Underwater acoustic communication equalization method based on leakage normalized least mean square algorithm

Abstract

The invention discloses an underwater acoustic communication equalization method based on a leakage normalized least mean square algorithm, and belongs to the technical field of underwater acoustic communication. The method comprises the following steps: acquiring a baseband sampling signal; initializing a weight vector w = 0, an instantaneous power vector p = 0 and an error vector e = 0; and taking the baseband sampling signal x as an input signal, taking the training sequence as an expected signal d, and carrying out gradual leakage updating on the weight of a filter through iterative calculation based on a leakage factor control strategy so as to complete self-adaptive adjustment in a signal equalization process. According to the method, the leakage factor control strategy is introduced, and the adaptive step parameter design is adopted, so that the stability and the convergence speed of the weight of the filter are effectively balanced, finally, the received signal is accurately balanced, and the robustness and the signal quality of the underwater acoustic communication system are remarkably improved.

Description

Technical Field

[0001] The invention belongs to the technical field of underwater acoustic communication, and in particular relates to an underwater acoustic communication equalization method based on a leakage normalized least mean square algorithm. Background Art

[0002] In underwater acoustic communication technology, the minimum mean square error (LMS) and its variant algorithms are widely used in the design of channel equalizers in underwater acoustic communication systems. Bismor et al. (2023) proposed a partially updated leaky LMS algorithm (Bismor D. Leaky partial update LMS algorithms in application to structural active noise control [J]. Sensors, 2023, 23 (3): 1169.) and applied it to the field of structural active noise control. This method reduces the computational complexity by introducing a partial update strategy, and at the same time limits the growth of the filter weight by the leakage factor. However, although this method shows good computational efficiency and stability when dealing with structural noise control, in a complex underwater acoustic channel environment, due to the significant multipath effect and the rapid change of the signal over time, the partial update strategy may cause the system response to lag, thereby affecting the convergence speed and performance of the equalizer. Therefore, directly applying this technology to underwater acoustic communication may have the problem of insufficient adaptability.

[0003] Lopes et al. (2023) conducted a detailed analysis of the LMS and normalized least mean square (NLMS) algorithms (Lopes PA C. Analysis of the LMS and NLMS algorithms using the misalignment norm [J]. Signal, Image and Video Processing, 2023, 17 (7): 3623-3628.), and used the mismatch norm to evaluate the error convergence characteristics of the algorithm. The study shows that NLMS can better cope with the mismatch problem when there is a large change in the input signal amplitude, with faster convergence speed and lower steady-state error. However, this method fails to effectively solve the problem of unbounded growth of filter weights caused by the highly time-varying nature of the underwater acoustic channel in dynamic and complex environments. In addition, relying solely on normalization is not enough to cope with multipath interference and long-term channel changes, which may cause system instability in complex underwater acoustic environments. This deficiency is particularly prominent in underwater scenarios with severe multipath interference, making it challenging to directly apply the NLMS algorithm to underwater acoustic communication systems.

[0004] In response to the above problems, some researchers have introduced leakage factors to constrain the filter weights in order to improve the stability of the algorithm. However, the existing leakage factor setting methods are usually static and cannot be dynamically adjusted according to environmental changes, resulting in certain bottlenecks in the performance of the algorithm in rapidly changing underwater acoustic channels. Especially in complex underwater environments, where channel states change frequently and are difficult to predict, static leakage strategies can easily lead to weight update lags, affecting the tracking ability and convergence speed of the equalizer. Therefore, although traditional methods have made many improvements in improving convergence speed and system stability, they still have deficiencies in computational efficiency, convergence speed, and the ability to respond to channel changes in dealing with complex underwater channel characteristics. Summary of the invention

[0005] Technical issues to be solved:

[0006] In order to avoid the shortcomings of the prior art, the present invention provides an underwater acoustic communication equalization method based on the leakage normalized least mean square algorithm. The invention uses an adaptive filtering framework, combined with a leakage factor adjustment mechanism, to continuously update the filter weights, thereby achieving channel equalization in a dynamic underwater acoustic environment. By introducing a leakage factor control strategy and adopting an adaptive step size parameter design, the stability and convergence speed of the filter weights are effectively balanced, and finally the received signal is accurately equalized, significantly improving the robustness and signal quality of the underwater acoustic communication system. The present invention is suitable for underwater acoustic communications, ocean detection and other fields.

