User communication method and system

Abstract

The invention provides a user communication method and system. The method comprises the following steps: inputting at least one first service required by a user into a trained communication model by utilizing a satellite transmitting end to obtain a semantic feature vector corresponding to each first service and a total codebook corresponding to all the first services; determining an index of each semantic feature vector based on the total codebook by using a satellite transmitting end; modulating the index by using a satellite transmitting end, and sending the modulated index to a user of a ground receiving end; and utilizing a ground receiving end to recover and obtain the first service based on the received modulated index. According to the embodiment of the invention, the semantic communication technology and the satellite communication system are fused, the semantic codebook is generated by using the vector quantization technology, and multi-satellite user distinguishing is realized in the model information space.

Description

Technical Field

[0001] This application relates to the field of satellite communication technology, and particularly to a user communication method and system. Background Art

[0002] Due to the uniqueness of satellite channels, satellite communication is significantly different from terrestrial networks, including severe signal attenuation and long-time delay caused by the long distance between satellites and ground devices, as well as strong line-of-sight channel characteristics such as path loss, atmospheric attenuation, and rain fade. In addition, there are high Doppler frequency shifts and co-channel interference in satellite channels. Limited by Shannon's theorem, the above problems are difficult to solve through existing satellite communication paradigms.

[0003] Semantic communication is a brand-new communication paradigm. By integrating artificial intelligence and communication, and through the organic mapping of task requirements and information transmission, it can greatly improve communication efficiency and enhance users' service experience. Traditional communication systems focus on metrics such as channel capacity, bit error rate, and outage probability, which are irrelevant to data content, while semantic communication focuses on the content and meaning of data and customizes different performance metrics according to different data types, such as sentence similarity for text, peak signal-to-noise ratio and structural similarity metrics for images, and distortion rate for voice data. Semantic communication relies on the deep integration of technologies such as intelligence, communication, and network, and realizes the efficient and accurate transmission and precise control of information through semantic feature extraction, semantic information transmission, and semantic information recovery.

[0004] However, in the existing technology, only non-orthogonal multiple access technology is still used to distinguish users, without fully utilizing the source semantic domain features in a higher information dimension and without being able to eliminate the influence between signals of different nodes. Summary of the Invention

[0005] In view of this, the purpose of this application is to propose a user communication method and system.

[0006] Based on the above purpose, this application provides a user communication method, which includes:

[0007] Using a satellite transmitter to input at least one first service required by a user into a trained communication model, obtaining a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all the first services;

[0008] Using the satellite transmitter to determine the index of each semantic feature vector based on the total codebook;

[0009] Using the satellite transmitter to modulate the index and send the modulated index to the user at the ground receiving end;

[0010] Using a ground receiving end, based on the received modulated index, the first service is recovered.

[0011] In a possible implementation manner, the using the satellite transmitting end to input at least one first service required by a user into a trained communication model to obtain a semantic feature vector corresponding to each first service and a total codebook corresponding to all the first services includes:

[0012] Using the satellite transmitting end to input at least one first service required by a user into a trained communication model, and using a semantic encoder to encode the first service to obtain a semantic feature vector corresponding to each first service and a private codebook corresponding to each first service;

[0013] Combining all the private codebooks to obtain the total codebook.

[0014] In a possible implementation manner, the number of codebook vectors of each private codebook is the same; the number of codebook vectors of the total codebook is the same as the number of codebook vectors of the private codebook.

[0015] In a possible implementation manner, the using the satellite transmitting end to determine an index of each semantic feature vector based on the total codebook includes:

[0016] Using the satellite transmitting end to perform vector quantization processing on each semantic feature vector based on the total codebook to determine an index of each semantic feature vector.

[0017] In a possible implementation manner, the performing vector quantization processing on each semantic feature vector based on the total codebook to determine an index of each semantic feature vector includes:

[0018] Calculating the Euclidean distance between each semantic feature vector and each codebook vector of the total codebook;

[0019] Taking the index of the codebook vector of the total codebook corresponding to the minimum Euclidean distance as the index of the corresponding semantic feature vector.

[0020] In a possible implementation manner, the using the satellite transmitting end to modulate the index and sending the modulated index to a user of the ground receiving end includes:

[0021] Modulating the index to obtain the modulated index;

[0022] Sending the modulated index and the private codebook corresponding to the first service to the user of the ground receiving end.

[0023] In a possible implementation, the use of the ground receiving end to recover the first service based on the received modulated index includes:

[0024] Input the received modulated index into an index recovery device to obtain a recovered index;

[0025] Based on the recovered index, retrieve in the private codebook and combine them in order to obtain the semantic feature vector of the first service corresponding to the user;

[0026] Use a semantic decoder to recover the semantic feature vector to obtain the first service.

