Very low frequency modulation feature extraction method based on dechirp processing

Abstract

The invention discloses a very low frequency modulation feature extraction method based on dechirp processing, and the method comprises the steps: carrying out the LOFAR spectrum of a scattered signal received by a single array element, and obtaining a frequency change rate and a modulation frequency; selecting a signal duration T, a time delay tau and a signal initial moment t0, and intercepting a target signal of the corresponding duration and a reference signal of the corresponding time delay; performing conjugate multiplication on the target signal and the reference signal to obtain a demodulation signal, performing Fourier transform to obtain frequency domain information of the demodulation signal, drawing a demodulation spectrum, and preliminarily reading a modulation frequency; obtaining demodulation spectrums of the N array elements according to the method; superposing the obtained N array element demodulation spectrums to obtain a total demodulation spectrum, and reading very low frequency modulation frequency from the spectrum; calculating a corresponding relation between the modulation depth and the demodulation depth; and obtaining very low frequency modulation characteristics. According to the method, the influence of Doppler frequency shift on modulation feature extraction can be effectively removed; and the corresponding relation between the demodulation depth and the modulation depth is obtained, so that water surface / underwater target distinguishing can be performed by using the modulation depth subsequently.

Description

Technical Field

[0001] The invention belongs to the field of underwater acoustic signal processing, and in particular relates to a very low frequency modulation feature extraction method based on de-skewing processing. Background Art

[0002] When an underwater moving body is sailing, it will be affected by the periodic fluctuation force generated by the seawater flow, thus generating extremely low-frequency periodic motion. This periodic motion will excite the fluid medium around the moving body to generate water gravity waves of the same frequency. The water gravity wave will couple with the low-frequency sound emitted by the active sound source to generate a forward scattered sound signal with modulation characteristics (1Semenov AG. Ultra low frequency fields of moving bodies [M]. 2017: 1-240.). In 2021, Wang Huan et al. theoretically and experimentally verified the existence of very low frequency water gravity waves (2Wang H, Fang E, Wu M, et al. Ultra Low Frequency Wave Produced by Underwater Oscillating Sphere and Its Measurement [J]. Journal of Marine Science and Engineering, 2021, 9 (12): 1317.), but did not conduct further research on the modulation effect. In 2023, Jing Chenxuan (3 Jing Chenxuan, Shi Shengguo, Yang Desen, et al. Study on the mechanism and characteristics of acoustic scattering modulation of underwater low-frequency oscillating eddy current field [J]. Acta Physica Sinica, 2023, 72(01): 221-235.) and others studied the generation mechanism and time-space frequency characteristics of acoustic scattering modulation of underwater low-frequency oscillating eddy current field. So far, the research on underwater very low frequency modulated acoustic scattering has focused more on the generation mechanism, and there is no relevant research on the extraction of very low frequency modulation features.

[0003] When using bistatic sonar to detect underwater moving objects, the relative motion between the target and the receiving array will inevitably cause Doppler shift. Doppler shift and very low frequency modulation are very likely to appear simultaneously in the scattered signal. The Doppler shift and the modulation frequency are very close in value. Direct envelope spectrum extraction cannot extract the modulation features, so the Doppler shift needs to be removed first. Dechirp technology is mainly used in the field of radar signal processing to demodulate linear frequency modulation signals to obtain the distance information of the target. Since the underwater target does not move very fast, the Doppler shift generated when it crosses the transmit and receive line is similar in form to the linear frequency modulation signal.

[0004] Based on this special scenario and previous theoretical and methodological research, the present invention utilizes the principle of de-skewing to remove the influence of Doppler frequency shift in the scattered signal, thereby realizing the extraction of very low frequency modulation features. Summary of the invention

[0005] The object of the present invention is to provide a very low frequency modulation feature extraction method based on de-skewing processing, which extracts the modulation features carried in the target forward scattered signal with Doppler frequency shift, including the modulation frequency and the modulation depth.