[0007] The technical solution of the present invention is: an underwater acoustic communication equalization method based on leakage normalized least mean square algorithm, and the specific steps are as follows:

[0008] Get the baseband sampling signal x;

[0009] Initialize the weight vector w=0, the instantaneous power vector p=0 and the error vector e=0;

[0010] The baseband sampling signal x is taken as the input signal and the training sequence is taken as the expected signal d. Based on the leakage factor control strategy, the filter weights are gradually leak-updated through iterative calculation to complete the adaptive adjustment in the signal equalization process.

[0011] A further technical solution of the present invention is: the method for obtaining the baseband sampling signal x is to obtain the signal through the underwater acoustic communication receiving end, multiply the signal with the carrier after signal synchronization and resampling, and then process it through a low-pass filter to obtain the baseband sampling signal x.

[0012] A further technical solution of the present invention is: the specific process of the iterative calculation is as follows:

[0013] Calculate the current input vector u, where u = [x[n], x[n-1], ..., x[n-L+1]] T , where L is the filter length and n is the number of bands;

[0014] Calculate the error e[n], where e[n] = d[n] - w T u, e[n] is the signal error between the filter output and the expected signal under the current input;

[0015] Calculate the normalization factor v, where v = u T u+1e -6 Used for normalization to prevent numerical instability;

[0016] Update the weight vector Where λ0 is the initial leakage factor and μ0 is a constant.

[0017] A further technical solution of the present invention is: the number of iterative calculations is equal to the length N of the input signal.

[0018] A further technical solution of the present invention is: the leakage factor control strategy is to dynamically adjust the leakage factor according to the current error and signal changes, so that it maintains a small value in the stable section of the signal, and automatically increases when the signal is subjected to drastic changes or interference, so as to enhance the robustness against noise and sudden interference.

[0019] An underwater acoustic communication equalization system based on a leakage normalized least mean square algorithm comprises an underwater acoustic communication receiving module, a signal processing module, and an adaptive filtering module;

[0020] Acquiring an underwater acoustic signal through the underwater acoustic communication receiving module;

[0021] The signal processing module is used to expand the signal frequency into the range of the carrier frequency, and then the signal is processed by a low-pass filter to obtain a baseband sampling signal;

[0022] The filter weights are leak-updated step by step through the adaptive filtering module to complete the adaptive adjustment in the signal equalization process.

[0023] An electronic device comprises at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the underwater acoustic communication equalization method based on the leakage normalized least mean square algorithm.

[0024] A computer-readable digital storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the underwater acoustic communication equalization method based on the leakage normalized least mean square algorithm when executed.

[0025] Beneficial Effects

[0026] The beneficial effects of the present invention are as follows: the present invention proposes a method for designing an underwater acoustic communication equalizer based on a leakage normalized least mean square (Leaky NLMS) algorithm to achieve dynamic equalization and error control of signal transmission in an underwater acoustic channel. The present invention utilizes a leakage factor mechanism combined with a normalized least mean square algorithm to achieve adaptive adjustment in the signal equalization process by gradually leaking and updating the filter weights, thereby effectively improving the stability and convergence speed of the underwater acoustic channel equalizer. This method can still ensure real-time adjustment and stable transmission of signals under conditions of low signal-to-noise ratio and complex multipath fading, so that the equalizer exhibits higher anti-interference capabilities in harsh underwater acoustic environments. Thanks to the introduction of the leakage mechanism, the algorithm of the present invention exhibits stronger robustness and adaptability in highly dynamically changing underwater acoustic channels, and can more efficiently cope with complex underwater acoustic propagation environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The following is a comparison chart of the computing time of the four algorithms: LMS, VSS-NLMS, NLMS and Leaky NLMS.

