Marine environment-oriented fishing boat networking end-to-end semantic communication method and system

Abstract

The invention provides a fishing boat networking end-to-end semantic communication method and system oriented to an offshore environment, and relates to the technical field of wireless communication and artificial intelligence, and the method comprises the steps: obtaining ocean buoy data, and converting the ocean buoy data into an ocean buoy data matrix; converting the preprocessed ocean buoy data matrix into an ocean buoy network data tensor, inputting the ocean buoy network data tensor into a seaborne environment-oriented semantic communication network, extracting a feature map of the ocean buoy network data tensor for multiple times by using a convolution block, continuously adjusting the feature map according to a signal-to-noise ratio through a self-adaptive feature block to obtain a feature vector, and carrying out feature extraction on the feature vector to obtain a seaborne environment-oriented semantic communication network; and then compressing the feature vector according to a ratio based on a data compression ratio to obtain a semantic feature vector, carrying out attenuation processing on the semantic feature vector through total channel response of a channel model to obtain a final output signal, and generating recovery data from the final output signal through an inverse reasoning process of a semantic decoder.

Description

Technical Field

[0001] The present disclosure relates to the fields of wireless communication and artificial intelligence technologies, and particularly to an end-to-end semantic communication method and system for fishing boat networking in a maritime environment. Background Art

[0002] The statements in this section merely provide background technical information related to the present disclosure and do not necessarily constitute prior art.

[0003] With the rapid development of the marine economy, the demand for efficient and reliable communication in scenarios such as offshore fishing boat operations, fishery resource management, and maritime emergency rescue is increasing day by day. However, the maritime communication environment has unique complexities, mainly manifested as problems such as long-distance transmission, significant multipath effects, severe Doppler frequency shift, large signal attenuation, and interference from adverse weather. These factors cause traditional wireless communication technologies to face problems such as severe signal attenuation, high latency, and poor reliability in the maritime environment.

[0004] As an emerging communication paradigm, semantic communication can significantly reduce communication overhead and improve transmission efficiency by extracting and transmitting the semantic features of information rather than raw data, which makes semantic communication have obvious advantages in environments with limited communication bandwidth and latency sensitivity. At the same time, as a new communication technology that integrates communication and intelligence, semantic communication proposes a new idea of replacing "modular" separate optimization with "end-to-end" through-type optimization, and can achieve a significant improvement in the overall performance of the communication system with a simpler network structure.

[0005] Although semantic communication has great potential in the application of the maritime environment, existing semantic communication technologies are mainly oriented to the terrestrial environment and have not fully considered the particularity of the complex propagation environment at sea. There are still the following problems: 1) Due to the strong reflectivity of seawater and the interference of weather conditions (such as wind, tides, ocean currents, etc.) on signals, signals will encounter severe attenuation and multipath effects during propagation, and it is impossible to simulate the real maritime environment and put it into the learning paradigm.

[0006] 2) The complex marine environment leads to the variability of marine data, and it is impossible to effectively process the complex semantic information under marine data, resulting in problems such as high latency and low transmission accuracy.

[0007] 3) In the maritime environment, fishing boat formation operations usually require multi-ship coordination. Traditional communication modes are difficult to meet the networking requirements, and existing networking technologies lack support for semantic communication, resulting in low communication efficiency and insufficient resource utilization. Summary of the Invention

[0008] To solve the above problems, the present disclosure proposes an end-to-end semantic communication method and system for fishing boat networking in a marine environment, extracts the semantic information of marine data, constructs a semantic communication environment model based on the Rice channel model, simulates problems such as attenuation, delay, and multipath effects caused by marine communication, and incorporates them into the learning paradigm. Through multi-scale feature learning and simultaneously increasing the depth of the generated feature maps, the ability of the model to extract complex semantic features is significantly enhanced, achieving a significant improvement in the overall performance of the communication system.