[0006] The purpose of the present invention is achieved through the following technical solutions:

[0007] A very low frequency modulation feature extraction method based on de-skewing processing, the specific steps are as follows:

[0008] Step 1: Make a LOFAR spectrum for the scattered signal received by a single array element and estimate the frequency change rate and modulation frequency;

[0009] Step 2: According to step 1, select the signal duration T, the delay τ and the initial time t0 of the signal, and intercept the target signal of the corresponding duration and the reference signal of the corresponding delay;

[0010] Step 3: According to step 2, the target signal is conjugate-multiplied with the reference signal to obtain a demodulated signal;

[0011] Step 4: According to step 3, perform Fourier transform on the demodulated signal to obtain the frequency domain information of the demodulated signal, draw the demodulation spectrum, and preliminarily read the modulation frequency from it;

[0012] Step 5: Repeat steps 2 to 4 for the received signals of the N array elements to obtain the demodulation spectra of the N array elements;

[0013] Step 6: Superimpose the N array element demodulation spectra obtained in step 5 to obtain a total demodulation spectrum, and read the very low frequency modulation frequency from the spectrum;

[0014] Step 7: Calculate the corresponding relationship between modulation depth and demodulation depth;

[0015] Step 8: Based on the modulation frequency obtained in step 6 and the demodulation depth obtained in step 7, the very low frequency modulation characteristics are obtained comprehensively.

[0016] Furthermore, in step 1, the modulation depth of the received signal of the single array element is m, 0<m<1, and the modulation frequency is f0. Due to the movement of the target, Doppler frequency shift will be caused; a LOFAR spectrum is made, and the frequency change rate and the range of the modulation frequency are read from the spectrum.

[0017] Furthermore, in step 2, the signal segment from t0 to t0+T is intercepted as the target signal, and the signal segment from t0+τ to t0+τ+T is intercepted as the reference signal.

[0018] Furthermore, in step 3, since the signal form with Doppler frequency shift is very close to the linear frequency modulation signal, the received scattered signal is simplified to a linear frequency modulation signal carrying modulation, that is:

[0019] r(t)=exp[j·2π(f l +0.5kt)t]

[0020] +m·exp[j·2π(f l -f0+0.5kt)t]

[0021] +m·exp[j·2π(f l +f0+0.5kt)t]

[0022] Among them, the frequency modulation range of the linear frequency modulation pulse is f l -f h , the pulse width is T, then its linear frequency modulation slope is k = (f h -f l ) / T, f0 is the modulation frequency, m is the modulation depth, and 0<m<1; the delay difference between the reference signal and the target signal is τ, then the form of the reference signal is:

[0023] r ref (t) = exp[j·2πf l (t-τ)+jπk(t-τ) 2 ]

[0024] +m·exp[j·2π(f l -f0)(t-τ)+jπk(t-τ) 2 ]

[0025] +m·exp[j·2π(f l +f0)(t-τ)+jπk(t-τ) 2 ]

[0026] Multiplying the formula for the chirp signal carrying the modulation conjugate with the formula for the reference signal yields:

[0027] y(t)=r(t)·r ref * (t)

[0028] =exp[j·2πkτt]exp[j·2π(f l τ-0.5kτ 2 )]·[1+m 2 ·exp(-j·2πf0τ)+m 2 ·exp(j·2πf0τ)]

[0029] +m·exp[j·2π(kτ-f0)t]exp[j·2π(f l τ-0.5kτ 2 )]·[1+exp(-j·2πf0τ)]

[0030] +m·exp[j·2π(kτ+f0)t]exp[j·2π(f l τ-0.5kτ 2 )]·[1+exp(j·2πf0τ)]

[0031] +m 2 ·exp[j·2π(kτ-2f0)t]exp[j·2π(f l τ+f0τ-0.5kτ 2 )]

[0032] +m 2 ·exp[j·2π(kτ+2f0)t]exp[j·2π(f l τ-f0τ-0.5kτ 2 )]

[0033] The linear frequency modulation signal carrying modulation is demodulated to obtain a single-frequency signal carrying two modulation frequencies. Since m is small, m 2 smaller, the kτ-2f0 and kτ+2f0 components are ignored; the obtained carrier is a single-frequency amplitude modulated signal, which is the demodulated signal.

