JPH05196717A - Sonar-signal detecting and processing method - Google Patents

Abstract

PURPOSE:To perform high-resolution analysis based on a small number of data by computing a time-frequency spectrum by using a linear prediction method in frequency analysis processing. CONSTITUTION:When sound signals are emitted from a ship and the like, the signals are received with a receiver group 20. The phases of the received signals are changed in a phasing device 30, and the arriving bearings are limited. The output signal is inputted into a signal processing device 40. A correction operating part 41 in the processing device 40 receives the output signal from the phasing device 30, performs correlation operation and outputs the correlation signal into a low-pass filter (LPF) 42. The LPF 42 obtains the correlation coefficient, wherein suitable weighting coefficient is applied based on the correlation coefficient and outputs the correlation coefficient into a prediction-coefficient operating part 43. The operating part 43 obtains the prediction coefficient by using the correlation coefficient and sends the prediction coefficient into a frequency operating part 44. The operating part 44 computes the frequency spectrum based on the coefficient and sends the spectrum into a display device 50. The high-resolution analysis is performed based on the few data by using the linear prediction method in the high-resolution frequency analysis.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a sonar signal detection processing system for finely and magnifying frequency analysis of a sonar output signal to extract and display a spectrum of a running sound of a ship.

[0002]

2. Description of the Related Art Conventionally, as a technique in such a field,
For example, some documents were described in the following documents. Reference: Oki Electric Research and Development, 53 [4] (Sho 61-10), Niijima and Igarashi, "Digital Signal Processing Technology in Underwater Acoustics," p. The 53-58 sonar device corresponds to a radar device in the air, and performs search, position measurement, and classification of a target moving in a three-dimensional space (underwater) using underwater acoustics. The sonar device is classified into a passive sonar device and an active sonar device according to the operation mode. The passive sonar device uses sound waves emitted by the target, and the active sonar device applies sound waves toward the target, and uses the reflected waves (echo) to search for the target, measure the position, and classify the target.

The signal detection processing used in the sonar device includes the temporal processing used to detect the temporal characteristics (waveform, spectrum, etc.) of the signal and the spatial characteristics (position, shape, moving speed, etc.) of the signal. It can be divided into the spatial processes used for detection. Temporal processing includes matched filters, Wiener filters, and spectrum estimation. Spectral estimation is for estimating the intensity of a signal with respect to frequency, and is used for detecting a periodic signal (line spectrum) buried in noise and classifying a target that emits the signal. On the other hand, the beam forming (B
F), and delay / phase estimation. The BF uses the output signals of a large number of wave receivers that form a wave receiver group, and utilizes the directionality of waves propagating in space to improve the S / N of the signal, the incident direction of the signal, and the intensity (space (Spectrum) estimation. Delay / phase estimation is a problem of estimating a delay or a phase difference occurring between signals received by a small number of receivers, and is mainly used for measuring a position of a target.

Of the passive sonar device and the active sonar device, for example, the passive sonar device is mainly used for detecting a running sound and a transient sound. The running sound is a sound generated when the ship runs, and is composed of a continuous random sound component and a pure tone line spectrum component. Transient and random tones are detected by wideband analysis and line spectral components are detected by narrowband analysis. FIG. 2 shows a configuration example of this type of passive sonar device.

FIG. 2 is a schematic block diagram showing a conventional passive sonar device. This passive sonar apparatus has a wave receiver group 10 including a plurality of wave receivers that receive sound waves emitted by the ship 1, and a sonar receiver 11 is connected to the output side thereof. In the sonar receiver 11, the received signals from the receiver group 10 are subjected to phasing processing, a directional beam is formed on the entire circumference to receive underwater sound from the ship 1,
The azimuth display 12a and the frequency display 12b are performed by detecting the azimuth of the ship 1 around the reception point and the characteristic (for example, frequency) of the signal of the ship 1.

In the reception processing of the sonar receiver 11,
In order to detect the characteristics of the signal of the ship 1, a frequency analysis in the time domain is performed by the fast Fourier transform method (FFT method), or a high resolution frequency analysis capable of detecting a fine frequency structure by increasing the number of data is performed. The frequency display 12b is displayed on the two-dimensional plane of frequency-time on the display at a concentration according to the intensity of the spectrum. further,
The sonar receiver 11 also has a function of performing ultra-narrow band analysis capable of detecting a fine frequency structure by narrowing the analysis frequency.

