An underwater acoustic communication receiving method based on superimposed interference signals

Abstract

The present application relates to the technical field of underwater acoustic communication, and particularly relates to an underwater acoustic communication receiving method based on superimposed interference signals.According to prior information of the interference signals, a time-domain signal reconstruction scheme based on a fast soft threshold iteration algorithm (FISTA) is proposed, which has higher reconstruction accuracy for sonar interference signals; in view of the influence of residual sonar interference signals, multipath effects and environmental noise, a VTRM equalization method based on FISTA sparse estimation and a non-parametric Viterbi decoding algorithm based on a Gaussian kernel are proposed, which can well solve the influence of multipath effects and environmental noise, and thus effectively improve the performance and signal receiving accuracy of underwater acoustic communication.

Description

Technical Field

[0001] This invention relates to the field of underwater acoustic communication technology, and more specifically to an underwater acoustic communication receiving method based on superimposed interference signals. Background Technology

[0002] Traditional superposition communication receiver technology employs a time-domain orthogonal superposition of communication and sonar signals. The received superimposed signals are processed without any signal separation methods, and the receiver then demodulates and makes a decision. While traditional methods attempt to avoid interference between signals by using orthogonal signals, severe interference between the superimposed signals still prevents the receiver from resolving the signal.

[0003] The Viterbi Decoder algorithm is a maximum likelihood decoding algorithm based on a convolutional code trellis graph. Its process can be divided into two directions: forward and backtracking. The backtracking method searches for the maximum likelihood path in the trellis graph using a path metric (PM), calculating each possible path and storing them as branch metrics (BMs), which are then summed to obtain the PM. To ensure that only one optimal path exists, only the path with the minimum PM value is retained and passed on in each iteration.

[0004] In Viterbi decoding algorithms, the performance of the system is significantly affected by the severe multipath effects and non-Gaussian noise characteristics of the ocean channel. Therefore, signal processing for multipath equalization and non-Gaussian noise processing are essential steps for the receiver. The accuracy of channel estimation is directly related to sonar signal reconstruction and the communication system's resistance to multipath fading. To improve the communication system's resistance to multipath fading, the receiver employs channel estimation and equalization techniques to enhance the signal-to-noise ratio (SNR) of the transmitting channel. Simultaneously, various impulse noise suppression techniques are proposed to suppress ocean impulse noise and improve the SNR of the communication system.

[0005] Marine environmental noise is predominantly non-Gaussian noise, widely present in the ocean, especially in shallow waters. Because the ocean is a complex scattering channel, marine environmental noise sources are numerous and extremely complex. Most noise environments (including wind and rain noise, shipping noise, biological noise, seismic noise, industrial noise, noise from biological activities such as krill, and various geophysical sources such as Arctic ice and seismic activity) result in a severe non-Gaussian nature of the overall noise, primarily consisting of impulse noise, characterized by its short duration, suddenness, and high amplitude. Compared to traditional Gaussian noise, impulse noise has energy that is almost tens of times greater. Furthermore, the probability density function of non-Gaussian noise is extremely complex, exhibiting a severe heavy tail effect.

[0006] Channel estimation and equalization are crucial. Channel estimation is a technique that estimates the parameters of a channel, such as a wireless channel or an underwater acoustic channel, based on known transmitted and received data. The accuracy of channel estimation directly affects the performance of the entire system. Channel equalization is a fading mitigation method that improves the transmission performance of communication systems in fading channels. It uses known channel estimation results to reduce the impact of multipath delay in underwater acoustic communication, thus compensating for system characteristics.

[0007] Existing underwater acoustic channel estimation typically employs conventional least squares (LS) channel estimation. The LS algorithm relies on the characteristics of newly received channels, but suffers from poor estimation accuracy and susceptibility to noise. Summary of the Invention

[0008] The technical problem to be solved by this invention is to propose an underwater acoustic communication receiving method based on superimposed interference signals, which can solve the interference problem between superimposed interference signals and reduce the impact of ocean multipath delay and non-Gaussian noise on the receiver, thereby effectively improving the performance of underwater acoustic communication and signal reception accuracy.

