RU2368919C1 - Method of identifying targets from sonar signals - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: invention relates to sonar and can be used for identifying targets with irregular geometrical shape. Method of identifying targets from sonar signals is based on analysing sound pulses with high resolution on range, reception of signals reflected by a target, matched filtering of composite signals, determination of location angle of the target with error of not more than 15°. Standard values of time positions of essential maximums in reflected signals for different targets and a defined range of location angle, corresponding to the value of measured location angle, are selected from memory. Probe signals are generated in form of a sum of probe pulses with standard delay times. Delays of probe signals relative the first pulse in the generated signal correspond to time positions of essential maximums relative the last maximum in the reflected signal. The generated signals are emitted in the direction of the target, reflected signals are received, envelopes of these signals are detected, average level of each envelope of reflected signal is recorded and the target is identified, corresponding to the probe composite signal, giving rise to a reflected signal the maximum of the envelope of which exceeds the average level by value of a given threshold.

EFFECT: increased noise-immunity for target identification.

2 cl, 2 dwg

Description

The invention relates to sonar and can be used to identify real objects of complex geometric shapes (for example, marine underwater objects or sections of the bottom).

Known methods of identification (or classification) by sonar and radar signals (see RF patents No. 2176401, No. 2161319, No. 2000892, MKI G01S 9/66, US patent No. 3523292, MKI G01S 9/02, NCI 346-5, etc. )

Known methods use radiation from narrowband and broadband probe pulses. The spectra and correlation characteristics of the reflected signals are analyzed. Identification is performed using reference frequency and time characteristics of the reflected signals.

The implementation of the body recognition method based on the relative position of the resonance maxima {f _j } in the spectrum of the echo signal is presented in US patent No. 3716823. In this invention, the low-frequency resonance maximum f _p characterizes the overall size of the recognized body, and high-frequency resonances the thickness of the shells (or, possibly resonant properties of individual structural elements of the body).

Long probing pulses with tonal filling at frequencies f _l , f ₂ , ..., f _n are generated using generators and fed to a power amplifier. In the power amplifier, the signals from the generators are amplified and fed through the receive-receive switch to the hydroacoustic antenna.

The antenna must be broadband to provide radiation and reception of signals in the frequency band from f ₁ to f _n . The frequency f _n may exceed the frequency f ₁ tens of times or more. It is quite difficult to use one hydroacoustic antenna to emit these signals (for example, it is possible to use a multi-element antenna consisting of several piezoelectric transducers with different resonant frequencies). As such an antenna, parametric radiators can be used. Apparently, in the present invention, it is proposed to use a hydroacoustic antenna consisting of a set of piezoelectric transducers with different resonant frequencies.

The reflected signal (or several signals at different frequencies) from the recognized body goes to the receiver, in which the amplification of signals in a wide frequency band is performed. From the output of the receiver, the echo under investigation is fed to bandpass filters tuned to separate individual frequency components f ₁ , f ₂ , -, f _n . The frequency bands of the filters are consistent with the bandwidth of the emitted signals. If the recognizable body (object) and (or) the vessel on which the device is located are moving, then the filters should take into account the Doppler frequency shift of the echo signal (this device does not provide for this, apparently, the vessel and the recognizable body are considered stationary or inactive). The signals from the outputs of the filters enter the computer. The computer computes the ratio of the amplitudes of the output signals kS ₁ , kS ₂ , ..., kS _n filters, excluding their dependence on the transfer coefficient of the hydroacoustic channel k. The calculated amplitude ratios kS ₁ , kS ₂ , ..., kS _{n are} compared with their reference values for different bodies entering the computer from the storage device. The computer decides whether the sample kS ₁ , kS ₂ , ..., kS _n belongs to one of the standards (the decision rule is not indicated). The decision signal for each probe package is sent to the registration device, and then to digital indicators (fluorescent indicators, a printing device, or other indicators). Along with recognition, the coordinates of the irradiated body are measured. For this purpose, a probe pulse is formed by a separate generator at a frequency f _ω , amplified in a power amplifier, through a receive-receive switch, is supplied to one of the transducers of a hydroacoustic antenna and is emitted simultaneously with the radiation of the main polychromatic signal in the direction of a recognizable body (additional circuit nodes necessary for measuring body coordinates).

The reflected signal from the antenna output through the transceiver enters a band-pass filter with a central frequency f _a and with a bandwidth corresponding to the spectrum width of the emitted signal. From the output of the filter, the signal enters the channel for determining the coordinates of the irradiated body, where its location is measured using one of the known methods. The measured coordinates are recorded in the registration device. Thus, the registration device simultaneously receives information about the type (class) of the recognized body (object), its range and angular coordinates.

As a prototype, US patent No. 4084148 is adopted, which proposes a method and apparatus for recognizing bodies of complex shape by an echo signal using short probe pulses that provide high resolution recognition device for the range of individual reflecting elements of the body 1.

