RU2555194C1 - Processing of hydroacoustic signal from noise-emitting object - Google Patents

Abstract

FIELD: physics.

SUBSTANCE: proposed method comprises reception of noise-emission signal, determination of noise signal and interference spectrum, accumulation of results and their comparison. Noise-radiation signal is received by signal antenna output to isolate the spectrum actual section. Then, performed are isolation of spectrum virtual section, reiteration of spectrum actual section isolation and reiteration of spectrum virtual section for N serial sets. Sid actual section and virtual sections are summed for N sets. Sum of actual section is squared as well as that of virtual sections. Energy spectrum of sum signal is defined. Simultaneously, the same initial data are used to define the energy spectrum of said difference as the sum of squares of N sets of the difference of actual sections and the sum of squares of N seta of the difference of virtual sections. Decision on the object noise-ration signal availability is make in case the sum energy spectrum exceeds that of the difference energy spectrum.

EFFECT: automatic detection of hydroacoustic signals.

1 dwg

Description

The invention relates to hydroacoustics, technical acoustics and can be used in the development of hydroacoustic systems for automatic detection of noise signals of objects.

The tasks of processing hydroacoustic signals of object noise include determining the signal level and the level of interference, determining the signal-to-noise ratio, which is one of the main parameters that determine the noise immunity of detection systems and their main characteristics, such as the probability of correct detection and the probability of false alarm. Therefore, there is a need to determine these values when working in real conditions with a specific noise signal situation and to determine the fact of detection of a noise signal (A.M. Tyurin Introduction to the theory of statistical methods in hydroacoustics. L. 1963, ed. VMOLA p.116).

, где S₁ - площадь автокорреляционной функции помехи и сигнала до точки перегиба; S₂ - площадь автокорреляционной функции сигнала до точки перегиба; S₃ - площадь автокорреляционной функции сигнала после точки перегиба.A known method of measuring the signal-to-noise ratio according to RF patent No. 2466416, in which a noise signal and a noise signal are received, a spectrum of a received mixture of a noise signal and a noise is determined, a secondary spectrum of a mixture of a noise signal and a noise or an autocorrelation function (ACF) is determined, an inflection point of an ACF is determined; determine S ₁ as the sum of the ACF samples from the zero reference to the reference at the inflection point, determine S ₃ as the sum of the ACF samples from the reference of the inflection point to the end of the ACF, determine S ₂ as the product of the ACF at the inflection point by the number of samples from the zero reference to the point inflection, and the signal-to-noise ratio is determined by the formula $$\frac{S_3+S_2}{S_{\mathrm{one}}-S_2}$$

where S ₁ is the area of the autocorrelation function of the noise and the signal to the inflection point; S ₂ - the area of the autocorrelation function of the signal to the inflection point; S ₃ is the area of the autocorrelation function of the signal after the inflection point.

The disadvantage of this method is the complexity of processing the envelope of the autocorrelation function and the difficulty of obtaining an estimate of the signal-to-noise ratio, which substantially depends on the sampling frequency of the input process. In addition, the processing complexity does not allow for automatic decision making on the presence of an object of noise emission.

A known method of processing a hydro-acoustic noise signal from an object, which comprises receiving a mixture of a noise signal and interference by two identical receivers or semi-antennas, the distance between which exceeds the interference correlation interval, determining the sum of signals received by two receivers, determining the difference of signals received by two receivers, spectral analysis, detection, integration and comparison of the result (A.M. Tyurin Introduction to the theory of statistical methods in hydroacoustics L.1963, ed. VMOLA p.172). It is assumed that the interference acting on the receivers is statistically independent, then the difference in the signals received by the two receivers will determine the interference. The sum of the signals received by the two receivers will determine the sum of the signal + interference. The difference in energy of the sum of signals + interference and interference energy will determine the signal, and the signal-to-noise ratio will be determined as the quotient of dividing the difference in the sum of signals + interference and the difference in signals received by two receivers, the decision on the presence of a signal is made by the difference of these signals and the magnitude of the ratio .