[0009] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows:

[0010] A method for receiving underwater acoustic communication signals based on superimposed interference signals includes the following steps:

[0011] S101, the bit data is channel-encoded by convolutional code and then digitally modulated by BPSK and carrier to obtain the communication signal, and generates a sonar signal with the same frequency.

[0012] S102, the communication signal and the sonar signal are superimposed, and a synchronization head is added for signal synchronization. At this time, the signal is subjected to marine environmental noise through the actual underwater acoustic channel.

[0013] S103, perform sonar interference signal cancellation on the received signal sequence;

[0014] S104. Based on the channel estimation results, a time-reversal mirror is used to perform VTRM channel equalization on the signal after sonar interference cancellation in order to enhance the energy of the main path.

[0015] S105 employs a Gaussian kernel-based soft-limiting impulse noise suppression algorithm combined with a Viterbi decoder to improve the signal-to-noise ratio of the receiving system.

[0016] Furthermore, in step S101, the ocean channel and environmental noise in the target water area are first collected to establish an environmental noise model and a channel model under underwater acoustic communication.

[0017] The generated bit data is then encoded using a (2, 1, 3) convolutional code and modulated using BPSK digital modulation to generate a BPSK communication signal. This signal is then carrier-modulated and superimposed with a sonar signal of the same frequency to serve as the transmitting end data.

[0018] Furthermore, the BPSK communication signal, after pulse shaping, is superimposed on the sonar's LMF signal, and its mathematical formula is expressed as follows:

[0019] s(n)=s b (n)+s c (n)

[0020] Wherein, [S(0),S(1),…,S(N-1)] T For the transmitted time-domain superimposed signal, S b =[S b (0),S b (1),...,S b (N-1)] T For the BPSK communication signal after convolutional coding, S c =[S c (0),S c (1),…,S c (N-1)] T This is the sonar chirp signal being transmitted.

[0021] Furthermore, in step S102, the transmitting signal is processed by the generated BELLHOP underwater acoustic channel and Middleton Class A impulse noise sequence to obtain the received signal.

[0022] Furthermore, in step S103, the sonar interference signal cancellation of the received signal sequence includes:

[0023] First, the synchronization header is extracted from the receiver signal. Then, the FISTA algorithm is used to estimate the BELLHOP underwater acoustic channel value and convolves it with the sonar signal to obtain the reconstructed sonar interference signal. Finally, the reconstructed sonar interference signal is subtracted from the received data.

[0024] Furthermore, the FISTA algorithm's process steps include:

[0025] 1) Initialize q (0) =0,y (1) =q (0) ,t (1) =1;

[0026] 2)

[0027] 3)

[0028] 4)y (n+1) =q (n) +((t (k) -1) / t (k+1) )(q (n) -q (n-1) );

[0029] Where, ρ + For the Euclidean projection of the non-negative quadrant. For the descent gradient, L is the Lipschose continuous gradient operator, numerically equal to A. T The largest eigenvalue of A.

[0030] Furthermore, in step S104, the step of using a time-reversal mirror to perform VTRM channel equalization on the signal after sonar interference cancellation includes:

[0031] S41, Signal reception: Receive the signal after sonar interference cancellation and perform preprocessing;

[0032] S42, Time Reversal: The received signal is time-reversed to compensate for distortion caused by the channel;

[0033] S43, Spectrum Equalization: Performs spectrum equalization on the inverted signal to enhance the energy of the main path of the signal;

[0034] S44, Demodulation and Data Recovery: Demodulate and recover the data after equalization to obtain the processed data transmission result.

[0035] Furthermore, in step S105, the impulse noise suppression algorithm based on Gaussian kernel soft limiting includes:

[0036] S51, Preprocessing: Perform filtering preprocessing on the data transmission results;

[0037] S52, Gaussian kernel smoothing: The preprocessed signal is smoothed using a Gaussian kernel, specifically by convolving the Gaussian kernel with the preprocessed signal.

[0038] S53, Wavelet Transform and Soft Limiting: Perform wavelet transform on the smoothed signal to obtain wavelet coefficients of different scales, and remove noise components from the wavelet coefficients by setting a preset threshold.

[0039] S54, Inverse Wavelet Transform: Perform inverse wavelet transform on the processed wavelet coefficients to obtain the denoised signal.

