A method and system for underwater acoustic interference-resistant transmission based on sparse time-frequency feature mapping

Abstract

This invention discloses an underwater acoustic anti-interference transmission method and system based on sparse time-frequency feature mapping. The method includes: at the transmitting end, adaptively generating a fractional-order linear frequency-modulated waveform with a specific time-frequency shear slope based on the Doppler state of the underwater acoustic channel, achieving physical-layer focusing of transmitted energy; at the receiving end, constructing a hybrid observation model containing wide-block sparse channel components and narrow-block sparse burst noise components, and using a dual-channel variational Bayesian algorithm to jointly iteratively infer the posterior probability distribution of environmental burst noise in the fractional-order domain; finally, recovering the original signal through soft-threshold interference cancellation and fractional-order channel equalization. This invention effectively solves the communication failure problem caused by high-dynamic Doppler diffusion and marine biological impulse noise interference in underwater acoustics by actively mapping the waveform and using heterogeneous sparse joint inference at the receiving end, significantly improving transmission reliability in harsh underwater acoustic environments.

Description

Technical Field

[0001] This invention relates to the technical field of underwater wireless communication and signal processing, and in particular to an underwater acoustic anti-interference transmission method and system based on sparse time-frequency feature mapping. Background Technology

[0002] Underwater acoustic communication channels are widely recognized as one of the most complex wireless communication channels in the world, facing two major challenges: first, the large-scale Doppler effect caused by the relative motion of underwater equipment and ocean waves; and second, the widespread burst impulse noise in the marine environment (such as the sound of pistol shrimp closing and bubble breaking).

[0003] Currently, Orthogonal Frequency Division Multiplexing (OFDM) technology is widely used due to its high spectral efficiency. For underwater acoustic environments, existing technologies attempt to estimate channel parameters at the receiver using fractional Fourier transform (FRFT) or to utilize compressed sensing algorithms to eliminate impulse noise by leveraging signal sparsity. However, the inventors of this application have discovered the following deep-seated technical shortcomings in existing technologies, which are often overlooked but are bottlenecks limiting performance improvement:

[0004] The waveform-channel characteristic mismatch problem: Most existing technologies only perform Doppler compensation at the receiver. However, when the Doppler frequency shift is too large, the traditional sinusoidal subcarriers have already experienced severe energy divergence during physical transmission. This waveform destruction at the physical layer is irreversible, and post-transmission compensation at the receiver alone is insufficient to recover the complete signal energy, leading to a sharp deterioration in the signal-to-noise ratio.

[0005] The “sparse structure confusion” problem: Existing sparse denoising algorithms (such as OMP or traditional Bayesian learning) utilize the “sparseness” of the channel and noise, but they treat both as homogeneous sparse variables.

[0006] The inventors noted that underwater acoustic multipath channels typically exhibit "wide-block sparseness" (arriving in clusters with relatively long durations), while underwater impulse noise exhibits "narrow-block sparseness" (sharp bursts with extremely short durations). Because existing algorithms fail to distinguish between the two at this microstructural level, in high-noise environments, they are prone to misclassifying a cluster of strong multipath signals as noise and attempting to eliminate it, or misclassifying continuous impulse noise as channel multipath and attempting to equalize it. This "structural misjudgment" is the fundamental reason why existing systems suffer from persistently high bit error rates in high sea states.

[0007] Therefore, there is an urgent need for an underwater acoustic anti-interference transmission method that can actively adapt to Doppler characteristics at the transmitting end and accurately distinguish heterogeneous sparse components at the receiving end. Summary of the Invention

[0008] To address the aforementioned technical problems in the existing technology, this invention proposes an underwater acoustic anti-interference transmission method and system based on sparse time-frequency feature mapping to solve the above-mentioned technical problems.

[0009] A method for underwater acoustic interference-resistant transmission based on sparse time-frequency feature mapping includes:

[0010] S1: The transmitter obtains the Doppler spread factor of the current underwater acoustic channel and establishes the relationship between the Doppler spread factor and the fractional rotation order. Linear mapping relationship; based on fractional rotation order Construct a discrete fractional Fourier inverse transform matrix, multiply the frequency domain data vector to be transmitted by the discrete fractional Fourier inverse transform matrix, generate a linear frequency modulated underwater acoustic signal with time-frequency shearing characteristics and transmit it.

[0011] S2: The receiver acquires the time-domain mixed signal after transmission through the underwater acoustic channel and constructs a linear observation equation containing the transmission matrix, channel response vector, and environmental burst noise vector. The transmission matrix is ​​synthesized by the discrete fractional Fourier inverse transform matrix and the underwater acoustic channel convolution matrix. The channel response vector is configured to follow a first-class block sparse prior distribution, and the environmental burst noise vector is configured to follow a second-class block sparse prior distribution.

[0012] S3: Run the variational inference process for the linear observation equation. Under the condition of fixing the hyperparameters of the first type of block sparse prior distribution, update the posterior probability distribution of the environmental burst noise vector; under the condition of fixing the hyperparameters of the second type of block sparse prior distribution, update the posterior probability distribution of the channel response vector; alternately execute the update process of the posterior probability distribution until the parameters of the two posterior probability distributions converge, and output the posterior mean of the environmental burst noise vector.