[0028] Figure 2 These are the constellation diagrams of the four algorithms, LMS, VSS-NLMS, NLMS and Leaky NLMS, after signal equalization.

[0029] Figure 3 This is a percentage diagram of the error amplitude distribution of the four algorithms, LMS, VSS-NLMS, NLMS and Leaky NLMS, after equalization. DETAILED DESCRIPTION

[0030] The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, but should not be construed as limiting the present invention.

[0031] Based on the existing leakage factor setting method, in a complex underwater environment, the channel state changes frequently and is difficult to predict. The static leakage strategy easily leads to a lag in weight update, which affects the tracking ability and convergence speed of the equalizer. The present invention provides an underwater acoustic communication equalization method based on a leakage normalized least mean square algorithm, and the specific steps are as follows:

[0032] Step 1: At the receiving end of underwater acoustic communication, after signal synchronization and resampling, the processed received signal is multiplied by the carrier and then processed by a low-pass filter to obtain a baseband sampling signal x, which is used as the input signal of the equalizer.

[0033] Step 2: Set the parameter μ0 to a constant (the value is 0.01 in this embodiment), set the leakage factor λ0 to a constant, initialize the weight vector w=0, the instantaneous power vector p=0 and the error vector e=0;

[0034] Step 3: Given an input signal x, a training sequence as the desired signal d, and a filter length L, repeat the following for n iterations (n=1, 2, …, N):

[0035] Calculate the current input vector u, where u = [x[n], x[n-1], ..., x[n-L+1]] T ;

[0036] Calculate the error e[n], where e[n] = d[n] - w T u, e[n] signal error under current input;

[0037] Calculate the normalization factor v, where v = u T u+1e -6 Used for normalization to prevent numerical instability;

[0038] Update the weight vector This formula includes the attenuation of the leakage factor and the weight update based on the error and input vector.

[0039] The number of iterations above is equal to the length N of the input signal.

[0040] The leakage factor control strategy is to dynamically adjust the leakage factor according to the current error and signal changes, so that it maintains a small value in the stable section of the signal, and automatically increases when the signal is subjected to drastic changes or interference, so as to enhance the robustness against noise and sudden interference.

[0041] The present invention discloses an underwater acoustic communication equalization system based on a leakage normalized least mean square algorithm, comprising an underwater acoustic communication receiving module, a signal processing module, and an adaptive filtering module;

[0042] Acquiring an underwater acoustic signal through the underwater acoustic communication receiving module;

[0043] The signal processing module is used to expand the signal frequency into the range of the carrier frequency, and then the signal is processed by a low-pass filter to obtain a baseband sampling signal;

[0044] The filter weights are leak-updated step by step through the adaptive filtering module to complete the adaptive adjustment in the signal equalization process.

[0045] The present invention provides an electronic device, comprising at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the underwater acoustic communication equalization method based on the leakage normalized least mean square algorithm.

[0046] The present invention provides a computer-readable data storage medium, wherein the computer-readable data storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the underwater acoustic communication equalization method based on the leakage normalized least mean square algorithm when executing the computer instructions.

[0047] The present invention discloses an underwater acoustic channel equalization method based on a leakage normalized least mean square (Leaky NLMS) algorithm, which is used for adaptive equalization of underwater communication signals. The method dynamically adjusts the filter weights by introducing a leakage factor control strategy, and adaptively adjusts the step size according to the current error size, thereby improving the signal equalization effect in a complex multipath underwater acoustic channel environment. Compared with traditional methods, the present invention exhibits stronger anti-interference performance in underwater environments with severe multipath interference and noise. Through adaptive leakage control, it can effectively suppress the unbounded growth of weights and accurately equalize received signals. This enhanced robustness plays an important role in improving the accuracy of channel equalization, thereby providing a more stable and reliable performance guarantee for underwater acoustic communication systems. The method is applicable to different channel conditions, especially to complex and changeable underwater acoustic environments, and provides reliable technical support for underwater acoustic signal transmission.

[0048] The above technical solution is further described below in conjunction with the accompanying drawings.