[0007]

However, the conventional sonar signal detection processing system has the following problems. 3A to 3D are diagrams showing an example of the display result of the frequency analysis, in which the horizontal axis represents frequency and the vertical axis represents time. When the actual input signal is roughly classified, FIG.
As shown in (A), the frequency does not fluctuate with time, as shown in (B), the frequency changes periodically with time, and as shown in (C), there is no periodicity and time changes. On the other hand, it is divided into a signal whose frequency largely changes and a signal which appears transiently discontinuously or irregularly as shown in FIG.
Output results of frequency analysis using the conventional FFT method for such a signal are shown in FIGS. This Figure 4
In (a) to (d), the horizontal axis represents frequency and the vertical axis represents time.

In the conventional sonar signal detection processing method, in the case of the signal of FIG.
Good processing results as shown in FIG. 4A are obtained. However, the signals in FIGS. 3B and 3C are the same as those in FIGS. 4B and 4C.
As shown in FIG. 4, the spectrum characteristics of the original signal cannot be detected, and the spectrum characteristics of the original signal cannot be detected, and the spectrum characteristics of the signal of FIG. 3D cannot be detected at all as shown in FIG. 4D. This is because the frequency resolution of the FFT method depends on the analysis time (number of data).

FIGS. 5 (a) to 5 (c) are diagrams showing examples of spectrum analysis by Fourier transform. FIG. 5 (a) shows 120 data, FIG. 5 (b) shows 60 data, and FIG. FIG. 6C shows the power with respect to the frequency when the number of data is 30. As is clear from this figure,
When the number of data is reduced, the frequency spectrum tends to be gentle.

The dependence of the frequency resolution on the number of data in the FFT method will be further considered. For example, consider a case where the 2048-point FFT method is used and the basic sampling frequency of data is 2048 Hz. For analysis of the frequency band up to 1024 Hz, one second of data is enough. Therefore, when performing high-resolution analysis up to 64 Hz, it is necessary to use data for 16 seconds. That is, in order to perform high resolution analysis, the data length must be increased accordingly. Therefore, the output becomes an averaged value within the analysis time, and the frequency fluctuation in a short time cannot be captured.

In order to prevent this, the analysis data is overlapped in the conventional method. However, since the dependence of the frequency resolution on the number of data, which is the principle drawback of the FFT method, has not been eliminated, for example, in passive sonar signal detection, it is unclear for a signal whose frequency changes with time or a transient signal. It will be a detection result.

The present invention solves the problem of the prior art by solving the problem that it is difficult to obtain a high-resolution analysis output with short-time data in the extraction of frequency characteristics such as ship running sound. It provides a processing method.

[0013]

[Means for Solving the Problems] FIGS.
It is a figure which shows the example of spectrum analysis using the linear prediction method which is a frequency analysis system completely different from the FFT method. The figure (a) has 120 data and the figure (b) has 6 data.
0 and (c) of the figure are waveform diagrams of power with respect to frequency when the number of data is 30. It is well known empirically that the adjacent sample values of the sampled analog signal x _n are not independent of each other and have a strong correlation with the linear prediction method. This is a method of predicting the current sample value from some past sample values. As shown in FIGS. 6A to 6C, there is almost no change in the frequency spectrum even if the number of data (sample values) is changed. Therefore, this linear prediction method is a method in which the frequency resolution does not depend on the number of data, and it is possible to perform high-resolution frequency analysis with a small number of data by using this linear prediction method.

Based on such a principle, in the first invention, the acoustic signal is received by the wave receiver group consisting of a plurality of wave receivers, the frequency spectrum is calculated from the received signal by frequency analysis processing, and the frequency is calculated. In the sonar signal detection processing method for displaying the intensity of the spectrum as an image, the frequency analysis processing uses a linear prediction method to calculate the time-frequency spectrum.

Further, in the second invention, in the frequency analysis processing of the first invention, the spatial frequency spectrum is calculated using a linear prediction method, and the arrival direction of the acoustic signal is estimated from the spatial frequency spectrum. ing.