[0034] Furthermore, when the signal-to-noise ratio is low, the modulation component of the total demodulation spectrum in step 6 is obvious and the modulation frequency is easy to read.

[0035] Furthermore, there is a certain quantitative relationship between the amplitudes of the frequency components after demodulation in step 7. When the influence of the phase is ignored, the amplitude of the kτ component is (1+2m 2 ), the corresponding magnitude of the amplitude of the kτ-f0 component and the kτ+f0 component is 2m, and the corresponding magnitude of the amplitude of the kτ-2f0 component and the kτ+2f0 component is m 2 ; Define the average difference between the modulation component and the peak value of the array modulation spectrum as the demodulation depth, that is:

[0036]

[0037] Among them, Δ is the difference caused by the phase influence.

[0038] The beneficial effects of the present invention are:

[0039] 1. The present invention can effectively remove the influence of Doppler frequency shift on modulation feature extraction;

[0040] 2. The present invention can accurately obtain the modulation frequency;

[0041] 3. The present invention can obtain the corresponding relationship between demodulation depth and modulation depth, which is convenient for subsequent use of modulation depth to distinguish surface / underwater targets;

[0042] 4. The calculation method of the present invention is simple and the calculation result is accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 It is a flow chart of the modulation feature extraction method.

[0044] Figure 2 The LOFAR spectrum of a single array element signal is given.

[0045] Figure 3 The demodulated spectrum of a single array element signal is given.

[0046] Figure 4 The total demodulated spectrum of all array element signals is given.

[0047] Figure 5 The total demodulated spectrum with a modulation depth of 0.05 is given.

[0048] Figure 6 The total demodulated spectrum with a modulation depth of 0.5 is given.

[0049] Figure 7 A comparison of demodulation conditions at different modulation depths is given. DETAILED DESCRIPTION

[0050] The present invention is further described below in conjunction with the accompanying drawings.

[0051] The present invention provides a very low frequency modulation feature extraction method based on de-skewing processing, and the specific steps are as follows:

[0052] Step 1: Make a LOFAR spectrum for the scattered signal received by a single array element and estimate the frequency change rate and modulation frequency;

[0053] The modulation depth of the received signal of a single array element is m (0<m<1), and the modulation frequency is f0. Due to the movement of the target, Doppler frequency shift will be caused. The LOFAR spectrum is made for it, and the approximate range of the frequency change rate and the modulation frequency can be read from the spectrum. In the present invention, the modulation depth is 0.2, and the modulation frequency is 0.25Hz. The LOFAR spectrum of the received scattered signal of a single array element obtained by simulation is as follows: Figure 2 As shown, due to the existence of Doppler frequency shift, the received signal is close to a linear frequency modulation signal, the frequency change rate is about 0.006 Hz / s, and the modulation frequency is less than 0.5 Hz.

[0054] Step 2: Select the signal duration T, delay τ and the initial time t0 of the signal, and intercept the target signal of the corresponding duration and the reference signal of the corresponding delay;

[0055] When the signal duration is 100s and the delay is 0.5s, the signal segment with obvious modulation phenomenon is intercepted according to the LOFAR spectrum. In the present invention, the 200s-300s segment signal is intercepted as the target signal, and the 200.5s-300.5s segment signal is intercepted as the reference signal.