[0040] Compared with the prior art, the present invention has the following main advantages:

[0041] Based on prior information about the interference signal, this invention proposes a sparse channel time-domain signal reconstruction scheme based on the Fast Soft Threshold Iteration Algorithm (FISTA), which has higher reconstruction accuracy for sonar interference signals. To address the effects of residual sonar interference, multipath effects, and environmental noise, this invention proposes a VTRM equalization method based on FISTA sparse estimation and a nonparametric Viterbi decoding algorithm based on Gaussian kernels. These methods effectively solve the problems caused by multipath effects and environmental noise, thereby significantly improving the performance and signal reception accuracy of underwater acoustic communication. Attached Figure Description

[0042] Figure 1 This is a block diagram of an underwater acoustic communication receiver system based on superimposed interference signals in an embodiment of the present invention;

[0043] Figure 2 This is a block diagram illustrating the specific principle of the sonar cancellation scheme based on the FISTA sparse channel estimation algorithm in this embodiment of the invention.

[0044] Figure 3 This is a schematic diagram of the impulse response of the BELLHOP underwater acoustic channel in an embodiment of the present invention;

[0045] Figure 4 This is a schematic diagram of the underwater acoustic channel estimation results based on matched correlation estimation in an embodiment of the present invention;

[0046] Figure 5 This is a schematic diagram of the underwater acoustic channel estimation results based on the FISTA sparse algorithm in an embodiment of the present invention;

[0047] Figure 6 This is a block diagram of the VTRM equalization principle based on FISTA sparse estimation in an embodiment of the present invention;

[0048] Figure 7 This is a block diagram of the Viterbi decoding algorithm based on Gaussian kernel soft limiting in an embodiment of the present invention. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention. Furthermore, the technical features involved in the various embodiments of this invention described below can be combined with each other as long as they do not conflict with each other.

[0050] It should be noted that, depending on the implementation needs, the various steps / components described in this application can be broken down into more steps / components, or two or more steps / components or parts of the operation of steps / components can be combined into new steps / components to achieve the purpose of this invention.

[0051] Example 1: This example provides a method for receiving underwater acoustic communication based on superimposed interference signals, mainly including:

[0052] 1) To address the serious problem of self-interference from superimposed signals, based on prior information about the interference signal and combined with the full-duplex digital domain self-interference elimination theory and compressed sensing theory, a sparse channel time-domain signal reconstruction scheme based on the Fast Soft Threshold Iteration Algorithm (FISTA) is proposed, which eliminates the interference of sonar signals.

[0053] 2) To address the impact of residual sonar interference signals, multipath effects in underwater acoustic channels, and non-Gaussian noise, a VTRM equalization scheme based on FISTA sparse estimation and a non-parametric Viterbi decoding algorithm based on Gaussian kernels are proposed. The effectiveness of sonar cancellation is measured by the reliability index of the communication system, considering both cases with strong and weak non-Gaussianity.

[0054] To verify the proposed technology, the following sea trials and simulation experiments were conducted:

[0055] S1: Collect ocean channel and environmental noise data of the target water area, and establish an environmental noise model and channel model for underwater acoustic communication;

[0056] S2: Encode the generated bit data using a (2, 1, 3) convolutional code, and use BPSK modulation to generate a BPSK signal. Then, perform carrier modulation and superimpose a sonar signal of the same frequency as the transmitting end data.

[0057] S3: The transmitted data passes through the real ocean channel and ocean noise to obtain the received data. After extraction and signal processing such as sonar interference signal separation, VTRM channel equalization and impulse noise suppression, the bit data before decision is obtained.

[0058] S4: The data that has undergone signal processing is demodulated, detected, and Viterbi decoded to recover the bit data, thereby calculating the bit error rate.

[0059] In step S1, the generated BPSK communication signal is pulse-shaped and then superimposed with the sonar LMF signal. The mathematical formula is as follows:

[0060] s(n)=s b (n)+s c (n)(1)

[0061] Wherein, [S(0),S(1),…,S(N-1)] T For the transmitted time-domain superimposed signal, S b =[S b (0),S b (1),…,Sb (N-1)] T For the BPSK communication signal after convolutional coding, S c =[S c (0),S c (1),…,S c (N-1)] T The received signal after the transmitted sonar chirp signal passes through the underwater acoustic channel is y = [y(0), y(1), ..., y(N-1)]. T It can be represented as:

[0062]

[0063] Where y is the signal received by the receiver.