[0016]

According to the first aspect of the invention, since the sonar signal detection processing method is configured as described above, the linear prediction method used for frequency analysis processing does not depend on the number of data, and therefore the number of data is high. The frequency spectrum can be analyzed with high resolution. Therefore, in the extraction of time-frequency characteristics such as ship running sound, if frequency estimation is performed using the linear prediction method, high-resolution analysis output can be obtained with short-time data.

According to the second invention, by calculating the spatial frequency spectrum using the linear prediction method, a high-resolution analysis output can be obtained with short-time data. By estimating the arrival direction of the acoustic signal from the spatial frequency spectrum, the arrival direction of the acoustic signal can be estimated with high resolution regardless of the number of receivers. Therefore, the above problem can be solved.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS First Embodiment FIG. 1 is a functional block diagram of a passive sonar apparatus showing a sonar signal detection processing system according to a first embodiment of the present invention. This passive sonar device is configured by, for example, an individual circuit using a semiconductor integrated circuit or the like, or a program control using a processor, and the like, from a plurality of wave receivers that convert an acoustic signal from a ship or the like into an electric signal. The wave receiver group 20 is provided, and the phase adjuster 30 is connected to the output side thereof. The phase shifter 30 has a function of limiting the arrival direction of the received signals by changing the phase of the received signals of the wave receiver group 20, and the signal processor 40 is connected to the output side thereof. The signal processor 40 has a function of performing frequency analysis using a linear prediction method on the output of the phase adjuster 30, and at the output side of the signal processor 40, a display 50 for displaying the intensity of the frequency spectrum as the shade of color is provided. It is connected.

In the signal processor 40, a frequency analysis unit 4 for performing frequency analysis by the linear prediction method, which is a feature of this embodiment.
0A is provided. The frequency analysis unit 40A performs a correlation operation on the output signal (sample value) x _n of the phase adjuster 30 to obtain a correlation signal rik (n) (i = 1, ..., M, k = 1, ..., m).
The correlation coefficient calculation unit 41 outputs a prediction coefficient via a low-pass filter 42 that outputs a correlation coefficient Rik (n) by performing a convolution calculation on the correlation signal rik (n) on the output side. The calculation unit 43 is connected. The prediction coefficient calculation unit 43 has a function of calculating the correlation coefficient Rik (n) and outputting the prediction coefficient ai (n), and the output side thereof calculates the prediction coefficient ai (n) to calculate the frequency. Spectrum F
A frequency calculation unit 44 that outputs (f, n) (where f: frequency) to the display device 50 is connected.

Next, a sonar signal detection processing method using the passive sonar device of FIG. 1 will be described. When an acoustic signal is radiated from a ship or the like, the acoustic signal is received by the receiver group 2
0, the received signal is received by the phase rectifier 30 by changing the phase of the received signal, the arrival direction of the received signal is limited, and the limited output signal x _n is sent to the signal processor 40. Be done.

In the correlation calculation unit 41 in the signal processor 40,
The output signal x _n of the phase adjuster 30 is input, correlation calculation based on the following equation (1) is performed, and the correlation signal rik (n) is output to the low-pass filter 42. rik (n) = x (n−i) · x (n−k) (1) where i = 1, ..., m, k = 1, ..., m n−i; i sample before (past ) Sample value n−k; sample value before k samples (past) In the low-pass filter 42, an appropriate weighting factor Wt is convoluted according to the following equation (2) based on the correlation coefficient rik (n), The number Rik (n) is calculated and given to the prediction coefficient calculator 43. The prediction coefficient calculation unit 43 uses the correlation coefficient Rik (n),
By solving the normal equation of the following equation (3), the prediction coefficient a
i (n) (i = 1, ..., M) is obtained and given to the frequency calculation unit 44.

[0022]

[Equation 1] The frequency calculation unit 44 calculates the frequency spectrum F (f, n) according to the following equation (4) based on the prediction coefficient ai (n), and sends the frequency spectrum F (f, n) to the display unit 50. However, i = 1, ..., M f; Frequency fs; Sampling frequency In the display unit 50, the frequency spectrum F (f, n) is input, and the intensity of the frequency spectrum F (f, n) is displayed as the shade of color. To do.