[0009] According to some embodiments, the present disclosure adopts the following technical solutions: An end-to-end semantic communication method for fishing boat networking in a marine environment, comprising: Obtain ocean buoy data, convert the ocean buoy data into matrix data according to time granularity and space granularity to obtain an ocean buoy data matrix, and perform preprocessing; Convert the preprocessed ocean buoy data matrix into an ocean buoy network data tensor, and input it into a semantic communication network for the marine environment to obtain restored data of end-to-end semantic communication; Wherein, after the ocean buoy network data tensor is input into the semantic communication network for the marine environment, the feature maps of the ocean buoy network data tensor are extracted multiple times by a convolutional block, and then the feature maps are continuously adjusted according to the signal-to-noise ratio by an adaptive feature block to obtain feature vectors. Then, the feature vectors are compressed according to a ratio based on the data compression rate to obtain semantic feature vectors. The semantic feature vectors are attenuated by the overall channel response of the channel model to obtain a final output signal, and the final output signal generates restored data through the inverse inference process of a semantic decoder.

[0010] According to some embodiments, the present disclosure adopts the following technical solutions: An end-to-end semantic communication system for fishing boat networking in a marine environment, characterized by comprising: A data acquisition module, configured to obtain ocean buoy data, convert the ocean buoy data into matrix data according to time granularity and space granularity to obtain an ocean buoy data matrix, and perform preprocessing; A semantic restoration module, configured to convert the preprocessed ocean buoy data matrix into an ocean buoy network data tensor, and input it into a semantic communication network for the marine environment to obtain restored data of end-to-end semantic communication; Among them, after the ocean buoy network data tensor is input into the semantic communication network for the marine environment, the convolutional block is used to extract the feature maps of the ocean buoy network data tensor multiple times, and then the adaptive feature block continuously adjusts the feature maps according to the signal-to-noise ratio to obtain feature vectors. Then, based on the data compression rate, the feature vectors are compressed by a ratio to obtain semantic feature vectors. The semantic feature vectors are attenuated through the overall channel response of the channel model to obtain the final output signal. The final output signal generates the restored data through the inverse inference process of the semantic decoder.

[0011] According to some embodiments, the present disclosure adopts the following technical solutions: A computer program product includes a computer program, and when the computer program is executed by a processor, it implements the end-to-end semantic communication method for fishing boat networking in a marine environment as described above.

[0012] According to some embodiments, the present disclosure adopts the following technical solutions: A non-transitory computer-readable storage medium is used to store computer instructions, and when the computer instructions are executed by a processor, they implement the end-to-end semantic communication method for fishing boat networking in a marine environment as described above.

[0013] According to some embodiments, the present disclosure adopts the following technical solutions: An electronic device includes: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device runs, the processor executes the computer program stored in the memory so that the electronic device executes and implements the end-to-end semantic communication method for fishing boat networking in a marine environment as described above.

[0014] Compared with the prior art, the beneficial effects of the present disclosure are: The end-to-end semantic communication method for fishing boat networking in a marine environment of the present disclosure is based on the end-to-end fishing boat communication network of semantic communication, combined with the end-to-end optimization design of a deep neural network, to construct a semantic communication network for the marine environment, significantly improving the reliability of marine communication; and training and learning are carried out in the learning paradigm based on the Rice channel model to simulate the marine multipath effect, signal attenuation, and time delay, enhancing the robustness of the system in a complex marine environment; the semantic encoder and decoder in the network combine multi-scale feature learning technology, improving the feature extraction and restoration ability of ocean buoy data, and being able to effectively cope with complex marine environments.

[0015] An end-to-end semantic communication method for fishing boat networking in a maritime environment according to the present disclosure includes two main innovative improvements: on the one hand, adding a Rice fading channel learning paradigm, proposing a semantic communication network model for the maritime environment, and realizing the extraction and restoration of semantic information of maritime data. On this basis, modeling the semantic communication environment based on the Rice channel model, simulating problems such as attenuation, delay, and multipath effects brought by maritime communication, and adding them to the learning paradigm; on the other hand, through multi-scale feature learning technology, combining multiple cyclic extractions of feature maps, and combining with increasing the depth of the generated feature maps, enhancing the model's ability to extract complex semantic features, significantly improving the overall performance of the communication system, realizing end-to-end semantic communication for fishing boat networking in a complex maritime propagation environment, and achieving efficient, reliable, and low-latency maritime communication. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] The accompanying drawings forming a part of this disclosure are used to provide a further understanding of the disclosure. The illustrative embodiments and descriptions thereof of the disclosure are used to explain the disclosure and do not constitute an improper limitation of the disclosure.