[0056] Step 3: Conjugate and multiply the target signal and the reference signal to obtain a demodulated signal;

[0057] Since the signal form with Doppler frequency shift is very close to the linear frequency modulation signal, the received scattered signal can be simplified to the linear frequency modulation signal carrying modulation, that is,

[0058]

[0059] In which, it is assumed that the frequency modulation range of the linear frequency modulation pulse is f l -f h , the pulse width is T, then its linear frequency modulation slope is k = (f h -f l ) / T, f0 is the modulation frequency, m is the modulation depth, and 0<m<1. Assuming the delay difference between the reference signal and the target signal is τ, the form of the reference signal is:

[0060]

[0061] Multiplying (1) by the conjugate of (2) yields:

[0062]

[0063] The linear frequency modulation signal carrying modulation can be demodulated to obtain a single-frequency signal carrying two modulation frequencies. However, since m is small, m 2 Even smaller, the components kτ-2f0 and kτ+2f0 can be almost ignored.

[0064] The present invention conjugate-multiplies the intercepted target signal and the reference signal, and the obtained carrier is a single-frequency amplitude modulation signal, which is the demodulation signal.

[0065] Step 4: Perform Fourier transform on the demodulated signal to obtain the frequency domain information of the demodulated signal, draw the demodulation spectrum, and preliminarily read the modulation frequency from it;

[0066] The single-element demodulation spectrum obtained by the present invention is as follows: Figure 3 As shown, it can be clearly seen that the modulation frequency is about 0.25Hz.

[0067] Step 5: Repeat steps 2 to 4 to obtain the demodulation spectra of all array elements;

[0068] Repeat steps 2 to 4 for the received signals of N array elements to obtain N demodulation spectra. The present invention simulates 32 array elements, so 32 single-element demodulation spectra can be obtained.

[0069] Step 6: Superimpose the demodulated spectra of the 32 array elements obtained in step 5 to obtain a total demodulated spectrum, from which the very low frequency modulation frequency can be read;

[0070] The total demodulation spectrum of the present invention is as follows Figure 4 As shown in the figure, compared with the single-element demodulation spectrum, the modulation component is more obvious and the modulation frequency is easier to read. When the signal-to-noise ratio is low, the advantage of the total demodulation spectrum will be more obvious.

[0071] Step 7: Calculate the corresponding relationship between modulation depth and demodulation depth;

[0072] There is a certain quantitative relationship between the amplitudes of the frequency components after demodulation. When the influence of the phase is ignored, the amplitude of the kτ component is (1+2m 2 ), the corresponding magnitude of the amplitude of the kτ-f0 component and the kτ+f0 component is 2m, and the corresponding magnitude of the amplitude of the kτ-2f0 component and the kτ+2f0 component is m 2 The present invention simulates the received signals with modulation depths of 0.05 and 0.5, and the array demodulation spectrum is as follows: Figure 5 and Figure 6 As shown in the figure, the demodulation comparison of different modulation depths is as follows: Figure 7 The average difference between the modulation component and the peak value of the array modulation spectrum is defined as the demodulation depth, that is:

[0073]

[0074] Among them, Δ is the difference caused by the phase influence, which is determined by the actual situation.

[0075] Step 8: Based on the modulation frequency obtained in step 6 and the demodulation depth obtained in step 7, the very low frequency modulation characteristics are obtained comprehensively.

[0076] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A very low frequency modulation feature extraction method based on de-skewing processing, characterized in that: The specific steps are as follows: Step 1: Make a LOFAR spectrum for the scattered signal received by a single array element and estimate the frequency change rate and modulation frequency; Step 2: According to step 1, select the signal duration T, the delay τ and the initial time t0 of the signal, and intercept the target signal of the corresponding duration and the reference signal of the corresponding delay; Step 3: According to step 2, the target signal is conjugate-multiplied with the reference signal to obtain a demodulated signal; Step 4: According to step 3, perform Fourier transform on the demodulated signal to obtain the frequency domain information of the demodulated signal, draw the demodulation spectrum, and preliminarily read the modulation frequency from it; Step 5: Repeat steps 2 to 4 for the received signals of the N array elements to obtain the demodulation spectra of the N array elements; Step 6: Superimpose the N array element demodulation spectra obtained in step 5 to obtain a total demodulation spectrum, and read the very low frequency modulation frequency from the spectrum; Step 7: Calculate the corresponding relationship between modulation depth and demodulation depth; Step 8: Based on the modulation frequency obtained in step 6 and the demodulation depth obtained in step 7, the very low frequency modulation characteristics are obtained comprehensively.