[0013] S4: Perform interference cancellation operation, subtract the posterior mean of the environmental burst noise vector from the time-domain mixed signal to obtain the denoised underwater acoustic signal; use the converged channel response vector to perform fractional-order domain equalization processing on the denoised underwater acoustic signal, and output the transmission data.

[0014] In some specific embodiments, in S1, the Doppler spread factor and the fractional rotation order are established. A linear mapping relationship, specifically satisfying: ,in, The relative velocity, The speed of sound in water, For carrier frequency, The preset scaling factor; the rotation angle of the discrete fractional Fourier inverse transform matrix is ​​determined by... This ensures that the slope of the energy distribution of the generated linear frequency modulated underwater acoustic signal in the time-frequency plane is consistent with the slope of the Doppler frequency shift trajectory of the underwater acoustic channel.

[0015] In some specific embodiments, in S2, the first type of block sparse prior distribution and the second type of block sparse prior distribution are distinguished and defined by the following structural parameters:

[0016] The first type of block sparse prior distribution is determined by the length of the first block. and the first sparsity control parameter Definition, where Time broadening corresponding to underwater acoustic multipath clusters;

[0017] The sparse prior distribution of the second type of block is determined by the length of the second block. Second sparsity control parameter Definition, where The duration corresponding to the underwater burst pulse;

[0018] Set constraints Furthermore, in the variational inference process, all elements within the same block share the same sparsity control parameter.

[0019] In some specific embodiments, in S3, the variational inference process includes specific signal statistical reconstruction operations: in each iteration, channel estimation and noise estimation are performed alternately: based on the posterior mean of the environmental burst noise vector obtained in the previous iteration. Calculate the posterior covariance matrix of the channel response vector for this round. with posterior mean : , Based on the posterior mean of the channel response vector obtained in this round of calculation Calculate the posterior covariance matrix of the environmental burst noise vector in this round. with posterior mean : , ,in, It is the identity matrix. It is a time-domain mixed signal. and These are respectively controlled by the first sparsity parameter Second sparsity control parameter The diagonal matrix formed is used to initialize the posterior mean of the burst noise vector in the environment. It is a vector of all zeros.

[0020] In some specific embodiments, in S3, the variational inference process further includes a specific sparse structure learning operation: utilizing the posterior mean of the channel response vector. and posterior covariance matrix Update the first sparsity control parameter for the diagonal elements. ; Utilizing the posterior mean of the environmental burst noise vector and posterior covariance matrix Update the second sparsity control parameter for the diagonal elements. When the value of any sparsity control parameter exceeds the preset threshold, the sub-block signal component corresponding to the parameter is determined to be zero, and it is removed from the non-zero support set of the corresponding channel response vector or environmental burst noise vector.

[0021] In some specific embodiments, S1 also includes pilot embedding based on sparse time-frequency feature mapping for underwater acoustic anti-interference transmission: before constructing the discrete fractional Fourier inverse transform matrix, a preset fractional pilot sequence is inserted at the zero-frequency position and high-frequency boundary position of the frequency domain data vector to be transmitted; the fractional pilot sequence is represented in the time domain as a Chirp signal with a predetermined frequency modulation slope, which is used by the receiver to perform blind estimation of the Doppler spread factor.

[0022] In some specific embodiments, in S4, performing the interference cancellation operation specifically involves:

[0023] S41: Construct a soft threshold decision function;

[0024] S42: Input the posterior mean of the environmental burst noise vector into the soft threshold decision function;

[0025] S43: Output the posterior mean as the interference estimate only if the magnitude of the posterior mean is greater than 3 times the standard deviation of the background noise; otherwise, output zero.

[0026] S44: Subtract the interference estimate from the time-domain mixed signal.

[0027] According to a second aspect of the invention, a computer-readable storage medium is provided on which one or more computer programs are stored, which, when executed by a computer processor, implement the method described above.

[0028] According to a third aspect of the present invention, an underwater acoustic anti-interference transmission system based on sparse time-frequency feature mapping is proposed, comprising:

[0029] An adaptive fractional-order transmitter, configured to acquire the Doppler spread factor of the underwater acoustic channel, establishes its relationship with the fractional-order rotational order. The linear mapping relationship is included, and a discrete fractional Fourier inverse transform logic circuit is included to generate and transmit a linear frequency modulated underwater acoustic signal with time-frequency shearing characteristics.

[0030] The hybrid acoustic field acquisition and modeling unit is configured to acquire time-domain hybrid signals through an underwater acoustic transducer and construct a linear observation equation containing a transmission matrix, a channel response vector, and an environmental burst noise vector; wherein the channel response vector and the environmental burst noise vector are configured to follow a first-class block sparse prior distribution and a second-class block sparse prior distribution, respectively.

[0031] The dual-channel variational inference engine is configured to run a variational inference process for linear observation equations. It has an internal alternating iteration logic to alternately update the posterior probability distribution of the environmental burst noise vector and the posterior probability distribution of the channel response vector under fixed hyperparameter conditions until convergence and output of the noise posterior mean.