[0064]

[0065] Where h is the time-domain impulse response of the underwater acoustic channel, and M is the channel impulse response length. i Here, h is the channel tap coefficient, and τ is the delay spread, because the tap coefficient h i The non-zero values ​​are far less than m, therefore the impulse response is said to be sparse with a sparsity of M, and is randomly distributed. Let represent the convolution symbol, and w represent the marine environmental noise. The above formula can be expressed in the form of a cyclic matrix-vector:

[0066] y = X·h + w(3)

[0067] Here, X is a cyclic matrix with a Toeplitz structure transformed from the pilot signal to be estimated (insufficient dimensions are padded with zeros), and y and w are M+L-1 dimensional data sequences. At this point, all formula operations are performed in the time domain, and matrix X is a square matrix of dimension M+L-1.

[0068]

[0069] In step S3, the specific principles of the proposed sonar cancellation scheme are detailed in the appendix. Figure 2 As shown, it illustrates the reconstruction block diagram based on time-domain sparse multipath channel signals. The channel estimate is obtained through the proposed FISTA-based time-domain sparse multipath channel reconstruction algorithm. The known sonar signal is convolved with this estimate to generate the reconstructed sonar interference signal, S. I It can be represented as:

[0070]

[0071] The residual communication signal can be obtained by subtracting the sonar estimated signal from the received signal. This residual communication signal can then be written as:

[0072]

[0073] As can be seen from the above formula, given the circular matrix X, the channel tap coefficients h can be solved from the received observation data y. i Since h is a sparse vector with sparsity M, solving this problem is transformed into a signal reconstruction problem under compressed sensing theory. The solution for vector h can be written mathematically as:

[0074]

[0075] Here, ||h||0 represents the L0 norm of the vector, and minimizing the L0 norm is an NP-hard problem. Regularization is a method that can alleviate ill-conditioned features, stabilize matrix inversion, and improve the accuracy of sparse estimation. We use the L1 norm as the regularization term, and the solution to this problem can be transformed into:

[0076]

[0077] Here, λ represents the regularization factor. The advantage of the L0 norm as a regularization term is that it can produce sparse solutions and is insensitive to outliers. Solving this formula is a convex optimization problem. Considering the computational complexity, the commonly used solution method is gradient-based, which has a simple structure and low complexity. The Fast Soft Thresholding Iterative Algorithm (FISTA) is adopted, which has relatively low complexity and good convergence.

[0078] The basic principle of the FISTA algorithm is as follows:

[0079]

[0080] Where ρ+ is the Euclidean projection (non-negative quadrant), For the descent gradient, L is the Lipschose continuous gradient operator, numerically equal to A. T The largest eigenvalue of A. The gradient can be expressed by the formula:

[0081]

[0082] The core idea of ​​the projection algorithm is to make the vector field decrease the fastest along the negative gradient direction, so that the objective function repeatedly searches for the q value in the negative gradient direction with a specific step size.

[0083] Appendix Figure 4 Channel estimation results based on matched correlation are presented, describing the estimation of amplitude and multipath delay using the matched correlation estimation scheme, as well as the actual Bellhop channel impulse response (the actual Bellhop channel impulse response will be appended). Figure 3(Given). With a signal-to-noise ratio (SNR) of 5dB, it can be seen that the peak value of the matched correlation estimation almost coincides with the channel impulse response location. However, some time delay estimations still have deviations, and the estimation deviation of the amplitude increases with the increase of time delay. This is a result of noise interference, and the periodic integrity of the chirp signal also affects the time delay of the matched estimation. For example, multiple energy paths exist between the main path and the first delay, which will have a serious impact on sonar signal cancellation.