The disadvantage of this method is the need for two identical signal processing channels received in the sum channel and in the difference channel, which in real conditions causes significant difficulties, since over time due to aging the characteristics of the elements change and the channels become non-identical. In addition, it is necessary to have two processing channels, the receivers of which are spaced apart by an interval exceeding the interference correlation interval, which is not always possible. The identity of the measurement channels can be caused by spatial anisotropy of the interference or reception of signals of high-noise objects by the side lobes of the directivity characteristic, which, with further accumulation of the energy spectra of the sum and difference, will introduce an error in determining the energy spectrum of the signal and the interference spectrum and make it difficult to make an automatic decision about the presence of a noise signal at the input .

The objective of the invention is to remedy these disadvantages, and the technical result of the method is to simplify equipment that implements a method for processing hydroacoustic noise emission of an object, and ensuring automatic decision-making on the presence of a hydroacoustic signal of noise emission from an object when using a single-channel signal processing system

The claimed technical result is achieved by the fact that in the method of processing the hydroacoustic signal of the noise of the object, containing the reception of the hydroacoustic signal of the noise in the mixture with interference, determining the spectrum of the electrical signal of the noise of the object, determining the electrical spectrum of the noise, accumulation, comparison of the results, new features are introduced, namely: reception the hydroacoustic signal of the noise emission of an object is produced from the output of a single antenna; the electrical signal of noise emission is sampled object, fast Fourier transform of a set of a discretized electrical signal with the formation of a complex spectrum, for all N consecutive sets, select the real part of the complex spectrum of this signal and select the imaginary part of the complex spectrum of the same signal, sum the real parts of the complex spectrum over N sets and sum the imaginary parts of the complex spectrum in N sets, squared the sum of the real parts of the complex spectrum, squared the sum of the imaginary parts of the complex spectrum, sum the square of the sum of the real parts of the complex spectrum and the square of the sum of the imaginary parts of the complex spectrum, at the same time using the same source data, sequentially subtract the real parts of the complex spectrum from N sets, subtract the imaginary parts of the complex spectrum from N sets, square the obtained difference of the real parts of the complex spectrum of N sets, squared the obtained difference of the imaginary parts of the complex spectrum of N sets, sum the square of the difference real parts of the complex spectrum and the square of the difference of the imaginary parts of the complex spectrum, and the decision on the presence of an object noise signal is made if the sum of the square of the sum of the real parts of the complex spectrum and the square of the sum of the imaginary parts of the complex spectrum is greater than the sum of the square of the difference of the real parts of the complex spectrum and the square the difference of the imaginary parts of the complex spectrum, while the number N must be even and determined by the noise power of the object, the value of which is chosen to be detected the use of low-noise objects and the detection of low-noise objects after making a decision about the absence of a noise signal.

The physical basis of the proposed method for processing the hydroacoustic signal of the noise emission of an object is based on the well-known fact that the variance of the total or difference stationary process will be equal to the sum of the variances of the components (V. T. Goryainov, A. G. Zhuravlev, V. I. Tikhonov “Examples and tasks on statistical radio engineering "M. Sov. radio. 1970. p.111). Dispersion in the applied sense is the power of the process (A.M. Zaezdny, “Fundamentals of Calculations for Statistical Radio Engineering.” M. Svyaz 1969, p. 37). Thus, the variance of the sum of two independent uncorrelated random processes is equal to the sum of the variances of the terms and the variance of the difference of independent uncorrelated random processes is equal to the sum of the variances subtracted and they will be equal to each other (A.M. Tyurin “Introduction to the theory of statistical methods in hydroacoustic” L.1963 Vol. 11, p. 12). In the proposed technical solution, when processing sonar signals, it is not the energy spectra that are used in modern systems that are used, but the complex spectra, which are a separate real part and an imaginary part, each of which is an alternating process with zero mean. And the prototype uses the energy spectra of the signal and interference, the average value of which is not equal to zero after accumulation. The energy spectrum is the end product of spectral processing and is defined as the sum of the squares of the real and imaginary parts, and the complex spectrum is the intermediate component, which later turns into the energy spectrum (“The use of digital signal processing”, Oppenheim M. Mir 1980) pg. 389-436).

