Coefficient adaptive updating method of underwater acoustic target frequency domain line spectrum intensifier

Abstract

The invention discloses a coefficient adaptive updating method of an underwater acoustic target frequency domain line spectrum intensifier (ALE), which comprises the following steps of: initializing parameters of the underwater acoustic target frequency domain line spectrum intensifier, including order, time delay, step length, kernel width, reference value of sparse constraint threshold, intensity of sparse constraint and frequency domain coefficient vector; updating a sparse constraint threshold according to the frequency domain coefficient vector; carrying out time delay and Fourier transform on the input time domain signal, and calculating the output of the underwater acoustic target frequency domain line spectrum enhancer and the error between the output and the input by combining the frequency domain coefficient vector; according to the input and output error and the sparse constraint threshold, updating the frequency domain coefficient vector of the underwater acoustic target frequency domain line spectrum enhancer; and performing loop iteration, and processing all time domain input signals. According to the method, the frequency domain coefficient of the ALE is updated by using a new iterative formula, and the model parameters of the underwater acoustic target frequency domain line spectrum enhancer are adaptively adjusted, so that the performance of the ALE is further improved.

Description

Technical Field

[0001] The invention relates to a technique for extracting and identifying line spectrum features of underwater acoustic target radiation noise, and in particular to a coefficient adaptive updating method of an underwater acoustic target frequency domain line spectrum enhancer (ALE). Background Art

[0002] Underwater acoustic target detection is one of the core issues in the field of underwater acoustic signal processing. According to different processing methods, target detection can be divided into active and passive. Passive detection mainly uses the noise received by passive sonar to determine the presence or absence of the target. As the object of passive sonar detection, the radiation noise of the underwater acoustic target is composed of a broadband continuous spectrum and a narrowband discrete line spectrum. The line spectrum component is usually generated by factors such as the mechanical vibration of the ship and the operation of the propeller. It contains important characteristic information of the target and has the advantages of long propagation distance and high signal-to-noise ratio (SNR). Using the line spectrum in the radiation noise for passive target detection can improve the detection distance, which is of great significance for detecting underwater targets.

[0003] Nowadays, there are still many difficulties in line spectrum detection of target radiation noise. Due to the continuous development of vibration reduction and noise reduction technology in recent years, the concealment of underwater acoustic targets has been continuously improved, and the line spectrum may be submerged in strong background noise and difficult to extract. In addition, the marine environment is complex and changeable, and its background is usually non-white noise with fluctuations. Some traditional technologies are difficult to apply to non-white backgrounds, resulting in a decrease in line spectrum detection performance. Therefore, accurately extracting line spectra in an environment with low signal-to-noise ratio and non-white background is a major problem in underwater acoustic signal processing technology.

[0004] Adaptive line spectrum enhancer (ALE) is a common method for enhancing line spectrum. It uses the correlation of line spectrum and the uncorrelation of broadband noise to improve the line spectrum signal-to-noise ratio through an adaptive algorithm. It has been widely used in various fields. The adaptive algorithm can adjust itself as the external input changes without any prior information, and always maintain the optimal filtering result. At present, conventional ALE uses the least mean square (LMS) algorithm, which has a significant performance degradation as the signal-to-noise ratio decreases. Recently, some scholars have proposed migrating ALE to the frequency domain for iterative updating, and further improving the line spectrum signal-to-noise ratio by imposing constraints on the frequency domain coefficients, but there is no optimization for non-white backgrounds, and it is sensitive to model parameters. Summary of the invention

[0005] The purpose of the present invention is to propose a coefficient adaptive updating method of a frequency domain line spectrum enhancer for underwater acoustic targets, so as to solve the problem that the current frequency domain ALE has insufficient ability to suppress non-white background.