[0084] Appendix Figure 5 Channel estimation results based on the FISTA sparse estimation algorithm are presented, describing the estimation of system multipath delay and amplitude by FISTA sparse estimation, and the actual Bellhop channel impulse response, with a signal-to-noise ratio (SNR) of 5 dB and a regularization parameter λ of 0.1. As can be seen from the figures, compared to FISTA sparse estimation, the peak value and channel impulse response location almost coincide, the amplitude estimation is more accurate, the sensitivity to noise is lower, and it is unaffected by the periodicity of the chirp signal. Therefore, it is more suitable as a parameter estimation scheme based on VTRM channel equalization.

[0085] Furthermore, the block diagram of the VTRM equalization principle based on FISTA sparse channel estimation proposed in step S3 is attached. Figure 6 The following is given, along with the Viterbi decoding algorithm based on Gaussian kernel soft-limiting, in the appendix. Figure 7 The information is provided in the text.

[0086] Example 2: This example provides a method for receiving underwater acoustic communication based on superimposed interference signals, such as... Figure 1 As shown, it includes the following steps:

[0087] S101: Bit data is encoded by convolutional code channel and then subjected to BPSK digital modulation and carrier modulation to obtain communication signals, while simultaneously generating sonar signals of the same frequency.

[0088] In a specific embodiment, the sampling frequency is 100KHz, the symbol length is 1ms, the transmission rate is 1000bps, the carrier frequency is 30KHz, the sonar signal length is also 1ms, and the pulse interval is 1ms.

[0089] S102: The communication signal and the sonar signal are superimposed, and a synchronization header is added for signal synchronization. At this time, the signal passes through the actual underwater acoustic channel and marine environmental noise is added.

[0090] In the specific implementation simulation, the transmitting end signal is processed through the generated BELLHOP underwater acoustic channel and Middleton Class A impulse noise sequence to obtain the received signal; in addition, in the sea trial semi-physical simulation experiment, the transmitting end obtains the received signal through the actual ocean channel and ocean noise.

[0091] S103: The received signal sequence is subjected to sonar interference signal cancellation. The specific scheme is to first extract the synchronization header, perform FISTA sparse estimation on the synchronization header to obtain the channel estimation result, reconstruct the sonar interference signal at the receiving end, and finally subtract the reconstructed sonar interference signal from the received data.

[0092] In the simulation example, the received signal (including noise) is estimated using the FISTA algorithm to obtain an estimate of the Bellhop underwater acoustic channel. This estimate is then convolved with the sonar signal to obtain the reconstructed sonar interference signal. The Bellhop underwater acoustic channel estimate can be compared with the original Bellhop channel, thus reflecting the accuracy of the algorithm. In the semi-physical simulation experiment during sea trials, the received signal (including noise) is estimated using the FISTA algorithm to obtain an estimate of the actual underwater acoustic channel. This estimate is then convolved with the sonar signal to obtain the reconstructed sonar interference signal.

[0093] In the simulation, the performance of the FISTA algorithm for channel estimation is related to the signal-to-noise ratio (SNR). In low SNR environments, the estimation results are poor, and the recovery of sonar interference signals is also poor. In medium and high SNR environments, the channel estimation results and the recovery of sonar interference signals are better.

[0094] S104: Based on the channel estimation results, a time-reversal mirror is used to perform VTRM channel equalization on the signal after sonar interference cancellation to enhance the energy of the main path.

[0095] In specific embodiments, the channel estimation results based on matched correlation estimation and the channel estimation results based on the FISTA sparse algorithm are analyzed. The superiority of the proposed algorithm and the accuracy of channel estimation are verified under the same signal-to-noise ratio.

[0096] S105: Due to the influence of non-Gaussian noise from the ocean, the signal is still affected by impulse noise after equalization. Therefore, we adopt an impulse noise suppression algorithm based on Gaussian kernel soft limiting combined with a Viterbi decoder to improve the signal-to-noise ratio of the system.

[0097] In specific embodiments, the performance of Viterbi decoding hard-decision algorithm, Viterbi decoding soft-decision algorithm, LLR-based impulse noise suppression algorithm, and Gaussian kernel-based soft-limiting impulse noise suppression algorithm were analyzed. Simulation results and sea trial results verified the superiority of the Gaussian kernel-based soft-limiting scheme.

[0098] Furthermore, all parts of this application that are not described in detail are the same as or implemented using existing technology.

