each of the wavenumber data adjacent to each of the wavenumber data of the wavenumber spectrum; and performing eigenvalue decomposition or singular value decomposition on the covariance matrix to obtain a plurality of eigenvectors. The plurality of eigenvectors are divided into a signal vector and a noise vector, and the steering vector of the signal component obtained from the horizontal distance between the transmitter and the receiver and the wave number of the received signal. High resolution processing of calculating the reciprocal of the inner product with the noise vector, and the reciprocal of the inner product finds a wave number having a peak, and a peak detection processing that makes the wave number having the peak a normal mode eigenvalue, Mode eigenvalue measurement method to be used. 2. The discrete Fourier transform process is performed after dividing the sound wave propagation characteristic obtained in the sound wave propagation characteristic calculation process into a plurality of distance sections, and the covariance matrix generation process is performed by the discrete Fourier transform. 2. The mode eigenvalue measurement method according to claim 1, wherein after generating a covariance matrix for each of the distance sections generated from each of the obtained wavenumber spectra, the covariance matrix is calculated by performing addition. 3. A transmitter and a receiver, one of which is fixed and the other of which moves linearly at a constant speed, are provided in seawater, and a sound wave of a constant frequency output from the transmitter is received by the receiver. Received signal generated by receiving is subjected to discrete Fourier transform to obtain a frequency spectrum of a plurality of frequencies, and phase-compensated for a moving time to perform a phase compensation between the transmitter and the receiver. Is a sound wave propagation characteristic calculation processing for obtaining sound wave propagation characteristics in the case of equidistant intervals, performs Hankel transform on the sound wave propagation characteristics, restricts the wavenumber band to a band including a mode wave component, and performs inverse Hankel transformation. Performing a band-limited process for obtaining a band-limited sound propagation characteristic; and a covariance matrix generation process for obtaining a covariance matrix between the data for the band-limited sound propagation characteristic data. High-resolution processing described and A method for measuring a mode eigenvalue, comprising: performing the peak detection processing according to claim 1. 4. The covariance matrix generation processing is performed after dividing the band-limited sound propagation characteristic obtained in the band restriction processing into a plurality of distance sections, and thereafter, the covariance matrix is calculated by performing accumulation. The mode eigenvalue measuring method according to claim 3, wherein the mode eigenvalue is calculated. 5. The sound wave propagation characteristic calculating process, wherein the received signal sequence is divided into a plurality of distance sections, and then the discrete Fourier transform is performed to obtain sound wave propagation characteristics for each of the distance sections. Is performed for each of the sound wave propagation characteristics for each of the distance sections, and the covariance matrix generation process calculates the covariance matrix by performing addition of the covariance matrix for each of the distance sections. Item 3. The mode eigenvalue measurement method according to item 3.