[0006] The technical solution to achieve the purpose of the present invention is: a coefficient adaptive updating method of an underwater acoustic target frequency domain line spectrum enhancer, comprising the following steps:

[0007] Step 1, initializing the parameters of the underwater acoustic target frequency domain line spectrum enhancer, including order, delay, step size, kernel width, reference value of sparse constraint threshold, strength of sparse constraint and frequency domain coefficient vector;

[0008] Step 2, updating the sparse constraint threshold according to the frequency domain coefficient vector;

[0009] Step 3, delay and Fourier transform the input time domain signal, combine the frequency domain coefficient vector, calculate the output of the underwater acoustic target frequency domain line spectrum enhancer and the error between it and the input;

[0010] Step 4, updating the frequency domain coefficient vector of the underwater acoustic target frequency domain line spectrum enhancer according to the error between the input and the output and the sparse constraint threshold;

[0011] Step 5, iterate steps 2-4 in a loop to process all time domain input signals.

[0012] Further, step 1, initialize the parameters of the underwater acoustic target frequency domain line spectrum enhancer, as follows:

[0013] (1.1) Set the order L and delay Δ of the frequency domain line spectrum enhancer of the underwater acoustic target;

[0014] (1.2) Set the parameters μ, σ, and ρ of the underwater acoustic target frequency domain line spectrum enhancer, where μ is the step size and is set to 1

[0015] ×10 -5 ; σ is the kernel width, and its value is determined by Silverman's rule; ρ controls the strength of the sparse constraint, and its value range is 10 -4 ~10 -3 μ;

[0016] (1.3) Set the reference value β0 of the sparse constraint threshold β of the underwater acoustic target frequency domain line spectrum enhancer, and its reciprocal

[0017] The value range of 1 / β0 is 1 / 5 to 1 / 20 of the maximum value of the ideal frequency response;

[0018] (1.4) Initialize the frequency domain coefficient vector of length L Where n is the index of the current moment, represents the L-1th element of the frequency domain coefficient vector, [·] T Represents the transpose of a vector.

[0019] Further, in step 2, the sparse constraint threshold is updated according to the frequency domain coefficient vector, as follows:

[0020] The sparse constraint threshold β is adaptively adjusted through formula (1):

[0021]

[0022] Furthermore, in step 3, the input time domain signal is delayed and Fourier transformed, and the output of the underwater acoustic target frequency domain line spectrum enhancer and the error between the output and the input are calculated by combining the frequency domain coefficient vector, as follows:

[0023] Assume that the original time domain signal at time n is x(n), where n = 1, 2, ..., N, N is the number of data points, and after taking the delay of point Δ, we get x(n-Δ). Then the delayed input vector x(n-Δ) of the underwater acoustic target frequency domain line spectrum enhancer is expressed as

[0024] x(n-Δ)=[x(n-Δ),x(n-Δ-1),...,x(n-Δ-L+1)] T (2)

[0025] Perform Fourier transform on x(n-Δ) and get Its expression is as follows:

[0026]

[0027] Then the output y(n) of the underwater acoustic target frequency domain line spectrum enhancer can be written as

[0028]

[0029] in,(·) H represents the conjugate transpose operator;

[0030] The calculation formula of error e(n) is

[0031] e(n)=x(n)-y(n) (5).

[0032] Furthermore, in step 4, the frequency domain coefficient vector of the underwater acoustic target frequency domain line spectrum enhancer is updated according to the error between the input and the output and the sparse constraint threshold, as follows:

[0033] The sparse constraint threshold β is substituted into the following formula to calculate

[0034]

[0035] Among them, g(·) represents the function operator, |·| represents taking the modulus of the vector element by element, and sgn(·) represents taking the sign of the vector element by element:

[0036]

[0037] Combining equations (3), (5) and (6), the following iterative formula is used to update the frequency domain coefficient vector of the underwater acoustic target frequency domain line spectrum enhancer:

[0038]

[0039] Further, in step 5, steps 2-4 are iterated in a loop to process all time domain input signals, as follows:

[0040] If the input data has not been iterated, return to step 2 to continue updating the frequency domain coefficient vector; otherwise, the iteration ends.

[0041] A coefficient adaptive updating system for a frequency domain line spectrum enhancer of an underwater acoustic target is provided. The system implements the coefficient adaptive updating method for a frequency domain line spectrum enhancer of an underwater acoustic target to achieve the coefficient adaptive updating of the frequency domain line spectrum enhancer of an underwater acoustic target. The system is divided into five modules and executes steps 1 to 5 respectively.