[0049] Figure 1 The comparison of the operation time of LMS, VSS-NLMS, NLMS and Leaky NLMS algorithms is shown. As can be seen from the figure, the operation time of all algorithms is concentrated around 2.4×10 -4 to 3.8×10 -4 In contrast, the computation time of the LMS algorithm is relatively stable with a small fluctuation range, while the computation time of VSS-NLMS and NLMS varies greatly, indicating that these two algorithms require additional computations in some cases, which may be related to the adjustment of their adaptive parameters. The computation time distribution of Leaky NLMS is relatively concentrated, indicating that the algorithm has a relatively stable control over the computation time after the introduction of the leakage factor. Therefore, Leaky NLMS has certain advantages in maintaining the stability of computation time, and is particularly suitable for application scenarios with limited computing resources and requiring stable response time.

[0050] Figure 2The constellation diagrams of the four algorithms after signal equalization are shown. The distribution of constellation points of the LMS algorithm is relatively scattered, especially in the edge areas, indicating that the algorithm still has large errors after equalizing the underwater acoustic channel. The distribution of the constellation diagram of the VSS-NLMS algorithm has improved slightly, but the overall error is still obvious, especially under low signal-to-noise ratio conditions, the signal points show significant offsets. The constellation diagram of the NLMS algorithm shows a relatively uniform distribution of signal points, but some points in some directions are still not accurately gathered. The constellation diagram of Leaky NLMS is the most concentrated, especially around the ideal position point, the signal point concentration is high, indicating that it can better suppress noise and multipath effects when equalizing complex multipath underwater acoustic channels. Therefore, Leaky NLMS shows the best equalization effect in the constellation diagram performance, verifying its good suppression ability of channel distortion and noise.

[0051] Figure 3 The percentage of error amplitude distribution of different algorithms after equalization is shown. For the LMS algorithm, it can be seen that a large proportion of errors are concentrated at 10 -2 to 10 -1 The error distribution of the VSS-NLMS algorithm has improved slightly, but there are still some large errors, especially in the case of 10 -3 to 10 -2 The error distribution of the NLMS algorithm shows an increase in the proportion of errors within the smaller error range, but there are still some large errors under low precision conditions. The Leaky NLMS algorithm has almost all errors concentrated in the 10 -3 The following shows that the dynamic adjustment of the weights by the leakage factor effectively controls the equalization error and significantly improves the accuracy of signal recovery. Therefore, Leaky NLMS performs best in error distribution and can better adapt to the equalization requirements of underwater acoustic signals in complex environments.

[0052] The box plot of the operation time further shows that the differences between the algorithms are reflected in stability. The upper and lower quartile difference of the LMS algorithm's operation time is small, which means that its calculation process is relatively stable, which is related to its simple weight update mechanism. However, its stability comes at the cost of lower equalization performance. Due to the introduction of step size adjustment, the fluctuation of VSS-NLMS's operation time has increased, which may be due to the additional computational burden involved in the calculation process of the adaptive step size. Although the NLMS algorithm has the blessing of normalization, it still leads to inconsistent operation time in some scenarios, especially when the signal characteristics change rapidly. In contrast, the Leaky NLMS algorithm introduces a leakage factor, which enables the weights to be better controlled during the update process, thereby effectively reducing the computational fluctuations caused by drastic changes in the channel, making it a choice with better operation time stability.

[0053] By comparing the four constellation diagrams, we can see the differences in the ability of each algorithm to deal with multipath effects and noise interference. Due to the lack of an effective weight control mechanism, the LMS algorithm is easily affected by multipath signals and noise, resulting in the inability of signal points to gather at the ideal position. The VSS-NLMS algorithm improves this situation to a certain extent through adaptive step size adjustment, but its effect is still limited under complex channel conditions. After normalization, the NLMS algorithm can better handle changes in the input signal amplitude and make the signal points more concentrated. However, Leaky NLMS combines the two mechanisms of leakage factor and normalization to effectively prevent the unbounded growth of weights and enhance the adaptability to dynamic environments. The point aggregation of its constellation diagram is significantly better than the other three algorithms, indicating that it has stronger robustness and adaptability in underwater acoustic channel equalization.