[0017] Figure 1 Schematic diagram of the maritime propagation environment for point-to-point fishing boat communication in an embodiment of the present disclosure; Figure 2 Processing flowchart of the end-to-end fishing boat networking semantic communication network model for a complex maritime propagation environment in an embodiment of the present disclosure; Figure 3 Effect diagram of ocean buoy data transmission and restoration in an embodiment of the present disclosure; Figure 4 Effect diagram of ocean image data transmission and restoration in an embodiment of the present disclosure; Figure 5 Schematic diagram of semantic encoder compression in an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The present disclosure will be further described below in conjunction with the accompanying drawings and embodiments.

[0019] It should be noted that the following detailed descriptions are all illustrative and are intended to provide a further description of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which the present disclosure belongs.

[0020] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0021] Example 1 An end-to-end semantic communication method for fishing boat networking in a marine environment according to the present disclosure aims to improve the reliability of a communication network in a complex and changeable marine environment. The steps include: Step 1: Obtain ocean buoy data, convert the ocean buoy data into matrix data according to time granularity and space granularity to obtain an ocean buoy data matrix, and perform preprocessing; Step 2: Convert the preprocessed ocean buoy data matrix into an ocean buoy network data tensor, and input it into a semantic communication network for a marine environment to obtain restored data for end-to-end semantic communication; Among them, after the ocean buoy network data tensor is input into the semantic communication network for a marine environment, the convolutional block is used to extract the feature map of the ocean buoy network data tensor multiple times, and then the adaptive feature block continuously adjusts the feature map according to the signal-to-noise ratio to obtain a feature vector. Then, based on the data compression rate, the feature vector is compressed at a ratio to obtain a semantic feature vector. The semantic feature vector is attenuated through the overall channel response of the channel model to obtain a final output signal. The final output signal generates restored data through the inverse inference process of the semantic decoder.

[0022] As an example, an end-to-end semantic communication method for fishing boat networking in a marine environment according to the present disclosure includes two main innovations: 1) Add a Rice fading channel learning paradigm and propose a semantic communication model for a marine environment. 2) Through multi-scale feature learning technology and combined with increasing the depth of the generated feature map, the model's ability to extract complex semantic features is enhanced. The specific training implementation process is as follows: Step 1: Obtain ocean buoy data. Obtain ocean buoy data through an open data platform, such as the buoy position (latitude and longitude) and environmental data (such as temperature, salinity, flow rate, etc.) provided by NOAA or GOOS; or obtain RGB marine image data through the kaggle platform.

[0023] As an example, in fact, whether it is for image data or ocean buoy data, the essence of semantic feature extraction is the same, that is, using a convolutional neural network to extract semantic features and only retaining key information during channel transmission, thereby reducing redundancy and improving communication efficiency. Therefore, the embodiment of this application takes the processing of ocean buoy data as an example to implement an end-to-end semantic communication method for fishing boat networking in a marine environment, but it still retains the basic image semantic transmission. Since its processing process is roughly the same, it will not be elaborated here.

[0024] Specifically, the current communication system in the marine environment mainly consists of key components such as coastal base stations, relay nodes, communication satellites, communication vessels, and unmanned aerial vehicles. Among them, the coastal base station, as an extension of the terrestrial communication network, undertakes tasks such as uploading marine monitoring data and transmitting operation scheduling instructions, realizing efficient information interaction between ships and the land. When a ship is far from the shore base station, the communication satellite provides long-distance data transmission to ensure that the operation fleet can maintain real-time connection with the command center. Relay nodes and unmanned aerial vehicles play auxiliary transmission and extended coverage roles in marine communication, enhancing the flexibility and reliability of the system. On this premise, as Figure 1 shown, this disclosure focuses on point-to-point communication between fishing vessels and proposes a fishing vessel networking system based on semantic communication. The system consists of a sending ship and a receiving ship. Among them, the sending ship serves as the communication core, responsible for coordinating information interaction between ships, and receiving scheduling instructions from the onshore command center through the coastal base station. The receiving ship mainly undertakes fishery operation tasks, including sensor data collection, fishery fishing, cruise monitoring, etc., and transmits relevant data to the sending ship.