[0042] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the coefficient adaptive updating method of the underwater acoustic target frequency domain line spectrum enhancer is implemented to realize the coefficient adaptive updating of the underwater acoustic target frequency domain line spectrum enhancer.

[0043] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for adaptively updating coefficients of an underwater acoustic target frequency domain line spectrum enhancer is implemented to achieve adaptive updating of coefficients of an underwater acoustic target frequency domain line spectrum enhancer.

[0044] Compared with the prior art, the present invention has the following significant advantages: 1) a new iterative formula is used to update the coefficients of the frequency domain ALE, so that the ALE also has good performance under a non-white background; 2) a parameter adaptive algorithm is used to reduce the sensitivity of the parameters to the ALE performance; 3) the implementation is simple, the physical meaning is clear, and the line spectrum enhancement effect is significant. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a flow chart for implementing the method of the present invention.

[0046] Figure 2 is the structure of ALE in the method of the present invention.

[0047] Figure 3 ] are input and output waveforms in the embodiment.

[0048] Figure 4 : is the input and output power spectrum in the embodiment. It is known that there are three line spectra of 40, 116 and 136 Hz in the frequency band shown. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] A method for adaptively updating coefficients of a frequency domain line spectrum enhancer for an underwater acoustic target comprises the following steps:

[0051] Step 1 initializes the parameters of the underwater acoustic target frequency domain line spectrum enhancer (ALE), which specifically includes the following steps:

[0052] (1.1) Set the order L and delay Δ of frequency domain ALE;

[0053] (1.2) Set the parameters μ, σ, and ρ of the frequency domain ALE, where μ is the step size, σ is the kernel width, and ρ controls the strength of the sparse constraint;

[0054] (1.3) Setting the reference value β0 of the sparse constraint threshold β of the underwater acoustic target frequency domain line spectrum enhancer;

[0055] (1.4) Initialize the frequency domain coefficient vector of length L Where n is the index of the current moment, represents the L-1th element of the frequency domain coefficient vector, [·] T Represents the transpose of a vector.

[0056] Step 2 uses the parameter adaptive algorithm in the present invention to update the parameter β, which specifically includes the following steps:

[0057] The sparse constraint threshold β is adaptively adjusted through formula (1):

[0058]

[0059] Step 3 delays and Fourier transforms the input time domain signal, combines the frequency domain coefficient vector, and calculates the output of the underwater acoustic target frequency domain line spectrum enhancer and the error between it and the input, which specifically includes the following steps:

[0060] Assume that the original time domain signal at time n is x(n), where n = 1, 2, ..., N, and N is the number of data points. After taking the delay of point Δ, we get x(n-Δ), and the delayed input vector x(n-Δ) of frequency domain ALE can be expressed as

[0061] x(n-Δ)=[x(n-Δ),x(n-Δ-1),...,x(n-Δ-L+1)] T (2)

[0062] Perform Fourier transform on x(n-Δ) and get Its expression is as follows:

[0063]

[0064] Then the output y(n) of ALE can be written as

[0065]

[0066] in,(·) H Represents the conjugate transpose operator.

[0067] The calculation formula of error e(n) is

[0068] e(n)=x(n)-y(n) (5).

[0069] Step 4 uses the iterative formula of the present invention to update the frequency domain coefficient vector of ALE, which specifically includes the following steps:

[0070] Substitute the parameter β obtained from equation (1) into the following equation to calculate

[0071]

[0072] Among them, g(·) represents the function operator, |·| represents taking the modulus of the vector element by element, and sgn(·) represents taking the sign of the vector element by element:

[0073]

[0074] Combining equations (3), (5) and (6), the following iterative formula is used to update the frequency domain coefficient vector of ALE:

[0075]

[0076] If the data has not been iterated in step 5, return to step 2 to continue iterating; otherwise, the iteration ends.

[0077] If the input data has not been iterated, return to step 2 to continue updating the frequency domain coefficient vector; otherwise, the iteration ends.

[0078] The present invention also proposes a coefficient adaptive updating system of an underwater acoustic target frequency domain line spectrum enhancer, implements the coefficient adaptive updating method of an underwater acoustic target frequency domain line spectrum enhancer, realizes the coefficient adaptive updating of the underwater acoustic target frequency domain line spectrum enhancer, and is divided into five modules, executing steps 1 to 5 respectively.