[0027] In a possible implementation, the training process of the trained communication model includes:

[0028] Determine the user group corresponding to the satellite transmitting end that receives services;

[0029] Based on the user group, train the communication model to be trained and the codebook to obtain the trained communication model and the private codebook;

[0030] Send the trained communication model to the corresponding user group, and send the private codebook to the corresponding users in the user group.

[0031] In a possible implementation, the determination of the user group corresponding to the satellite transmitting end that receives services includes:

[0032] Use the service control center to receive service requests sent by ground users who need services;

[0033] According to the utilization situation of the current satellite resources, the service request, and the service type of the user, determine the service satellite and the user group corresponding to the service satellite that receives services.

[0034] Based on the same inventive concept, an embodiment of the present application further provides a user communication system, including:

[0035] A satellite transmitting end, configured to input at least one first service required by a user into a trained communication model to obtain a semantic feature vector corresponding to each first service and a total codebook corresponding to all first services; based on the total codebook, determine the index of each semantic feature vector; modulate the index, and send the modulated index to the users of the ground receiving end;

[0036] A ground receiving end, configured to recover the first service based on the received modulated index.

[0037] As can be seen from the above, in the user communication method and system provided by the present application, at least one first service required by a user is input into a trained communication model by a satellite transmitter, to obtain a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all of the first services; the satellite transmitter determines an index of each of the semantic feature vectors based on the total codebook; the satellite transmitter modulates the index and sends the modulated index to a user at a ground receiving end; the ground receiving end restores the first service based on the received modulated index. In an embodiment of the present application, an original service data (such as text, image or other form of data) is extracted as a semantic feature vector by a semantic encoder, only the core meaning of the data is retained, and redundant information is removed. This semantic compression mechanism greatly reduces the amount of data, improves the transmission efficiency, and reduces the requirement for communication bandwidth. Vector quantization is performed on the semantic feature vectors by using the total codebook, to represent high-dimensional continuous data as discrete indexes. By transmitting compact indexes instead of complete feature vectors, the amount of data is further compressed and the communication efficiency is optimized. Each user has its own private codebook, and the characteristics of the codebook ensure that data transmission between different users does not interfere with each other. During the transmission process, the codebook mechanism can ensure that the data of each user corresponds one-to-one with its private codebook through the quantized indexes, so as to effectively distinguish between multiple users. Through a pre-trained semantic encoder, decoder and codebook communication model, which are distributed in advance to the satellite and ground user terminals, the computational overhead during real-time transmission is significantly reduced. This mechanism enables the update of the model and the codebook to be independent of the communication process, enhancing the flexibility of the system. The total codebook is generated by aggregating private codebooks of multiple services, and can be compatible with and support various different types of service requirements. This design meets the requirements of multi-service scenarios, serves multiple users under one communication resource, and has good scalability. The private codebook of each user is only distributed to the corresponding user, and other users cannot access it, thereby enhancing the privacy and security of the data. Even if the index data is intercepted, the semantic features cannot be restored without the corresponding private codebook. The private codebook, as a decoding tool exclusive to the user, effectively prevents eavesdropping or interference between different users, and improves the communication security of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the present application or related technologies, the following will briefly introduce the drawings required for use in the description of the embodiments or related technologies. Obviously, the drawings in the following description are only embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0039] Figure 1 It is a schematic flowchart of the user communication method according to an embodiment of the present application;

[0040] Figure 2 Schematic diagram of the detailed process of the user communication method according to the embodiment of the present application;

[0041] Figure 3 Schematic diagram of the information transmission process according to the embodiment of the present application;

[0042] Figure 4 Schematic diagram of the structure of the user communication system according to the embodiment of the present application. Detailed implementation manners

[0043] To make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to specific embodiments and the accompanying drawings.

[0044] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should have the ordinary meanings understood by those of ordinary skill in the field to which the present application belongs. The "first", "second" and similar terms used in the embodiments of the present application do not indicate any order, quantity or importance, but are only used to distinguish different components. The terms such as "include" or "comprise" mean that the elements or objects appearing before this term cover the elements or objects listed after this term and their equivalents, without excluding other elements or objects. The terms such as "connect" or "couple" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The terms such as "upper", "lower", "left" and "right" are only used to represent relative positional relationships, and when the absolute position of the object being described changes, the relative positional relationship may also change accordingly.

[0045] It can be understood that, before using the technical solutions of the various embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved will be informed to the user in an appropriate manner and the user's authorization will be obtained.

[0046] For example, when responding to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be executed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, an application program, a server or a storage medium that executes the operation of the technical solution of the present disclosure according to the prompt message.