[0025] In this system, any two ships can serve as the sending ship or the receiving ship, thereby enhancing the flexibility of the communication system, optimizing the data transmission efficiency, and meeting the requirements of the marine fishery for low-latency and high-reliability communication.

[0026] Furthermore, the seawater surface has high conductivity and dielectric constant, making it an efficient electromagnetic wave reflection medium. According to electromagnetic theory, the influence of seawater on radio waves is mainly reflected in: (1) Specular Reflection: For signals with relatively high frequencies (such as microwaves and millimeter waves), the sea surface can be approximated as a smooth surface, causing the signal to undergo specular reflection according to Snell's law, forming a predictable reflection path.

[0027] (2) Diffuse Reflection: Due to the fluctuations on the sea surface (such as ripples and waves), when a signal is incident on the sea surface, a part of the energy will be scattered in multiple directions, resulting in signal power attenuation and random phase change.

[0028] (3) Absorption: Part of the electromagnetic wave energy will be absorbed by the seawater, resulting in energy loss. In particular, wireless signals in the low-frequency band will be attenuated to a certain extent.

[0029] Figure 1 In , the reflection surface intuitively shows the specular reflection and diffuse reflection processes, and absorption is manifested as energy attenuation, that is, the amplitude decreases.

[0030] Step 2: Convert the ocean buoy data into matrix data according to the time granularity and space granularity to obtain the ocean buoy data matrix, and perform preprocessing; Specifically, the obtained original ocean buoy data is processed into matrix data according to time granularity and spatial granularity. The time granularity is in hours, and the spatial granularity is in the coordinates of different working stations, and it is divided based on three data types: temperature, wave, and wind. After processing, an ocean buoy data matrix is obtained.

[0031] Among them, represents the ocean buoy data matrix obtained within the t-th time period, H represents the height of the matrix, W represents the width of the matrix, C represents different representation dimensions under the same data type.

[0032] As an example, if it is image data, it is transmitted in the format of an RGB three-dimensional pixel matrix.

[0033] Furthermore, the obtained ocean buoy data matrix is preprocessed, including: Taking each ocean buoy data matrix as a unit to perform data normalization, missing value filling, and outlier processing to improve data quality and ensure the stability of the model.

[0034] (1) Since ocean buoys may have missing data due to equipment failures or communication interruptions, the resulting NaN data (Not a Number, a special floating-point number representing invalid or missing values, usually caused by abnormal operation of the sensor data acquisition system) will cause the neural network training gradient value to be NaN, and further prevent the normal update of model parameters. Therefore, it is necessary to fill the NaN values in the matrix. The mathematical representation of the filling process is formula (1): (1) Among them: is the data point at the t-th time step, the h, w position, and the c-th channel; N is the total number of non-NaN data in the current column. To prevent data filling failure due to large-scale data loss, when obtaining training data before model training in this disclosure, the NaN values existing in the data will be detected again, and the mean value of the non-NaN data in the corresponding channels of multiple ocean buoy data matrices will be used as the secondary filling process.

[0035] (2) Using the Min-Max normalization method, the data is scaled to the range [0, 1], and the calculation formula (2) is as follows: (2) Among them, and are the minimum and maximum values of this data set respectively. is the normalized ocean buoy data matrix.

[0036] Step 3: Convert the preprocessed ocean buoy data matrix into an ocean buoy network data tensor and input it into the semantic communication network for the marine environment to obtain the restored data of end-to-end semantic communication, including: Step 31: Construct and train a semantic communication network for the marine environment; Specifically, first, the ocean buoy data matrix is converted into a tensor, that is, the ocean buoy network data tensor , and the data set is divided into batches. For the ocean buoy data matrix, randomly selected data segments are used as training data, and the remaining data segments are used as test data, which are used as the training input of the subsequent semantic communication network.

[0037] Furthermore, based on the ocean buoy network data tensor , channel state information CSI, data compression ratio CR, and number of channels channels, construct a semantic communication network MS for the marine environment. The semantic communication network for the marine environment includes a speech encoder, a marine channel model, and a semantic decoder. Among them, the marine channel model is constructed based on the channel state information CSI, the semantic encoder and semantic decoder are constructed based on the data compression ratio CR and the number of channels channels, and the model parameters are updated based on the ocean buoy network data tensor .