[0079] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the coefficient adaptive updating method of the underwater acoustic target frequency domain line spectrum enhancer is implemented to realize the coefficient adaptive updating of the underwater acoustic target frequency domain line spectrum enhancer.

[0080] A computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the method for adaptively updating coefficients of an underwater acoustic target frequency domain line spectrum enhancer is implemented to achieve adaptive updating of coefficients of an underwater acoustic target frequency domain line spectrum enhancer.

[0081] In summary, the present invention uses a new iterative formula to update the frequency domain coefficients of ALE, and adaptively adjusts the model parameters of the underwater acoustic target frequency domain line spectrum enhancer, so as to further improve the performance of ALE.

[0082] Example

[0083] In order to verify the effectiveness of the scheme of the present invention, the following experiment was carried out.

[0084] like Figure 1 , Figure 2 As shown, the following steps are included:

[0085] Step 1 is as follows:

[0086] The processed data is the target radiation noise measured at sea. It is known that it contains three line spectra of 40, 116, and 136 Hz within 200 Hz, with a sampling rate of 1 kHz, a duration of 100 seconds, and the number of data points N = 100,000.

[0087] (1.1) The order L of ALE can generally be 1000~4000, and L=2000 is taken here; the delay Δ can generally be 50~200, and Δ=100 is taken here.

[0088] (1.2) Set the ALE step size μ = 1e-5, kernel width σ = 0.5, and parameter ρ = 1e-9.

[0089] (1.3) Set the reference value β0 of the sparse constraint threshold of the underwater acoustic target frequency domain line spectrum enhancer to 500.

[0090] (1.4) Initialize the frequency domain coefficient vector of ALE

[0091] Step 2 is as follows:

[0092] The parameter β is adaptively adjusted using formula (1).

[0093] Step 3 is as follows:

[0094] Formula (5) is used to calculate the error between the input and output of the underwater acoustic target frequency domain line spectrum enhancer.

[0095] Step 4 is as follows:

[0096] Use equation (8) to update the frequency domain coefficient vector of ALE.

[0097] Step 5 is as follows:

[0098] Let n=n+1. If the input data has not been iterated, that is, n≤100000, return to step 2 to continue updating the frequency domain coefficient vector; otherwise, the iteration ends.

[0099] Figure 3 1 and 2 are waveforms of the ALE input signal and output signal in the embodiment.

[0100] Figure 4 The power spectrum comparison of the ALE input and output in the embodiment. The analysis method used is to use the periodogram method to estimate the power spectrum of any 10 seconds of data and observe the line spectrum enhancement effect. The power spectrum estimation results of the 60-70 second data are shown here. Figure 4 It can be seen that the background noise has obvious non-white characteristics. The method of the present invention has good noise suppression ability and significant line spectrum enhancement effect.

[0101] The above embodiments show that the method of the present invention can effectively enhance the low-frequency line spectrum of the underwater acoustic target radiation noise, and has a significant effect on suppressing non-white background noise, which is beneficial to the subsequent line spectrum detection and extraction work.

[0102] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0103] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A coefficient adaptive updating method for a frequency domain line spectrum enhancer of an underwater acoustic target, characterized in that: The steps include: Step 1, initializing the parameters of the underwater acoustic target frequency domain line spectrum enhancer, including order, delay, step size, kernel width, reference value of sparse constraint threshold, strength of sparse constraint and frequency domain coefficient vector; Step 2, updating the sparse constraint threshold according to the frequency domain coefficient vector; Step 3, delay and Fourier transform the input time domain signal, combine the frequency domain coefficient vector, calculate the output of the underwater acoustic target frequency domain line spectrum enhancer and the error between it and the input; Step 4, updating the frequency domain coefficient vector of the underwater acoustic target frequency domain line spectrum enhancer according to the error between the input and the output and the sparse constraint threshold; Step 5, iterate steps 2-4 in a loop to process all time domain input signals.