In the first embodiment, the frequency analysis unit 40
In A, since the frequency analysis is performed using the linear prediction method, the frequency resolution does not depend on the number of data unlike the conventional FFT method. Therefore, it is not necessary to increase the number of data even in the case of high resolution frequency analysis, and the time can be shortened. Even a signal that fluctuates at can be analyzed without losing its characteristics. Therefore, FIG.
As shown in (D) to (D), it is possible to analyze the passive sonar output signal with high precision by fine and expanded frequency analysis, and extract and display the spectrum of the navigation sound of the ship.

The second embodiment in sonar signal detection processing system of the second embodiment, by providing a spatial element (installation position of the receivers) to the output signal x _n of the phasing device 30 of FIG. 1 , As in the first embodiment,
The spatial frequency spectrum can be obtained based on the equations (1) to (4). Then, in the analysis result of the spatial frequency spectrum, as shown in FIG. 6, if the horizontal axis indicates the azimuth instead of the frequency and the vertical axis indicates the power, the peak value of the power indicates the arrival direction of the acoustic signal. It will be. Therefore, if the frequency analysis unit 40A is provided with a function of detecting the peak value of the power, the arrival direction of the acoustic signal can be estimated, and the estimation result can be displayed on the display unit 50.

Thus, in this embodiment, the frequency analysis unit 4
By obtaining the spatial frequency spectrum using the linear prediction method with 0A, the arrival direction of the acoustic signal can be estimated with high resolution regardless of the number of receivers.

The present invention is not limited to the above embodiment,
If the passive sonar device of FIG. 1 is provided with a function of emitting a sound wave toward a target and receiving a reflected wave thereof, the above embodiment can be applied to a sonar signal detection processing method in an active sonar device. Almost the same action and effect can be obtained. Further, the display 50 of FIG. 1 displays the strength and weakness of the frequency spectrum as light and shade of color, but the strength and weakness may be displayed in different colors.

[0027]

As described in detail above, according to the first aspect of the invention, since the frequency analysis processing is performed using the linear prediction method, the frequency resolution is reduced to the number of data as in the conventional FFT method. Since it does not depend on it, it is not necessary to increase the number of data even at the time of high-resolution frequency analysis, and a time-frequency spectrum can be analyzed even for a signal that fluctuates in a short time without impairing its characteristics. Therefore, it becomes possible to analyze the sonar output signal with high precision in a fine and expanded frequency, and extract and display the spectrum of the navigation sound of the ship.

According to the second invention, in the frequency analysis processing, the spatial frequency spectrum is calculated by using the linear prediction method, so that the arrival direction of the acoustic signal can be detected with high resolution regardless of the number of receivers. It can be estimated and the applicable range of the sonar signal detection processing method can be expanded.

[Brief description of drawings]

FIG. 1 is a functional block diagram of a passive sonar device showing a first embodiment of the present invention.

FIG. 2 is a schematic configuration block diagram of a conventional passive sonar device.

FIG. 3 is a diagram showing a frequency analysis result.

FIG. 4 is a diagram showing a frequency analysis result using the FFT method.

FIG. 5 is a diagram showing an example of spectrum analysis by Fourier transform.

FIG. 6 is a diagram showing an example of spectrum analysis by a linear prediction method.

[Explanation of symbols]

 20 receiver group 40 signal processor 40A frequency analysis unit 41 correlation calculation unit 42 low-pass filter 43 prediction coefficient calculation unit 44 frequency calculation unit 50 indicator

Front page continued (72) Inventor Kiyohito Tokuda 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. (72) Inventor Takuro Sato 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. Company (72) Inventor Yuichi Shiraki 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd.

Claims (2)

[Claims] 1. A sonar signal detection process for receiving an acoustic signal by a wave receiver group including a plurality of wave receivers, calculating a frequency spectrum from the received signal by frequency analysis processing, and displaying the intensity of the frequency spectrum as an image. In the method, in the frequency analysis processing, a sonar signal detection processing method is characterized in that a time-frequency spectrum is calculated using a linear prediction method. 2. The sonar signal detection processing method according to claim 1, wherein in the frequency analysis processing, a spatial frequency spectrum is calculated using a linear prediction method, and the arrival direction of the acoustic signal is estimated from the spatial frequency spectrum. A sonar signal detection processing method characterized by the above.