[0038] Finally, the training process uses the gradient descent algorithm to train the semantic communication network for the marine environment, calculates the gradient of each parameter (i.e., the weight in the model) with respect to the loss based on backpropagation until the loop ends.

[0039] Step 32: Convert the normalized ocean buoy data matrix into a tensor to obtain the ocean buoy network data tensor. After inputting it into the semantic communication network for the marine environment, the specific processing process of the semantic communication network for the marine environment is as follows: First, the ocean buoy network data tensor enters the semantic encoder, and the spatial pattern or local correlation of the ocean buoy network data tensor is extracted based on the convolutional block (conv). The output feature map is normalized through the generalized divisive normalization layer (GDN) to remove redundant information in the input data, reduce overfitting to certain features, and then the feature matrix is nonlinearly activated through the PReLU layer to avoid "death" of neurons during training. The calculation process is shown in formula (3): (3) Furthermore, the features of the output feature map are further extracted. Meanwhile, to effectively alleviate the problem of gradient disappearance in deep networks, a residual structure is constructed as shown in Equation (4): (4) Furthermore, based on calculating the mean of features and combining the signal-to-noise ratio (SNR), an adaptive feature block AF is constructed. The adaptive feature block can adjust features according to the signal-to-noise ratio, thereby helping the model better handle noise or other environmental changes, as shown in Equation (5): (5) The feature matrix F output by Equation (5) is taken as the input and fed back into Equation (3). The processes of (3), (4), and (5) are repeated N times to obtain the output feature vector , and finally, based on the data compression rate CR, the feature vector is compressed at a ratio. A semantic encoder is designed, and its output semantic information is a one-dimensional tensor. The data compression rate is CR, and the compression process is as Figure 5 shown. Assuming that the encoder outputs semantic information of 1024 bits and the compression rate CR = 0.1, then we discard the latter 90% of the data, and the sender actually sends the first 1024 * 0.1 = 102 (rounded) bits of data. The semantic feature output of the semantic encoder, that is, the semantic feature vector is obtained.

[0040] Secondly, the semantic feature vector is attenuated through the total channel response of the maritime channel model to obtain the final output signal, including: First, a channel model is constructed based on the channel state information CSI, and there is , and the Rician channel is selected to simulate the maritime environment. The Rician fading channel is a common fading model in wireless communication. Its model assumes that there is a dominant direct-path signal, and it also includes multiple path signals generated by reflection, refraction, or scattering of other objects.

[0041] Since the sea surface is a strong reflecting surface, multiple transmission paths are usually generated in maritime communication, and there is usually a strong LOS direct-path signal in point-to-point fishing boat communication in the sea. Therefore, the Rician channel model is very suitable for simulating the signal propagation characteristics in maritime communication. By adjusting the K value, the Rician model can better adapt to different maritime propagation conditions. The specific process is as follows: Determine the gain coefficient of the channel based on the signal-to-noise ratio SNR , and there is Simulate the fading of the analog signal in the Rician channel. Divide the signal into two parts, the real part and the imaginary part, to form a complex signal. Calculate the LOS direct-path signal in complex form, where K is the power ratio of the direct path to the multipath scattering path in Rician fading, L is the characteristic length, and the LOS (direct path) component is shown in Equation (6): (6) The signal component of the scattering path is a random signal with a Gaussian distribution, as shown in Equation (7): (7) Perform random time delay processing on the signal of each scattering path, and accumulate the signals of all scattering paths to obtain the final multipath scattering signal Add the LOS direct-path signal and the multipath scattering signal to obtain the total channel response Semantic feature vector Through the total channel response Perform attenuation processing to obtain the output signal after the fading channel .

[0042] According to the set noise power P, generate Gaussian noise and add it to the signal to obtain the received signal , and then perform normalization processing on the received signal to remove the influence of the channel gain to obtain the final output signal , as shown in Equation (8): (8) Finally, generate the recovered data through the inverse inference process of the semantic decoder for the final output signal, including: To recover the output signal information compressed by the semantic communication network for the maritime environment, a semantic decoder is constructed based on the transposed convolutional layer. Feature extraction is performed on the semantic compressed data passing through the channel based on the transposed convolutional block (deconv), normalization is performed on the output feature map through the generalized divisive normalization layer (GDN), and then the feature matrix is non-linearly activated through the PReLU layer, as shown in Equation (9).