[0032] The signal purification and equalization receiver is configured to perform interference cancellation circuit logic, subtract the posterior mean of the environmental burst noise vector from the time-domain mixed signal, and perform fractional-order domain equalization using the converged channel response vector to output the transmitted data.

[0033] In some specific embodiments, the dual-channel variational inference engine is specifically configured to include a first sparse parameter register and a second sparse parameter register, which are respectively used to store the first block length corresponding to the underwater acoustic multipath cluster. and the second block length corresponding to the underwater burst pulse And it is embedded in the hardware logic. The comparison constraints are used to distinguish between channel features and noise features in iterative calculations.

[0034] Compared with the prior art, the present invention has the following significant advantages:

[0035] This invention breaks through the limitations of traditional OFDM fixed carriers by introducing fractional-order Chirp waveform mapping at the transmitter. This is achieved by establishing a Doppler factor and a fractional-order rotation order. The linear mapping relationship causes the transmitted waveform to produce shear deformation in the time-frequency plane that is consistent with the channel's Doppler trajectory. This design allows the signal energy to remain highly focused at the receiving end after transmission through a high dynamic range channel, physically avoiding energy leakage caused by Doppler spread, making it particularly suitable for high-speed mobile AUV communication scenarios.

[0036] This invention innovatively proposes a "heterogeneous biblock sparse model" that utilizes inequality constraints. (That is, the channel block length is greater than the noise block length) is used as prior knowledge. During variational Bayesian inference, the algorithm can automatically and accurately separate "signal clusters" from "noise spikes" based on the duration characteristics of the signal. This effectively avoids the shortcomings of existing algorithms that "inadvertently delete signals" or "miss noise deletions" under strong interference, and significantly reduces the bit error rate.

[0037] Unlike the step-by-step processing mode of "denoising first, then estimating" in existing technologies (which is prone to error propagation), this invention adopts a dual-channel variational inference mechanism, simultaneously iteratively updating channel parameters and noise parameters within the same probabilistic framework. This joint optimization mechanism enables the two to mutually correct each other using residual information, resulting in faster convergence and stronger robustness to non-Gaussian environments.

[0038] This invention utilizes fractional-order pilots for blind estimation and employs a soft-decision cancellation strategy to handle noise. It does not require additional guard bands or complex analog filter hardware and can be implemented on existing underwater acoustic transducer platforms simply by upgrading digital signal processing algorithms, making it highly valuable for engineering applications. Attached Figure Description

[0039] The accompanying drawings are included to provide a further understanding of the embodiments and are incorporated in and constitute a part of this specification. The drawings illustrate embodiments and, together with the description, serve to explain the principles of the invention. Other embodiments and many anticipated advantages of the embodiments will be readily recognized as they become better understood through reference to the following detailed description. Other features, objects, and advantages of this application will become more apparent from reading the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0040] Figure 1 This is a flowchart of an embodiment of the underwater acoustic anti-interference transmission method based on sparse time-frequency feature mapping of this application;

[0041] Figure 2 This is a time-frequency feature mapping comparison diagram of a specific embodiment of this application;

[0042] Figure 3 This is a framework diagram of an underwater acoustic anti-interference transmission system based on sparse time-frequency feature mapping, according to an embodiment of this application.

[0043] Figure 4 This is a schematic diagram of the structure of a computer system used to implement the electronic device of the present application. Detailed Implementation

[0044] The present application will now be described in further detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and not intended to limit it. Furthermore, it should be noted that, for ease of description, only the parts relevant to the invention are shown in the accompanying drawings.

[0045] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0046] Figure 1A flowchart illustrating an underwater acoustic anti-interference transmission method based on sparse time-frequency feature mapping according to an embodiment of this application is shown. Figure 1 As shown, the method includes the following steps:

[0047] S1: The transmitter obtains the Doppler spread factor of the current underwater acoustic channel and establishes the relationship between the Doppler spread factor and the fractional rotation order. Linear mapping relationship; based on fractional rotation order A discrete fractional Fourier inverse transform matrix is ​​constructed, and the frequency domain data vector to be transmitted is multiplied by the discrete fractional Fourier inverse transform matrix to generate a linear frequency modulated underwater acoustic signal with time-frequency shearing characteristics and then transmitted.

[0048] In a specific embodiment, the Doppler spread factor and the fractional rotation order are established. A linear mapping relationship, specifically satisfying: ,in, The relative velocity, The speed of sound in water, For carrier frequency, The preset scaling factor; the rotation angle of the discrete fractional Fourier inverse transform matrix is ​​determined by... This ensures that the energy distribution slope of the generated linear frequency modulated underwater acoustic signal in the time-frequency plane matches the slope of the Doppler frequency shift trajectory of the underwater acoustic channel. In this embodiment, the transmitter does not transmit a traditional sinusoidal subcarrier, but instead pre-distorts the waveform using the aforementioned mapping relationship. This setting maximizes the matched filter output gain at the receiver, avoiding energy leakage due to parameter mismatch. In a specific high-speed AUV communication scenario, assuming the underwater acoustic communication system operates in a shallow sea environment, the specific parameters are set as follows: underwater acoustic carrier frequency... =2kHz, speed of sound in water =1500m / s, relative velocity of the transmitting and receiving nodes =15m / s (approximately 30 stops, simulating a high-speed AUV), setting a scaling factor. The normalization constant (e.g., take) ,in (Sampling rate). If the sampling rate =48kHz, then the Doppler spread factor At this point, the calculated fractional order of rotation is... Corresponding to a specific rotation angle (e.g.) ). Implementation effect: The transmitter, based on this... The generated Chirp-OFDM signal has a frequency offset slope of approximately 120 Hz in the time-frequency plane. When this signal passes through the Doppler channel caused by motion at 15 m / s, the physical frequency shift generated by the channel "straightens" this slope, allowing the signal at the receiver to return to an orthogonal state, effectively avoiding the problem of traditional OFDM completely failing at this speed.