[0099] In summary:

[0100] Based on prior information about the interference signal, this invention proposes a sparse channel time-domain signal reconstruction scheme based on the Fast Soft Threshold Iteration Algorithm (FISTA), which has higher reconstruction accuracy for sonar interference signals. To address the effects of residual sonar interference, multipath effects, and environmental noise, this invention proposes a VTRM equalization method based on FISTA sparse estimation and a nonparametric Viterbi decoding algorithm based on Gaussian kernels. These methods effectively solve the problems caused by multipath effects and environmental noise, thereby significantly improving the performance and signal reception accuracy of underwater acoustic communication.

[0101] Those skilled in the art will readily understand that the above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for receiving an underwater acoustic communication based on a superimposed interference signal, characterized in that, The method comprises the following steps: S101, bit data is subjected to convolution code channel coding, then subjected to BPSK digital modulation and carrier modulation to obtain a communication signal, and a simultaneously same-frequency sonar signal is generated; S102, the communication signal is superimposed with the sonar signal, and a synchronization header is added for signal synchronization, at this time, the signal is subjected to actual underwater acoustic channel processing and ocean environment noise addition; S103, sonar interference signal elimination is performed on the received signal sequence, comprising: first, the synchronization header in the received signal is extracted, then the FISTA algorithm is used to estimate the BELLHOP underwater acoustic channel to obtain an estimated value, and the sonar signal is convolved to obtain a reconstructed sonar interference signal, and finally the reconstructed sonar interference signal is subtracted from the received data; the flow steps of the FISTA algorithm, comprising: 1) Initialization ; 2） ； 3） ； 4） ； wherein, is the Euclidean projection onto the non-negative quadrant, is the descent gradient, is the Lipschitz continuous gradient operator, numerically equal to the largest eigenvalue of S104, according to the channel estimation result, VTRM channel equalization is performed on the signal after sonar interference elimination by using a time reversal mirror to enhance the energy of the main path, comprising: S41, signal reception: receiving the signal after sonar interference elimination and performing preprocessing; S42, time reversal: time reversal processing is performed on the received signal to offset the distortion caused by the channel; S43, spectral equalization: spectral equalization processing is performed on the reversed signal to enhance the energy of the signal main path; S44, demodulation and data recovery: demodulation and data recovery are performed on the equalized signal to obtain the processed data transmission result; S105, a pulse noise suppression algorithm based on Gaussian kernel soft limiting is used in combination with a Viterbi decoder to improve the signal-to-noise ratio of the receiving system.

2. The method according to claim 1, wherein In step S101, the ocean channel and the environmental noise in the target water area are first collected, and the environmental noise model and the channel model under the water acoustic communication are established; then the generated bit data is subjected to (2, 1, 3) convolution code encoding, and a BPSK digital modulation method is used to generate a BPSK communication signal, and then carrier modulation is performed and superimposed on the simultaneously same-frequency sonar signal to serve as the transmitting end data.

3. The method according to claim 2, wherein, After pulse shaping, the BPSK communication signal is superimposed with the LMF signal of the sonar, and the mathematical formula is expressed as follows: S(n) = S b (n) + S c (n) wherein S = 1 is a time domain superimposed signal to be transmitted, is a BPSK communication signal after convolution code encoding, is a transmitted sonar Chirp signal.

4. The method of claim 1, wherein In step S102, the transmitting end signal passes through the generated BELLHOP underwater acoustic channel and Middleton Class A pulse noise sequence to obtain the received signal.

5. The method of claim 1, wherein In step S105, the pulse noise suppression algorithm based on Gaussian kernel soft limiting comprises: S51, preprocessing: filtering preprocessing is performed on the data transmission result; S52, Gaussian kernel smoothing: Gaussian kernel is used to smooth the preprocessed signal, specifically, the Gaussian kernel is convolved with the preprocessed signal; S53, wavelet transform and soft limiting: the smoothed signal is subjected to wavelet transform to obtain wavelet coefficients of different scales, and a preset threshold is set to remove noise components in the wavelet coefficients; S54, inverse wavelet transform: the processed wavelet coefficients are subjected to inverse wavelet transform to obtain the denoised signal.

Images