Short tonal sounding pulses, the spatial length of which cτ _ξ / 2 (where c is the speed of sound, τ _ξ is the duration of the sounding pulse) is much smaller than the overall dimensions of the recognized bodies, are formed using a generator and emitted into the aqueous medium using a radiator.

Signal s (t) reflected from the irradiated body arrives at the receiver. In the direct Fourier transform-FFT block, the spectrum F (ω) of the echo s (t) is determined. The probe pulse from the generator is supplied simultaneously to the emitter and to the direct Fourier-FFT converter, where its spectrum F _ξ (ω) is calculated. In the inverter, the function 1 / F _ξ (ω) is formed, the inverse of the spectrum of the probe pulse ξ (t). The multiplier computes the ratio of the spectrum of the echo signal F (ω) to the spectrum of the probe pulse F _ξ (ω). Thus, the frequency transfer function of the reflecting body H (ω) = F (ω) / F _ξ (ω) is determined. In the inverse Fourier-BOPF transform block, the transition function of the reflecting body h (t) is calculated. The function h (t) is represented as the sum of delta functions:

Where a _i - amplitude coefficients; τ _i - the time delay of the arrival of the individual reflected signals from the body of complex shape.

The calculated transition function characterizes the relative position of the reflecting elements on the body of complex shape using the values of τ _i and the ratio of the reflection levels of the probe pulse from these elements - amplitude coefficients a _i . The transition function h (t) is used directly to recognize bodies of complex shape. The device diagram does not show the block of standards and the block of decision-making. The description of the method and device does not indicate the type of decision rule. The calculation of the transfer function of the reflecting body H (ω) in the form of the ratio of the spectrum of the echo signal F (ω) to the spectrum of the probe pulse F _ξ (ω) is associated with certain difficulties during the movement of the carrier vessel of the present device relative to the recognized body. In this case, the spectrum of the echo signal F (ω) will be shifted and possibly expanded relative to the spectrum of the probe pulse F _ξ (ω). In the considered patent, the formulas for H (ω) and h (t) are not given taking into account the Doppler frequency shift of the echo signal. The disadvantage of this device is the need to perform the Fourier transform three times, which complicates the construction of a system for recognizing bodies of complex geometric shapes.

Apparently, it is advisable to isolate the transition function h (t) directly from the envelope of the echo signal, for example, using the algorithm for extracting significant local maxima (Davydov BC Algorithm for detecting significant maxima in the envelope of the signal. // Abstracts at the All-Union Acoustic Seminar. M. 1988, p. 64). In this case, it is not necessary to additionally take into account the Doppler shift of the spectrum of the echo signal when calculating h (t). To recognize bodies of complex shape, the most informative feature space of the temporal positions of the maxima in the echo signals {τ _j } should be used first, and the feature space of the amplitude values of the maxima {S _j } should be used as an additional one if it is necessary to increase the probability of correct recognition. In this case, it is possible to apply optimal decision rules (Davydov BC The optimal decision rule for recognizing bodies of complex geometric shape by reflected hydroacoustic signals. // Abstracts of the IV Far Eastern Conference. Acoustic methods and means of ocean research. // USSR Academy of Sciences, Vladivostok, 1986, p. 33-36).

This method uses classification features for identification that contain the most information about bodies of complex geometric shape. However, when using probing pulses with a high range resolution, the signal reflection level is lower than when using long probing pulses (without compression) due to the fact that individual echo maximums are formed when the pulses are reflected not from the entire target body, but from its individual reflective elements. For identification, it is necessary to exceed the maxima over the minima, and the minima must be greater than the level of interference, with a signal / noise ratio of less than 18 dB it is difficult to ensure reliable identification of the submarines.

The purpose of this proposal is to increase the noise immunity of identifying targets of complex geometric shapes according to the design of its body at a location angle determined with an error of not more than 15 °.

The goal is achieved in the same way as in the method of target identification by sonar signals, including the emission of probing pulses with high resolution in range, the reception of reflected signals from the target, the determination of the target location angle with an error of not more than 15 °, form the reference values of the temporal positions of the maxima in the reflected signals during training, during identification, the reference values of the temporal positions of significant maxima in reflected signals for various purposes and a certain range are selected from memory probing signals are generated in the form of a sum of probing pulses with reference time delays, and the delays of probing pulses relative to the first pulse in the generated signal correspond to the temporary positions of significant maxima relative to the last maximum in the reflected signal, the generated signals radiate in the direction target, receive the reflected signals, highlight the envelopes of these signals, remember the average level of each the envelope of the reflected signal and identifying the target, the corresponding sum signal probing, the caller reflected signal exceeding its maximum envelope above the mean level by an amount specified by the threshold.