2. The coefficient adaptive updating method of the underwater acoustic target frequency domain line spectrum enhancer according to claim 1 is characterized in that: Step 1: Initialize the parameters of the underwater acoustic target frequency domain line spectrum enhancer, as follows: (1.1) Set the order L and delay Δ of the frequency domain line spectrum enhancer of the underwater acoustic target; (1.2) Set the parameters μ, σ, and ρ of the frequency domain line spectrum enhancer of the underwater acoustic target, where μ is the step size, which is 1×10 -5 ; σ is the kernel width, and its value is determined by Silverman's rule; ρ controls the strength of the sparse constraint, and its value range is 10 -4 ~10 -3 μ; (1.3) Setting a reference value β0 of the sparse constraint threshold β of the underwater acoustic target frequency domain line spectrum enhancer, the value range of its reciprocal 1 / β0 is 1 / 5 to 1 / 20 of the maximum value of the ideal frequency response; (1.4) Initialize the frequency domain coefficient vector of length L Where n is the index of the current moment, represents the L-1th element of the frequency domain coefficient vector, [·] T Represents the transpose of a vector.

3. The coefficient adaptive updating method of the underwater acoustic target frequency domain line spectrum enhancer according to claim 1 is characterized in that: Step 2: Update the sparse constraint threshold according to the frequency domain coefficient vector, as follows: The sparse constraint threshold β is adaptively adjusted through formula (1):

4. The coefficient adaptive updating method of the underwater acoustic target frequency domain line spectrum enhancer according to claim 1 is characterized in that: Step 3, delay and Fourier transform the input time domain signal, combine the frequency domain coefficient vector, calculate the output of the underwater acoustic target frequency domain line spectrum enhancer and the error between it and the input, as follows: Assume that the original time domain signal at time n is x(n), where n = 1, 2, ..., N, N is the number of data points, and after taking the delay of point Δ, we get x(n-Δ). Then the delayed input vector x(n-Δ) of the underwater acoustic target frequency domain line spectrum enhancer is expressed as x(n-Δ)=[x(n-Δ),x(n-Δ-1),...,x(n-Δ-L+1)] T (2) Perform Fourier transform on x(n-Δ) and get Its expression is as follows: Then the output y(n) of the underwater acoustic target frequency domain line spectrum enhancer can be written as in,(·) H represents the conjugate transpose operator; The calculation formula of error e(n) is e(n)=x(n)-y(n) (5).

5. The method for adaptively updating coefficients of a frequency domain line spectrum enhancer for underwater acoustic targets according to claim 1, characterized in that: Step 4: Update the frequency domain coefficient vector of the underwater acoustic target frequency domain line spectrum enhancer according to the input and output error and the sparse constraint threshold, as follows: The sparse constraint threshold β is substituted into the following formula to calculate Among them, g(·) represents the function operator, |·| represents taking the modulus of the vector element by element, and sgn(·) represents taking the sign of the vector element by element: Combining equations (3), (5) and (6), the following iterative formula is used to update the frequency domain coefficient vector of the underwater acoustic target frequency domain line spectrum enhancer:

6. The method for adaptively updating parameters of a frequency domain line spectrum enhancer for underwater acoustic targets according to claim 1, characterized in that: Step 5, loop iterate steps 2-4 to process all time domain input signals, as follows: If the input data has not been iterated, return to step 2 to continue updating the frequency domain coefficient vector; otherwise, the iteration ends.

7. A coefficient adaptive updating system for underwater acoustic target frequency domain line spectrum enhancer, characterized in that: The method for adaptively updating coefficients of the underwater acoustic target frequency domain line spectrum enhancer described in any one of claims 1 to 6 is implemented to realize adaptive updating of coefficients of the underwater acoustic target frequency domain line spectrum enhancer, which is divided into five modules and executes steps 1 to 5 respectively.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the coefficient adaptive updating method of the underwater acoustic target frequency domain line spectrum enhancer according to any one of claims 1 to 6 is implemented to realize the coefficient adaptive updating of the underwater acoustic target frequency domain line spectrum enhancer.

9. A computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the coefficient adaptive updating method of the underwater acoustic target frequency domain line spectrum enhancer according to any one of claims 1 to 6 is implemented to realize the coefficient adaptive updating of the underwater acoustic target frequency domain line spectrum enhancer.

Images