The noise signal from the object is stationary in nature and the spectral composition that is constant over a long period of time and the sequential sets of this process are dependent. The spectral properties of the signal are determined over the entire time interval of observation and are coherently connected in time-sequential sets of discretized samples that are sequentially fed to the processing system. With the accumulation of the real and imaginary parts of the complex spectrum of the signal for discretized sets of temporal realization, the position of the spectral component of the signal along the frequency axis will be determined. In this case, the summation of the positive real parts and the summation of the negative real parts of the signal component occurs. The spectral relations of interference between sets of time realizations spaced by a time interval greater than the correlation interval for the interference will turn out to be independent and uncorrelated, which almost always takes place in real conditions. If there is no signal in the dialed implementations, but there is only interference, then random summation of the interference will occur. Since the interference in adjacent time sequences is not correlated, the real part of the interference will not add up coherently with the real part of the interference of the subsequent set. Similarly, the imaginary part of the interference will not add up coherently with the imaginary part of subsequent sets. Also, in the difference channel, the interference in adjacent time sequences is not correlated, and the real part of the interference will not be subtracted coherently with the real part of the interference of the subsequent set. If there is a signal in the difference channel in adjacent time sequences, the signal is correlated, and the real part of the signal will be subtracted coherently with the real part of the subsequent signal, and the interference is not coherent. The average value of the total process and the average value of the difference process will be equal to zero, since the average values of the individual components of the initial process are equal to zero. If the output energy spectrum (the sum of the square of the real part and the sum of the squares of the imaginary part of the complex spectrum) (“Using Digital Signal Processing”, Oppenheim M. Mir 1980, p. 389-436) will be larger than the output energy spectrum of the difference of the complex spectra, then this means that there is noise emission from the object at the input of the receiver. If the output energy spectrum of the sum of the complex spectra is equal to the output energy spectrum of the difference of the complex spectra, then this means that there is no noise emission from the object at the input of the receiver, and the number of time series sets must be even. As a rule, objects of noise emission are subdivided into high noise and low noise. For a highly noisy object, a decision on the presence of a noise signal can be made with a small number of accumulations, which will immediately allow us to proceed to the solution of the classification problem and measure the motion parameters, but if the solution is not adopted, this does not mean that there is no low-noise object in this direction. To detect low-noise objects, it is necessary to increase the accumulation time. Therefore, a two-stage procedure is provided, containing a short accumulation time and a quick decision, and continued accumulation to make a decision about the presence of a low-noise object.

The invention is illustrated in figure 1, which shows a block diagram of a device that implements this method.

In figure 1, the antenna and the receiving device 1 through the sampler 2 are connected to the input of the FFT unit 3. The first, second, third and fourth outputs of block 3 are connected to the inputs of drive 4 of the sum of the real part of the complex spectrum, drive 5 of the sum of the imaginary part of the complex spectrum, drive 7 of the difference of the imaginary part of the complex spectrum and drive 8 of the difference of the imaginary part of the complex spectrum, respectively. The outputs of drives 4 and 5 are connected to the inputs of block 6 for calculating the sum of squares of the accumulated sum of the real part of the complex spectrum and the accumulated sum of the imaginary part of the complex spectrum. The outputs of blocks 7 and 8 are connected to the inputs of block 9 for calculating the sum of squares of the accumulated difference of the real part of the spectrum and the accumulated difference of the imaginary part of the spectrum. The outputs of blocks 6 and 9 are connected to the inputs of the comparison and decision block 10, and the output of block 10 is connected to the second input of the FFT block 3.