[0049] Figure 2 A time-frequency feature mapping comparison diagram of a specific embodiment of this application is shown, such as... Figure 2 As shown, the energy of a traditional OFDM signal will dissipate like fog after passing through a high-velocity underwater acoustic channel (e.g., Figure 2 The energy leakage shown in the left-middle figure leads to inter-carrier interference. However, using the method described in this step, the shearing characteristics of the transmitted signal precisely cancel out the physical shearing effect of the channel, allowing the energy to be refocused within the optimal fractional-order domain when the signal reaches the receiver (e.g., ...). Figure 2 As shown in the middle right figure, this maximizes the signal-to-noise ratio at the physical level, laying a high-quality waveform foundation for subsequent signal processing.

[0050] In a specific embodiment, the method further includes pilot embedding based on sparse time-frequency feature mapping for underwater acoustic anti-interference transmission: before constructing the discrete fractional-order Fourier inverse transform matrix, preset fractional-order pilot sequences are inserted at the zero-frequency position and high-frequency boundary position of the frequency domain data vector to be transmitted, respectively. The fractional-order pilot sequences are represented in the time domain as chirp signals with a predetermined frequency modulation slope, which are used by the receiver to perform blind estimation of the Doppler spread factor. By inserting fractional-order pilots at the zero-frequency and high-frequency boundaries, and utilizing the autocorrelation characteristics of the chirp signal for blind detection, a safe mode is constructed when the feedback link fails. Even if the underwater acoustic feedback channel is interrupted due to deep fading, the receiver can still independently estimate the Doppler factor and perform demodulation, ensuring the survivability of the communication link.

[0051] In a specific embodiment, the transformation matrix is ​​constructed directly using a kernel function based on the sampled discrete fractional Fourier transform. Assuming the data length is The matrix of the first Line number Column elements The calculation formula is as follows (after normalization): ,in, For rotation angle, , Represents the imaginary unit. Represented by natural constant An exponential function with base 0.

[0052] S2: The receiver acquires the time-domain mixed signal transmitted through the underwater acoustic channel and constructs a linear observation equation including the transmission matrix, channel response vector, and environmental burst noise vector. The transmission matrix is ​​synthesized from the discrete fractional Fourier inverse transform matrix and the underwater acoustic channel convolution matrix. The channel response vector is configured to follow a first-type block sparse prior distribution, and the environmental burst noise vector is configured to follow a second-type block sparse prior distribution. The core of this step is to solve the problem of "signal and noise structure confusion" in underwater acoustic environments. Existing mathematical models often assume that noise is an isolated point or Gaussian white noise, but in shallow sea environments, for example, the noise generated by pterodactyls is a continuous burst of pulses with a certain duration.

[0053] In a specific embodiment, the first type of block sparse prior distribution and the second type of block sparse prior distribution are distinguished and defined by the following structural parameters:

[0054] The first type of block sparse prior distribution is determined by the length of the first block. and the first sparsity control parameter Definition, where Time broadening corresponding to underwater acoustic multipath clusters;

[0055] The sparse prior distribution of the second type of block is determined by the length of the second block. Second sparsity control parameter Definition, where The duration corresponding to the underwater burst pulse;

[0056] Set constraints Furthermore, in the variational inference process, all elements within the same block share the same sparsity control parameter. This setting solidifies the underwater acoustic physical characteristics at the probabilistic model level (the delay spread of multipath clusters is much larger than the pulse width of a single puffball noise). It provides robust structure discrimination capabilities, preventing the algorithm from misclassifying continuous strong pulses as channel paths or weak channel multipaths as noise under low signal-to-noise ratio conditions, thus ensuring the system's stability in harsh sea conditions. In a specific application example, targeting a shallow sea environment with significant puffball noise interference, the system sampling rate is set to... =20kHz (i.e., sampling interval) =0.05ms), Type I block sparsity parameter (channel): The multipath delay spread of underwater acoustic channels is typically between 5ms and 10ms, and exhibits a clustered distribution. Therefore, the channel block length is set to... =20 sampling points (corresponding to a cluster width of 1ms). This means the algorithm considers channel energy to appear in clusters of 1ms. Second type of block sparsity parameter (noise): The pulse duration generated by the collapse of a single cavitation bubble is extremely short, typically between 0.1ms and 0.2ms. Therefore, the noise block length is set... =4 sampling points (corresponding to a pulse width of 0.2ms). Satisfies... (20) (4) Constraints. During variational inference, if a high-energy spike lasting 0.2 ms appears in the received signal, due to its length matching... And much smaller It will be classified as "noise" and categorized as such. Vector; if an energy fluctuation lasting 1ms occurs, the algorithm will classify it as a "channel" and categorize it. Vectors. This physical-scale-based automatic classification enables the system to accurately recover signals even with signal-to-interference ratios as low as -5dB.