[0043] (9) Construct the residual structure again, as shown in Equation (10): (10) Corresponding to the encoder part, construct the adaptive feature block AF to perform inverse inference on the encoded data, as shown in Equation (11): (11) The recovered feature matrix of Equation (11) It is fed back into formula (9) as input, and the processes of (9), (10), and (11) are repeated N times to obtain the restored data. .

[0044] Embodiment 2 In an embodiment of the present disclosure, a fishing boat networking end-to-end semantic communication system for a marine environment is provided, including: A data acquisition module, configured to acquire ocean buoy data, convert the ocean buoy data into matrix data according to time granularity and space granularity to obtain an ocean buoy data matrix, and perform preprocessing; A semantic restoration module, configured to convert the preprocessed ocean buoy data matrix into an ocean buoy network data tensor, and input it into a semantic communication network for the marine environment to obtain end-to-end semantic communication restored data; Among them, after the ocean buoy network data tensor is input into the semantic communication network for the marine environment, the convolutional block is used to extract the feature maps of the ocean buoy network data tensor multiple times, and then the adaptive feature block is used to continuously adjust the feature maps according to the signal-to-noise ratio to obtain feature vectors. Then, based on the data compression ratio, the feature vectors are compressed according to the ratio to obtain semantic feature vectors. The semantic feature vectors are attenuated by the total channel response of the channel model to obtain the final output signal. The final output signal is used to generate restored data through the inverse inference process of the semantic decoder.

[0045] Embodiment 3 In an embodiment of the present disclosure, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, it implements the above-mentioned fishing boat networking end-to-end semantic communication method for a marine environment.

[0046] Embodiment 4 In an embodiment of the present disclosure, a non-transitory computer-readable storage medium is provided, and the non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, the above-mentioned fishing boat networking end-to-end semantic communication method for a marine environment is implemented.

[0047] Embodiment 5 In an embodiment of the present disclosure, an electronic device is provided, including: a processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device runs, the processor executes the computer program stored in the memory so that the electronic device executes and implements the above-mentioned fishing boat networking end-to-end semantic communication method for a marine environment.

[0048] This disclosure is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the disclosure. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processors of general-purpose computers, special-purpose computers, embedded processors, or other programmable data processing devices to generate a machine, such that the instructions executed by the processors of the computer or other programmable data processing devices produce means for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0049] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one flow Figure 1 one flow or multiple flows and / or blocks Figure 1 or multiple blocks.

[0050] Although the specific embodiments of the disclosure have been described above in conjunction with the accompanying drawings, they are not intended to limit the protection scope of the disclosure. Those skilled in the art should understand that, based on the technical solutions of the disclosure, various modifications or variations that can be made by those skilled in the art without creative efforts are still within the protection scope of the disclosure.

Claims

1. An end-to-end semantic communication method for fishing boat networking in a maritime environment, characterized in that, Including: Obtain ocean buoy data, convert the ocean buoy data into matrix data according to time granularity and spatial granularity to obtain an ocean buoy data matrix, and perform preprocessing; Convert the preprocessed ocean buoy data matrix into an ocean buoy network data tensor, and input it into a semantic communication network for the marine environment to obtain the restored data of end-to-end semantic communication; Among them, after the ocean buoy network data tensor is input into the semantic communication network for the marine environment, the convolutional block is used to extract the feature maps of the ocean buoy network data tensor multiple times, and then the adaptive feature block continuously adjusts the feature maps according to the signal-to-noise ratio to obtain feature vectors. Then, based on the data compression ratio, the feature vectors are compressed by a ratio to obtain semantic feature vectors. The semantic feature vectors are attenuated by the total channel response of the channel model to obtain the final output signal. The final output signal is used to generate the restored data through the inverse inference process of the semantic decoder.

2. The end-to-end semantic communication method for fishing boat networking in a marine environment according to claim 1, wherein Converting the ocean buoy data into matrix data according to time granularity and spatial granularity to obtain an ocean buoy data matrix includes: processing the original ocean buoy data into matrix data according to time granularity and spatial granularity. The time granularity is in hours, and the spatial granularity is in the coordinates of different working stations. And it is divided based on three data types of temperature, wave, and wind to obtain the ocean buoy data matrix.