When using complex sounding pulses after radiation of the generated total signals and receiving the reflected signals before the selection of the envelopes of the signals, a coordinated filtering of the reflected signals is performed.

Salient features of the proposed solution from the known technical solutions (see RF patents No. 2176401, No. 2161319, 2000892) and the prototype are:

- the formation of sounding signals in the form of a sum of sounding pulses with reference time delays corresponding to the inverse order reference values of the temporal positions of significant maxima in the reflected signals for different purposes and a certain range of location angles corresponding to the value of the measured target location angle;

- identification of the target corresponding to the emitted total probing signal, which caused the reflected signal with the excess of its maximum envelope above the average level by the value of the specified threshold value.

Thus, the claimed solution has significant differences.

The positive effect is an increase in the noise immunity of target identification is achieved by irradiating the target with a probing signal, which is the sum of the probe pulses, the delays between which correspond to the temporal positions of the significant maxima in the reflected signal from this target (i.e., the location of the most significant in reflectivity elements of the target’s body) in the range of location angles corresponding to the location angle of the target. In this case, the reflected signal is formed as a result of the addition of individual reflected pulses from the totality of the reflecting elements of the target. A maximum is formed in the reflected signal when the reflected pulses are added from all the elements of the target body that are most significant in reflectivity. The level of this maximum corresponds to the level of the reflected long sounding pulse from the entire reflecting surface of the target of complex geometric shape and exceeds the average level of the reflected signal from the target when using sounding pulses with high resolution in range (for example, for extended bodies of complex shape, usually 8-12 dB for all oblique angles of location).

To identify the target, it is enough to determine the excess of this maximum over the average level of the envelope of the reflected signal. Moreover, the level of interference can be commensurate with the average level of the envelope of the reflected signal. To identify goals in the prototype and other known methods of identification (or classification), it is necessary to obtain information about the structure of the reflected signal. In the prototype, it is required to identify significant maxima in relation to their limiting minima, having a level of 18-20 dB less than the maximum level. In the claimed solution, the interference level may correspond to the level of these maxima in the reflected signal.

Thus, in the claimed solution, the interference level is allowed to be 18-20 dB more than in the prototype. Therefore, the noise immunity of the proposed solution is 18-20 dB higher than the prototype and most known methods for identifying (classifying) targets.

Figure 1 shows an example of the formation of the reflected signal from the target, consisting of three reflective elements, when it is irradiated with a short probe pulse and the total signal, where

1 - envelope of the reflected signal;

2 - a probe signal representing the sum of the probe pulses;

3 - signal reflected from the first reflective element;

4 - signal reflected from the second reflective element;

5 - signal reflected from the third reflective element;

6 - total signal reflected from the target.

Figure 2 shows a functional diagram of a device that implements the proposed method, where

8 - probe pulse generator;

9 - power amplifier;

10 - receiving-emitting antenna;

11 - identifiable target;

12 - amplifier premium path;

13 - direction finder;

14 - meter angle of location;

15 - storage device;

16 - key schemes;

17 - adder;

18 - matched filter;

19 - detector;

20 - integrator;

21 is a storage device;

22 - delay line;

23 - subtractive device;

24 - a comparator;

25 - pulse counter;

26 - indicator.

Figure 1 shows an example of an envelope of a reflected signal from a target having a complex geometric shape, i.e. consisting of several reflective elements (in the example of three reflective elements).

When using short tonal sounding pulses or complex sounding signals with subsequent compression, the envelope of the reflected signal 1 contains several significant maxima. Temporary positions of these maxima are indicated by a dotted line.

The transfer function of a reflecting target of complex geometric shape is represented as a set of delta pulses delayed relative to each other for a time τ _i with amplitude factors a _i . When this target is irradiated with a probe pulse ξ (t), the reflected signal S ₁ (t) is formed as the sum of individual pulses ξ (t) reflected from the target elements and is determined as a convolution:

,

,

- переходная функция тела сложной формы;Where

- transition function of the body of complex shape;

a _i - amplitude coefficients characterizing the reflection levels of the probe pulse from individual parts of the body;

δ (t) is the Dirac delta function;

τ _i are the mutual delays between the individual reflected signals.

Knowing the reference averaged values of the temporal positions of the maxima {τ _i } in the reflected signals for various purposes and certain ranges of angles of location, it is possible to form a probe signal in the form of a sum of probe pulses with time delays {τ _j }, located in the opposite order with respect to {τ _i }. The generated probe signal 2 is shown in FIG. 1 in the form of three probe pulses, the delays between which {τ _j } correspond to the temporary positions of the maxima in the reflected signal 1, but are located in the opposite order with respect to {τ _i }.

The probe signal can be represented as:

1, for simplicity, α _i and α _{j are} depicted equal.