2. The very low frequency modulation feature extraction method based on de-skewing processing according to claim 1, characterized in that: In the step 1, the modulation depth of the received signal of the single array element is m, 0<m<1, and the modulation frequency is f0. Due to the movement of the target, Doppler frequency shift will be caused; LOFAR spectrum is made, and the frequency change rate and the range of the modulation frequency are read from the spectrum.

3. The very low frequency modulation feature extraction method based on de-skewing processing according to claim 1, characterized in that: In step 2, the signal from t0 to t0+T is intercepted as the target signal, and the signal from t0+τ to t0+τ+T is intercepted as the reference signal.

4. The method for extracting very low frequency modulation features based on de-skewing according to claim 1, characterized in that: In step 3, since the signal form with Doppler frequency shift is very close to the linear frequency modulation signal, the received scattered signal is simplified to a linear frequency modulation signal carrying modulation, that is: r(t)=exp[j·2π(f l +0.5kt)t] +m·exp[j·2π(f l -f0+0.5kt)t] +m·exp[j·2π(f l +f0+0.5kt)t] Among them, the frequency modulation range of the linear frequency modulation pulse is f l -f h , the pulse width is T, then its linear frequency modulation slope is k = (f h -f l ) / T, f0 is the modulation frequency, m is the modulation depth, and 0<m<1; the delay difference between the reference signal and the target signal is τ, then the form of the reference signal is: r ref (t)=exp[j·2πf l (t-τ)+jπk(t-τ) 2 ] +m·exp[j·2π(f l -f0)(t-τ)+jπk(t-τ) 2 ] +m·exp[j·2π(f l +f0)(t-τ)+jπk(t-τ) 2 ] Multiplying the formula for the chirp signal carrying the modulation conjugate with the formula for the reference signal yields: y(t)=r(t)·r ref * (t) =exp[j·2πkτt]exp[j·2π(f l t-0.5kt 2 )]·[1+m 2 ·exp(-j·2πf0τ)+m 2 ·exp(j·2πf0τ)] +m·exp[j·2π(kτ-f0)t]exp[j·2π(f l t-0.5kt 2 )]·[1+exp(-j·2πf0τ)] +m·exp[j·2π(kτ+f0)t]exp[j·2π(f l t-0.5kt 2 )]·[1+exp(j·2πf0τ)] +m 2 ·exp[j·2π(kτ-2f0)t]exp[j·2π(f l τ+f0τ-0.5kτ 2 )] +m 2 ·exp[j·2π(kτ+2f0)t]exp[j·2π(f l τ-f0τ-0.5kτ 2 )] The linear frequency modulation signal carrying modulation is demodulated to obtain a single-frequency signal carrying two modulation frequencies. Since m is small, m 2 smaller, the kτ-2f0 and kτ+2f0 components are ignored; the obtained carrier is a single-frequency amplitude modulated signal, which is the demodulated signal.

5. The method for extracting very low frequency modulation features based on de-skewing according to claim 1, characterized in that: Compared with the demodulation spectrum of a single array element in step 6, when the signal-to-noise ratio is low, the modulation component of the total demodulation spectrum is obvious and the modulation frequency is easy to read.

6. The method for extracting very low frequency modulation features based on de-skewing according to claim 1, characterized in that: There is a certain quantitative relationship between the amplitudes of the frequency components after demodulation in step 7. When the influence of the phase is ignored, the amplitude of the kτ component is (1+2m 2 ), the corresponding magnitude of the amplitude of the kτ-f0 component and the kτ+f0 component is 2m, and the corresponding magnitude of the amplitude of the kτ-2f0 component and the kτ+2f0 component is m 2 ; Define the average difference between the modulation component and the peak value of the array modulation spectrum as the demodulation depth, that is: Among them, Δ is the difference caused by the phase influence.

Images