3. The end-to-end semantic communication method for fishing boat networking in a maritime environment according to claim 1, wherein, The preprocessing process includes: performing normalization, missing value filling, and outlier processing for each ocean buoy data matrix. Among them, the Min-Max normalization method is used to scale the ocean buoy data matrix to the range of [0,1].

4. The end-to-end semantic communication method for fishing boat networking in a marine environment according to claim 1, characterized in that, Convert the normalized ocean buoy data matrix into a tensor to obtain the ocean buoy network data tensor. After inputting it into the semantic communication network for the maritime environment, it first enters the semantic encoder. Based on the convolutional block, the spatial pattern or local correlation of the ocean buoy data matrix is extracted to obtain a feature map. Then, the output feature map is normalized through the generalized division normalization layer, and the output feature map features are non-linearly activated through the PReLU layer and then the features of the output feature map are extracted again. An adaptive feature block is constructed based on calculating the mean of the features and combining the signal-to-noise ratio. The adaptive feature block adjusts the feature map according to the signal-to-noise ratio, re-extracts the feature map, and after N extraction processes, a feature vector is obtained. Then, based on the data compression ratio CR, the feature vector is compressed at a ratio to obtain the semantic feature vector.

5. The end-to-end semantic communication method for fishing boat networking in a maritime environment according to claim 1, characterized in that, Construct a channel model based on the channel state information CSI, and select the Rice channel to simulate the marine environment. The Rice channel is a fading model, and its model assumes that there is a dominant direct path signal, and it also includes multiple path signals generated by reflection, refraction, or scattering of other objects.

6. The end-to-end semantic communication method for fishing boat networking in a maritime environment according to claim 1, characterized in that Attenuating the semantic feature vectors through the total channel response of the channel model includes: determining the gain coefficient of the channel based on the signal-to-noise ratio SNR, simulating the fading of the signal in the Rice channel, dividing the signal into real and imaginary parts to form a complex signal, calculating the direct path signal in complex form, performing random time delay processing on the signals of each scattering path, and adding up the signals of all scattering paths to obtain the final multipath scattering signal. Adding the direct path signal and the multipath scattering signal to obtain the total channel response, attenuating the semantic feature vectors through the total channel response, and then performing normalization processing to obtain the final output signal.

7. An end-to-end semantic communication system for fishing boat networking in a maritime environment, characterized in that, Including: A data acquisition module for obtaining ocean buoy data, converting the ocean buoy data into matrix data according to time granularity and spatial granularity to obtain an ocean buoy data matrix, and performing preprocessing; A semantic restoration module for converting the preprocessed ocean buoy data matrix into an ocean buoy network data tensor, and inputting it into a semantic communication network for the marine environment to obtain the restored data of end-to-end semantic communication; Among them, after the ocean buoy network data tensor is input into the semantic communication network for the marine environment, the convolutional block is used to extract the feature map of the ocean buoy network data tensor multiple times, and then the adaptive feature block continuously adjusts the feature map according to the signal-to-noise ratio to obtain a feature vector. Then, based on the data compression rate, the feature vector is compressed by a ratio to obtain a semantic feature vector. The semantic feature vector is attenuated through the overall channel response of the channel model to obtain the final output signal, and the final output signal generates the restored data through the inverse inference process of the semantic decoder.

8. A computer program product comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the end-to-end semantic communication method for fishing boat networking in a marine environment according to any one of claims 1-6.

9. A non-transitory computer-readable storage medium, characterized in that, The non-transitory computer-readable storage medium is used to store computer instructions. When the computer instructions are executed by a processor, it implements the end-to-end semantic communication method for fishing boat networking in a marine environment according to any one of claims 1-6.

10. An electronic device, characterized in that, Including: A processor, a memory, and a computer program; wherein, the processor is connected to the memory, the computer program is stored in the memory, and when the electronic device runs, the processor executes the computer program stored in the memory, so that the electronic device executes and implements the end-to-end semantic communication method for fishing boat networking in a marine environment according to any one of claims 1-6.

Images