The signals reflected from the first, second and third elements of the target 3, 4, 5, are delayed relative to each other in accordance with the temporary positions of the maxima {τ _i } in the reflected signal 1.

In figure 1, the reflected signals 3, 4, 5 for simplicity are presented in the form of a set of individual pulses without taking into account the diffuse component of the reflected signal. The total signal 6 reflected from the target can also be represented as a convolution of the probe signal S ₂ (t) 2 with the transfer function of the reflecting target in the form of a set of delta pulses:

If {τ _j } and {τ _i } coincide, S _Σ (t) is equal to the sum of the reflected signals from all the reflecting elements of the object:

for a _j = 1, for j = 1, 2 ... n.

In the middle of the reflected signal 6, a pronounced maximum is visible, the amplitude of which is equal to the energy sum of the probe pulses reflected from all the reflecting elements of the target.

As mentioned above, the excess of this maximum over the average level of the envelope of the reflected signal, measured in the area that does not capture the maximum, is, for example, at oblique angles of irradiation of a body of complex shape 8 ÷ 12 dB.

In the device that implements the proposed method (figure 2), the generator 8 generates probing pulses with high resolution in range (short compared to the target length, tonal probe pulses or complex probe pulses, for example, with frequency modulation). The probe pulses are amplified by a power amplifier 9 and emitted by an antenna 10. The signals 11 reflected from the target are amplified in the amplifier of the receiving path 12. Using the direction finder 13 and the meter for the angle of location 14, the target location angle is determined by one of the known methods. Information about the angle of location with an error of not more than 15 ° enters the storage device 15, where the reference values of the temporal positions of the maxima {τ _i } in the reflected signals for various purposes and ranges of angles of location are stored. The reference values {τ _i } are selected from the storage device 15 in the reverse order with respect to the first maximum in the reflected signals and are supplied as control signals to the key circuits 16. The probing pulses from the generator 8 also arrive at the key circuits. The probe pulses delayed relative to each other are summed up adder 17, are formed in the total probe and received radiation to the power amplifier 9. The total probing signals generated in accordance with {τ _j} for different purposes, from uchayutsya via the antenna 10. The reflected signals are received via antenna 10, amplified in an amplifier receiving path 12 and fed to the matched filter 18 (matched filter is used only in the case of complex sounding signals). In the matched filter 18, compression of the reflected signals is performed, for example, by correlating them with a copy of the probe pulse (multichannel processing is possible taking into account the Doppler shift of the spectra of the reflected signals). In the detector 9, the envelope of each reflected signal is extracted. The signal from the detector enters the pulse counter 25. The counter counts each reflected signal. The number of each reflected signal coming from the target corresponds to the number of the radiated probe signal, and therefore to the number of the identified target. The signals from the counter are fed to the indicator 26. The average envelope levels of each reflected signal are determined using the integrator 20, stored in the memory 21 and fed to the subtractor 23. In the subtractor 23, the difference between the current values of the envelopes (received from the detector 19 through the delay line is calculated 22) and their average level. The magnitude of this difference is supplied to the comparator 24 and compared with a threshold value. When the threshold value is exceeded, the comparator 24 is triggered. The output signal of the comparator is sent to the indicator 26. When the comparator is activated, its output signal gives permission to display or register using the indicator 26 the identifiable target number.

The proposed method allows to increase the noise immunity of identifying targets of complex shape when using the most informative classification features - temporary positions of the maxima in the reflected signals. This method is resistant to possible distortions of the reflected signals due to the multipath propagation of signals in the sea, since in this case, not one, but several global maximums are formed in the total signal reflected from the target, but the target identification rule does not change. Improving the noise immunity of target identification allows you to increase the range of target recognition or increase the likelihood of their correct recognition.

Claims (2)

1. A method for identifying targets by sonar signals, based on the study of high-resolution sound pulses in range, receiving signals reflected from the target, coordinated filtering of complex signals, determining the target’s location angle with an error of no more than 15 °, characterized in that the reference values of time the positions of the maxima in the reflected signals during training; during identification, the reference values of the temporal positions of significant maxima in the reflected signals for various probing signals and a certain range of location angles corresponding to the value of the measured location angle, probing signals are formed in the form of a sum of probing pulses with reference time delays, and the delays of probing pulses relative to the first pulse in the generated signal correspond to the temporary positions of significant maxima relative to the last maximum in the reflected signal, emit the generated signals in the direction of the target, receive the reflected signals, highlight the envelopes of these signals c, remember the average level of each envelope of the reflected signal and identify the target corresponding to the probing total signal that caused the reflected signal in excess of the maximum of its envelope above the average level by the value of the specified threshold. 2. The method according to claim 1, characterized in that when using complex sounding pulses after emitting the generated total signals and receiving the reflected signals before selecting envelopes of the signals, a coordinated filtering of the reflected signals is performed.

Images