The implementation of the proposed method will be shown on the example of the system (figure 1). From the output of the antenna and the receiving device 1, the input process enters the input of the sampler 2, which converts the analog signal into a digital form and transfers it to the FFT unit 3 with fixed serial arrays, where, according to well-known algorithms, the complex spectrum of the input implementation of the original analog process is calculated. At the output of block 3, FFTs form during the calculation the real and imaginary parts of the spectrum, each of which contains the positive and negative parts. The obtained estimate of the real part of the complex spectrum from the first output of the FFT block 3 is transferred to drive 4, and from the second output of the FFT block 3, the estimate of the imaginary part of the complex spectrum is transferred to drive 5. The accumulated estimates of the real and imaginary part of the complex spectrum containing a mixture of the object noise signal and interference are served on block 6, where they are squared and summed, forming the energy spectrum of the sum. From the third output of the FFT block 3, the imaginary part of the complex spectrum is evaluated in block 7, where the real part of the previous set of the complex spectrum is subtracted from the real part of the subsequent set of the complex spectrum and so on as the sets arrive, and from the fourth output of the FFT block 3 to block 8 the imaginary part of the complex spectrum arrives, where the imaginary part of the complex spectrum is subtracted from the previous set of the next set of the imaginary part of the complex spectrum and the sum of the difference of the real parts is accumulated complex spectrum. The obtained estimates of the accumulated difference of the real parts and the accumulated difference of the imaginary parts go to block 9, where each of the components is squared and summed with each other, forming the energy spectrum of the difference. The obtained estimates of the accumulated difference and the accumulated sum of the final energy spectra are sent to the comparison unit 10, where a decision is made on the presence or absence of noise from the object in the received signal. If the received energies of the complex estimates of the sum and the difference are equal, then this means that only one interference acts at the input. In this case, a signal is transmitted to the FFT block 3 to increase the number of accumulations. If the level of the energy spectrum of the sum is greater than the level of the energy spectrum of the difference, then this means that the received signal contains noise emission from the object.

The principles of digital conversion and processing are described in sufficient detail in the work (“Application of Digital Signal Processing” p / r Oppenheim M. Mir 1980, p. 389-436) When using digital technology, the Fast Fourier Transform (FFT) procedures are used as spectral analysis, which provide the selection and measurement of the energy spectrum of a noise electrical process. Currently, almost all hydroacoustic equipment is performed on special processors that convert the acoustic signal into digital form and digitally generate directivity characteristics, multichannel processing and signal detection, as well as measuring the spectra of the noise signal, autocorrelation processing and spectral analysis procedures. Issues of the implementation of special processors are considered in sufficient detail in the book of Yu.A. Koryakin, S.A. Smirnov, G.V. Yakovlev “Ship hydroacoustic equipment” St. Petersburg “Science” 2004, p. 281.

Thus, a technical result is achieved associated with the determination of signal energy, interference energy using a single-channel receiving system, and when comparing the processing results, an automatic decision is made about the presence or absence of an object noise signal at the receiver input.

Claims (1)

A method for processing a hydroacoustic signal of an object’s noise radiation, comprising receiving a hydroacoustic signal of a noise emission in a mixture with interference, determining a spectrum of an electrical noise signal of an object, determining an electrical interference spectrum, accumulating, comparing results, characterized in that a hydroacoustic signal of an object noise emission is produced from the output of a single antenna, discretization of the electrical signal of the noise emission of an object, fast Fourier transform of a set of discretized electron a ctric signal with the formation of a complex spectrum, for all N sequential sets, the real part of the complex spectrum of this signal is extracted and the imaginary part of the complex spectrum of the same signal is extracted, the real parts of the complex spectrum are summed over N sets and the imaginary parts of the complex spectrum are summed over N sets, square the sum of the real parts of the complex spectrum, square the sum of the imaginary parts of the complex spectrum, sum the square of the sum of the real parts of the set xn spectrum and the square of the sum of the imaginary parts of the complex spectrum, at the same time using the same initial data, sequentially subtract the real parts of the complex spectrum from N sets, subtract the imaginary parts of the complex spectrum from N sets, square the resulting difference of the real parts of the complex spectrum of N sets , squared the resulting difference in the imaginary parts of the complex spectrum of N sets, summarize the square of the difference in the real parts of the complex spectrum and the square of the difference in the imaginary parts th complex spectrum, and the decision on the presence of an object noise signal is made if the sum of the square of the sum of the real parts of the complex spectrum and the square of the sum of the imaginary parts of the complex spectrum is greater than the sum of the square of the difference of the real parts of the complex spectrum and the square of the difference of the imaginary parts of the complex spectrum, the number N must be even and determined by the noise power of the object, the value of which is selected to detect high-noise objects and detect low-noise objects after Answered decision on the absence of noise emission signal.