[0057] S3: Run the variational inference process for the linear observation equation. Under the condition of fixing the hyperparameters of the first type of block sparse prior distribution, update the posterior probability distribution of the environmental burst noise vector; under the condition of fixing the hyperparameters of the second type of block sparse prior distribution, update the posterior probability distribution of the channel response vector; alternately execute the above update process until the parameters of the two posterior probability distributions converge, and output the posterior mean of the environmental burst noise vector.

[0058] In a specific embodiment, the variational inference process includes specific signal statistical reconstruction operations: in each iteration, channel estimation and noise estimation are performed alternately: based on the posterior mean of the environmental burst noise vector obtained in the previous iteration. Calculate the posterior covariance matrix of the channel response vector for this round. with posterior mean : , Based on the posterior mean of the channel response vector obtained in this round of calculation Calculate the posterior covariance matrix of the environmental burst noise vector in this round. with posterior mean : , ,in, It is the identity matrix. It is a time-domain mixed signal. and These are respectively controlled by the first sparsity parameter Second sparsity control parameter The diagonal matrix formed is used to initialize the posterior mean of the burst noise vector in the environment. It is a vector of all zeros.

[0059] In a specific embodiment, the variational inference process also includes a specific sparse structure learning operation: utilizing the posterior mean of the channel response vector. and posterior covariance matrix Update the first sparsity control parameter for the diagonal elements. ; Utilizing the posterior mean of the environmental burst noise vector and posterior covariance matrix Update the second sparsity control parameter for the diagonal elements. When the value of any sparsity control parameter exceeds the preset threshold, the sub-block signal component corresponding to the parameter is determined to be zero, and it is removed from the non-zero support set of the corresponding channel response vector or environmental burst noise vector.

[0060] This embodiment employs an alternating minimization strategy to decouple the complex joint inference into two low-dimensional sub-problems. The specific process is as follows: Set the iteration count. Initialize the posterior mean of the burst noise vector in the environment. It is a vector of all zeros (i.e., assuming no noise interference initially).

[0061] Step A (Update Channel Estimation): Use the noise estimate obtained in the previous round. From the total observed signal Subtracting this noise component from the data yields the "net channel observation": , obtained based on Calculate the current number Channel posterior mean of the round Covariance : , Using the channel estimate just calculated From the total observed signal Subtracting the channel component from the result yields the "noise residual observation": ,based on Calculate the current number posterior mean of wheel noise Covariance : , ,in, This represents the reciprocal of the variance of the background Gaussian white noise (i.e., the noise accuracy). Represents the transmission matrix; Representing a matrix Conjugate transpose.

[0062] Step B (Update Noise Estimation): Utilize the first... Update the sparsity control parameters using the mean and variance calculated in each round: ,in, These represent the element indices of a vector or matrix, respectively. Used for channel vectors Used for noise vectors; Indicates the updated number Channel sparsity control parameters; Represents the channel posterior mean vector The first in One element; Represents the channel posterior covariance matrix The first on the main diagonal One element; Indicates the updated number One noise sparsity control parameter; Represents the posterior mean vector of the noise The first in One element; Represents the posterior covariance matrix of the noise The first on the main diagonal Each element.

[0063] Check if the parameters converge. If not, let... Return to step A.

[0064] S4: Perform interference cancellation operation, subtract the posterior mean of the environmental burst noise vector from the time-domain mixed signal to obtain the denoised underwater acoustic signal; use the converged channel response vector to perform fractional-order domain equalization processing on the denoised underwater acoustic signal, and output the transmission data.

[0065] In a specific embodiment, the interference cancellation operation is performed as follows:

[0066] S41: Construct a soft threshold decision function;

[0067] S42: Input the posterior mean of the environmental burst noise vector into the soft threshold decision function;

[0068] S43: Output the posterior mean as the interference estimate only if the magnitude of the posterior mean is greater than 3 times the standard deviation of the background noise; otherwise, output zero.

[0069] S44: Subtract the interference estimate from the time-domain mixed signal. Compared to traditional hard cut-off (direct zeroing), soft decision preserves the weak signal components superimposed on the noise, maintains the continuity of the signal waveform, and reduces spectral regeneration interference caused by nonlinear operation.

[0070] In a specific embodiment, a standard piecewise shrinkage function is used as the soft threshold operator. For the input noise estimate and threshold The calculation logic is as follows: Among them, threshold Set to 3 times the amplitude of the background noise. This function directly sets the background noise with small amplitude to zero, while retaining the burst pulses with significant amplitude and scaling them down to ensure the smoothness of the signal waveform.