[0054] In the histogram of error range distribution, the Leaky NLMS algorithm shows excellent error control ability, and almost all the errors after equalization are within 10 -3 As follows. This is in stark contrast to the performance of LMS, VSS-NLMS, and NLMS, especially LMS and VSS-NLMS, which have a higher proportion of large errors, indicating that their robustness under noisy and multipath channel conditions is weak. Although NLMS reduces some errors through normalization, it still fails to completely eliminate the impact of channel changes. LeakyNLMS effectively suppresses the divergence of weights by introducing a leakage factor, ensuring that the algorithm can maintain a low steady-state error under different channel conditions. This improvement in robustness makes Leaky NLMS particularly suitable for underwater acoustic communication systems with complex environmental changes and high signal quality requirements, and has important practical application value.

[0055] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations on the present invention. A person skilled in the art may change, modify, replace and modify the above embodiments within the scope of the present invention without departing from the principles and purpose of the present invention.

Claims

1. A method for underwater acoustic communication equalization based on leakage normalized least mean square algorithm, characterized in that The specific steps are as follows: Get the baseband sampling signal x; Initialize the weight vector w=0, the instantaneous power vector p=0 and the error vector e=0; The baseband sampling signal x is taken as the input signal and the training sequence is taken as the expected signal d. Based on the leakage factor control strategy, the filter weights are gradually leak-updated through iterative calculation to complete the adaptive adjustment in the signal equalization process.

2. According to claim 1, a method for underwater acoustic communication equalization based on leakage normalized least mean square algorithm is characterized by: The method for acquiring the baseband sampling signal x is to acquire the signal through the underwater acoustic communication receiving end, multiply the signal with the carrier after signal synchronization and resampling, and then process it through a low-pass filter to obtain the baseband sampling signal x.

3. According to claim 2, a method for underwater acoustic communication equalization based on leakage normalized least mean square algorithm is characterized by: The specific process of the iterative calculation is as follows: Calculate the current input vector u, where u = [x[n], x[n-1], ..., x[n-L+1]] T , where L is the filter length and n is the number of bands; Calculate the error e[n], where e[n] = d[n] - w T u, e[n] is the signal error between the filter output and the expected signal under the current input; Calculate the normalization factor v, where v = u T u+1e -6 Used for normalization to prevent numerical instability; Update the weight vector Where λ0 is the initial leakage factor and μ0 is a constant.

4. According to claim 3, a method for underwater acoustic communication equalization based on leakage normalized least mean square algorithm is characterized by: The number of iterative calculations is equal to the length N of the input signal.

5. The underwater acoustic communication equalization method based on leakage normalized least mean square algorithm according to claim 4 is characterized by: The leakage factor control strategy is to dynamically adjust the leakage factor according to the current error and signal changes, so that it maintains a small value in the stable section of the signal, and automatically increases when the signal is subjected to drastic changes or interference, so as to enhance the robustness against noise and sudden interference.

6. An underwater acoustic communication equalization system based on a leakage normalized least mean square algorithm, used to implement the underwater acoustic communication equalization method based on a leakage normalized least mean square algorithm as described in any one of claims 1 to 5; characterized in that: It includes an underwater acoustic communication receiving module, a signal processing module, and an adaptive filtering module; Acquiring an underwater acoustic signal through the underwater acoustic communication receiving module; The signal processing module is used to expand the signal frequency into the range of the carrier frequency, and then the signal is processed by a low-pass filter to obtain a baseband sampling signal; The filter weights are leak-updated step by step through the adaptive filtering module to complete the adaptive adjustment in the signal equalization process.

7. An electronic device, characterized in that: It includes at least one processor and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the underwater acoustic communication equalization method based on the leakage normalized least mean square algorithm as described in any one of claims 1-5.

8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the underwater acoustic communication equalization method based on the leakage normalized least mean square algorithm as described in any one of claims 1-5 when executed.

Images