[0071] Figure 3 A framework diagram of an underwater acoustic anti-interference transmission system based on sparse time-frequency feature mapping according to an embodiment of this application is shown, as follows: Figure 3 As shown, the system mainly includes an adaptive fractional-order transmitter 301, a hybrid acoustic field acquisition and modeling unit 302, a dual-channel variational inference engine 303, and a signal purification and equalization receiver 304. The adaptive fractional-order transmitter 301 is configured to acquire the Doppler spread factor of the underwater acoustic channel and establish its relationship with the fractional-order rotational order. The system establishes a linear mapping relationship and includes discrete fractional Fourier inverse transform logic circuitry to generate and transmit linear frequency modulated underwater acoustic signals with time-frequency shearing characteristics. A hybrid acoustic field acquisition and modeling unit 302 is configured to acquire the time-domain hybrid signal via an underwater acoustic transducer and construct a linear observation equation containing a transmission matrix, channel response vector, and environmental burst noise vector. The channel response vector and environmental burst noise vector are configured to follow a first-class block sparse prior distribution and a second-class block sparse prior distribution, respectively. A dual-channel variational inference engine 303 is configured to run a variational inference process on the linear observation equation. It contains alternating iteration logic to alternately update the posterior probability distribution of the environmental burst noise vector and the posterior probability distribution of the channel response vector under fixed hyperparameter conditions until convergence and the posterior mean of the noise is output. A signal purification and equalization receiver 304 is configured to execute interference cancellation circuit logic, subtract the posterior mean of the environmental burst noise vector from the time-domain hybrid signal, and perform fractional domain equalization using the converged channel response vector to output transmitted data. The system can operate as described above. Figure 1 The specific methods in the text.

[0072] In a specific embodiment, the dual-channel variational inference engine 303 is specifically configured to include a first sparse parameter register and a second sparse parameter register, which are respectively used to store the first block length corresponding to the underwater acoustic multipath cluster. and the second block length corresponding to the underwater burst pulse And it is embedded in the hardware logic. The comparison constraints are used to distinguish between channel features and noise features in iterative calculations.

[0073] This application proposes an underwater acoustic anti-interference transmission method and system based on sparse time-frequency feature mapping, constructing an underwater acoustic communication architecture integrating "physical layer waveform adaptation" and "signal layer interference joint separation". At the transmitting end, by constructing a mapping mechanism between the Doppler factor and the fractional-order rotation order, the large-scale frequency shift of the underwater acoustic channel is actively canceled by utilizing the time-frequency shearing characteristics of the waveform, thus achieving physical focusing of signal energy. At the receiving end, this invention overcomes the bottleneck of traditional sparse algorithms being unable to distinguish between the channel and noise, and innovatively utilizes the structural differences between "wide-block multipath" and "narrow-block noise" at the time-domain physical scale, achieving accurate decoupling and removal of interference components in the mixed sound field through a dual-channel variational inference algorithm. This scheme effectively overcomes the dual technical barriers of high-dynamic Doppler diffusion and strong burst impulse noise in underwater acoustics without adding additional analog filtering hardware, significantly improving the robustness and spectrum utilization of the underwater communication system under harsh sea conditions.

[0074] The following is for reference. Figure 4 It shows a schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application. Figure 4 The electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0075] like Figure 4 As shown, the computer system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes based on programs stored in read-only memory (ROM) 402 or programs loaded into random access memory (RAM) 403 from storage section 408. RAM 403 also stores various programs and data required for system operation. The CPU 401, ROM 402, and RAM 403 are interconnected via bus 404. Input / output (I / O) interface 405 is also connected to bus 404.

[0076] The following components are connected to I / O interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a liquid crystal display (LCD) and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card and a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. Drive 410 is also connected to I / O interface 405 as needed. Removable media 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 410 as needed so that computer programs read from them can be installed into storage section 408 as needed.

[0077] Specifically, according to embodiments of this disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this disclosure include a computer program product comprising a computer program carried on a computer-readable storage medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit (CPU) 401, it performs the functions defined in the methods of this application. It should be noted that the computer-readable storage medium of this application can be a computer-readable signal medium or a computer-readable storage medium or any combination thereof. The computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in connection with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such propagated data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable storage medium other than a computer-readable storage medium that can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. Program code contained on a computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wireless, wire, optical fiber, RF, etc., or any suitable combination thereof.

[0078] Computer program code for performing the operations of this application can be written in one or more programming languages ​​or a combination thereof. Programming languages ​​include object-oriented programming languages—such as Java, Smalltalk, and C++—as well as conventional procedural programming languages—such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0079] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0080] The modules described in the embodiments of this application can be implemented in software or in hardware.

[0081] On the other hand, this application also provides a computer-readable storage medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to: acquire the Doppler spread factor of the current underwater acoustic channel at the transmitting end, and establish the Doppler spread factor and the fractional-order rotation order. Linear mapping relationship; based on fractional rotation order A discrete fractional-order inverse Fourier transform (IFFT) matrix is ​​constructed. The frequency domain data vector to be transmitted is multiplied by the IFFT matrix to generate a linear frequency-modulated underwater acoustic signal with time-frequency shearing characteristics, which is then transmitted. The receiving end acquires the time-domain mixed signal after transmission through the underwater acoustic channel and constructs a linear observation equation containing the transmission matrix, channel response vector, and environmental burst noise vector. The transmission matrix is ​​synthesized from the IFFT matrix and the underwater acoustic channel convolution matrix. The channel response vector is configured to follow a first-class block sparse prior distribution, and the environmental burst noise vector is configured to follow a second-class block sparse prior distribution. The linear observation equation is then applied... The variational inference process involves updating the posterior probability distribution of the environmental burst noise vector under the condition of fixed hyperparameters of the first type of block sparse prior distribution; updating the posterior probability distribution of the channel response vector under the condition of fixed hyperparameters of the second type of block sparse prior distribution; alternatingly executing the above update process until the parameters of the two posterior probability distributions converge, and outputting the posterior mean of the environmental burst noise vector; performing interference cancellation operation, subtracting the posterior mean of the environmental burst noise vector from the time-domain mixed signal to obtain the denoised underwater acoustic signal; and using the converged channel response vector to perform fractional-order domain equalization processing on the denoised underwater acoustic signal, outputting the transmitted data.

[0082] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A method for underwater acoustic anti-interference transmission based on sparse time-frequency feature mapping, characterized in that, include: S1: The transmitter obtains the Doppler spread factor of the current underwater acoustic channel and establishes the relationship between the Doppler spread factor and the fractional rotation order. The linear mapping relationship; based on the fractional order rotation order Construct a discrete fractional Fourier inverse transform matrix, multiply the frequency domain data vector to be transmitted by the discrete fractional Fourier inverse transform matrix, generate a linear frequency modulated underwater acoustic signal with time-frequency shearing characteristics, and transmit it. S2: The receiving end acquires the time-domain mixed signal after transmission through the underwater acoustic channel and constructs a linear observation equation containing a transmission matrix, a channel response vector, and an environmental burst noise vector; wherein, the transmission matrix is ​​synthesized by the discrete fractional Fourier inverse transform matrix and the underwater acoustic channel convolution matrix; the channel response vector is configured to follow a first type of block sparse prior distribution, and the environmental burst noise vector is configured to follow a second type of block sparse prior distribution; S3: Perform variational inference on the linear observation equation. Under the condition of fixing the hyperparameters of the first type of block sparse prior distribution, update the posterior probability distribution of the environmental burst noise vector; under the condition of fixing the hyperparameters of the second type of block sparse prior distribution, update the posterior probability distribution of the channel response vector; alternately execute the posterior probability distribution update process until the parameters of the two posterior probability distributions converge, and output the posterior mean of the environmental burst noise vector; and S4: Perform interference cancellation operation by subtracting the posterior mean of the environmental burst noise vector from the time-domain mixed signal to obtain the denoised underwater acoustic signal; perform fractional-order domain equalization processing on the denoised underwater acoustic signal using the converged channel response vector, and output the transmission data. In S1, the establishment of the Doppler expansion factor and the fractional rotation order... A linear mapping relationship, specifically satisfying: ,in, The relative velocity, The speed of sound in water, For carrier frequency, The preset scaling factor; the rotation angle of the discrete fractional Fourier inverse transform matrix is ​​determined by... This is determined so that the slope of the energy distribution of the generated linear frequency modulated underwater acoustic signal in the time-frequency plane is consistent with the slope of the Doppler frequency shift trajectory of the underwater acoustic channel; In S2, the first type of block sparse prior distribution and the second type of block sparse prior distribution are distinguished and limited by the following structural parameters: The first type of block sparse prior distribution is determined by the first block length. and the first sparsity control parameter Definition, where Time broadening corresponding to underwater acoustic multipath clusters; The second type of block sparse prior distribution is determined by the second block length. Second sparsity control parameter Definition, where The duration corresponding to the underwater burst pulse; Set constraints Furthermore, in the variational inference process, all elements within the same block share the same sparsity control parameter. In S3, the variational inference process includes specific signal statistical reconstruction operations: in each iteration, channel estimation and noise estimation are performed alternately: based on the posterior mean of the environmental burst noise vector obtained in the previous iteration. Calculate the posterior covariance matrix of the channel response vector for this round. with posterior mean : , Based on the posterior mean of the channel response vector obtained in this round of calculation Calculate the posterior covariance matrix of the environmental burst noise vector in this round. with posterior mean : , ,in, It is the identity matrix. It is a time-domain mixed signal. This represents the reciprocal of the variance of the background Gaussian white noise, i.e., the noise accuracy. Represents the transmission matrix. Representing a matrix Conjugate transpose and These are respectively controlled by the first sparsity parameter Second sparsity control parameter The diagonal matrix formed is used to initialize the posterior mean of the burst noise vector in the environment. It is a vector of all zeros.

2. The underwater acoustic anti-interference transmission method based on sparse time-frequency feature mapping according to claim 1, characterized in that, In S3, the variational inference process also includes specific sparse structure learning operations: utilizing the posterior mean of the channel response vector. and posterior covariance matrix Update the first sparsity control parameter for the diagonal elements. ; Utilizing the posterior mean of the environmental burst noise vector and posterior covariance matrix Update the second sparsity control parameter for the diagonal elements. ; When the value of any sparsity control parameter exceeds the preset threshold, the sub-block signal component corresponding to that parameter is determined to be zero, and it is removed from the non-zero support set of the corresponding channel response vector or environmental burst noise vector.

3. The underwater acoustic anti-interference transmission method based on sparse time-frequency feature mapping according to claim 1, characterized in that, S1 also includes underwater acoustic anti-interference transmission based on pilot embedding and sparse time-frequency feature mapping: before constructing the discrete fractional Fourier inverse transform matrix, a preset fractional pilot sequence is inserted at the zero-frequency position and high-frequency boundary position of the frequency domain data vector to be transmitted; the fractional pilot sequence is represented in the time domain as a Chirp signal with a predetermined frequency modulation slope, which is used by the receiver to perform blind estimation of the Doppler spread factor.

4. The underwater acoustic anti-interference transmission method based on sparse time-frequency feature mapping according to claim 1, characterized in that, In step S4, the interference cancellation operation specifically involves: S41: Construct a soft threshold decision function; S42: Input the posterior mean of the environmental burst noise vector into the soft threshold decision function; S43: Output the posterior mean as the interference estimate only if the magnitude of the posterior mean is greater than 3 times the standard deviation of the background noise; otherwise, output zero. S44: Subtract the interference estimate from the time-domain mixed signal.

5. A computer-readable storage medium having one or more computer programs stored thereon, characterized in that, When the one or more computer programs are executed by a computer processor, they perform the method according to any one of claims 1-4.

6. A water acoustic anti-interference transmission system based on sparse time-frequency feature mapping, characterized in that, include: An adaptive fractional-order transmitter, configured to acquire the Doppler spread factor of the underwater acoustic channel, establishes its relationship with the fractional-order rotational order. The linear mapping relationship is included, and a discrete fractional Fourier inverse transform logic circuit is included to generate and transmit a linear frequency modulated underwater acoustic signal with time-frequency shearing characteristics. The hybrid acoustic field acquisition and modeling unit is configured to acquire time-domain hybrid signals through an underwater acoustic transducer and construct a linear observation equation containing a transmission matrix, a channel response vector, and an environmental burst noise vector; wherein the channel response vector and the environmental burst noise vector are configured to follow a first-type block sparse prior distribution and a second-type block sparse prior distribution, respectively. The dual-channel variational inference engine is configured to run a variational inference process for the linear observation equation. It has an internal alternating iteration logic to alternately update the posterior probability distribution of the environmental burst noise vector and the posterior probability distribution of the channel response vector under fixed hyperparameter conditions until convergence and output of the noise posterior mean. The signal purification and equalization receiver is configured to execute interference cancellation circuit logic, subtract the posterior mean of the environmental burst noise vector from the time-domain mixed signal, and perform fractional-order domain equalization using the converged channel response vector to output transmitted data. In the adaptive fractional-order transmitter, the Doppler spread factor and the fractional-order rotation order are established. A linear mapping relationship, specifically satisfying: ,in, The relative velocity, The speed of sound in water, For carrier frequency, The preset scaling factor; the rotation angle of the discrete fractional Fourier inverse transform matrix is ​​determined by... This is determined so that the slope of the energy distribution of the generated linear frequency modulated underwater acoustic signal in the time-frequency plane is consistent with the slope of the Doppler frequency shift trajectory of the underwater acoustic channel; In the hybrid sound field acquisition and modeling unit, the first type of block sparse prior distribution and the second type of block sparse prior distribution are distinguished and limited by the following structural parameters: The first type of block sparse prior distribution is determined by the first block length. and the first sparsity control parameter Definition, where Time broadening corresponding to underwater acoustic multipath clusters; The second type of block sparse prior distribution is determined by the second block length. Second sparsity control parameter Definition, where The duration corresponding to the underwater burst pulse; Set constraints Furthermore, in the variational inference process, all elements within the same block share the same sparsity control parameter. In the dual-channel variational inference engine, the variational inference process includes specific signal statistical reconstruction operations: in each iteration, channel estimation and noise estimation are performed alternately: based on the posterior mean of the environmental burst noise vector obtained in the previous iteration. Calculate the posterior covariance matrix of the channel response vector for this round. with posterior mean : , Based on the posterior mean of the channel response vector obtained in this round of calculation Calculate the posterior covariance matrix of the environmental burst noise vector in this round. with posterior mean : , ,in, It is the identity matrix. It is a time-domain mixed signal. This represents the reciprocal of the variance of the background Gaussian white noise, i.e., the noise accuracy. Represents the transmission matrix. Representing a matrix Conjugate transpose and These are respectively controlled by the first sparsity parameter Second sparsity control parameter The diagonal matrix formed is used to initialize the posterior mean of the burst noise vector in the environment. It is a vector of all zeros.

7. The underwater acoustic anti-interference transmission system based on sparse time-frequency feature mapping according to claim 6, characterized in that, The dual-channel variational inference engine is specifically configured to include a first sparse parameter register and a second sparse parameter register, which are used to store the first block length corresponding to the underwater acoustic multipath cluster, respectively. and the second block length corresponding to the underwater burst pulse And it is embedded in the hardware logic. The comparison constraints are used to distinguish between channel features and noise features in iterative calculations.

Images