[0047] As an optional but non-limiting implementation manner, the way of sending a prompt message to the user in response to receiving an active request from the user may be, for example, in the form of a pop-up window, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0048] It is understandable that the above notification and the process of obtaining user authorization are only illustrative and do not limit the implementation manner of the present disclosure. Other manners that comply with relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0049] As described in the background art section, due to the uniqueness of satellite channels, satellite communication is significantly different from terrestrial networks, including severe signal attenuation and long-time delay caused by the relatively long distance between satellites and ground devices, and strong line-of-sight channel characteristics such as path loss, atmospheric attenuation, and rain fade. In addition, there are high Doppler frequency shifts and co-channel interference in satellite channels. Limited by Shannon's theorem, it is difficult to solve the above problems through existing satellite communication paradigms.

[0050] Semantic communication is a brand-new communication paradigm. By integrating artificial intelligence and communication, and through the organic mapping of task requirements and information transmission, it can greatly improve communication efficiency and enhance the user's service experience. Traditional communication systems focus on metrics that have nothing to do with data content, such as channel capacity, bit error rate, and outage probability, while semantic communication focuses on the content and meaning of data and customizes different performance metrics according to different data types, such as sentence similarity for text, peak signal-to-noise ratio and structural similarity metrics for images, and distortion rate for voice data. Relying on the deep integration of intelligence, communication, network and other technologies, semantic communication realizes the efficient and accurate transmission and precise control of information through semantic feature extraction, semantic information transmission, and semantic information recovery.

[0051] However, in the existing technology, only non-orthogonal multiple access technology is still used to distinguish users, without fully utilizing the source semantic domain features in a higher information dimension and without being able to eliminate the influence between signals of different nodes.

[0052] Considering the above, an embodiment of the present application proposes a user communication method. By using a satellite transmitter to input at least one first service required by a user into a trained communication model, a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all of the first services are obtained; using the satellite transmitter to determine an index of each of the semantic feature vectors based on the total codebook; using the satellite transmitter to modulate the index and send the modulated index to a user at a ground receiving end; using the ground receiving end to recover the first service based on the received modulated index. In the embodiment of the present application, the original service data (such as text, image or other forms of data) is extracted as a semantic feature vector through a semantic encoder, only the core meaning of the data is retained, and redundant information is removed. This semantic compression mechanism greatly reduces the amount of data, improves the transmission efficiency, and reduces the demand for communication bandwidth. The semantic feature vectors are vector quantized using the total codebook, and the high-dimensional continuous data is represented as discrete indexes. By transmitting the compact indexes instead of the complete feature vectors, the amount of data is further compressed and the communication efficiency is optimized. Each user has its own private codebook, and the characteristics of the codebook ensure that data transmission between different users does not interfere with each other. During the transmission process, the codebook mechanism can ensure that the data of each user corresponds one-to-one with its private codebook through the quantized indexes, so as to effectively distinguish between multiple users. Through the pre-trained semantic encoder, decoder and codebook communication model, which are distributed in advance to the satellite and the ground user terminal, the computational overhead during real-time transmission is significantly reduced. This mechanism enables the update of the model and the codebook to be independent of the communication process, enhancing the flexibility of the system. The total codebook is generated by aggregating the private codebooks of multiple services and can be compatible with and support various different types of service requirements. This design meets the requirements of multi-service scenarios, serves multiple users under one communication resource, and has good scalability. The private codebook of each user is only distributed to the corresponding user, and other users cannot access it, thereby enhancing the privacy and security of the data. Even if the index data is intercepted, the semantic features cannot be restored without the corresponding private codebook. The private codebook, as a decoding tool exclusive to the user, effectively prevents eavesdropping or interference between different users and improves the communication security of the system.

[0053] Hereinafter, the technical solutions of the embodiments of the present application will be described in detail through specific embodiments.

[0054] Reference Figure 1 , the user communication method of the embodiment of the present application includes the following steps:

[0055] Step S101, using a satellite transmitter to input at least one first service required by a user into a trained communication model, obtaining a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all of the first services;

[0056] Step S102: The satellite transmitting end determines the index of each semantic feature vector based on the total codebook;

[0057] Step S103: The satellite transmitting end modulates the index and sends the modulated index to the users at the ground receiving end;

[0058] Step S104: The ground receiving end restores the first service based on the received modulated index.

[0059] Reference Figure 2 , which is a detailed flowchart of the user communication method according to the embodiment of the present application.

[0060] As Figure 2 shown, the embodiment of the present application can be divided into three major parts. The first part is the service request. In this part, the satellite-ground network will first confirm the corresponding relationship between the user and the service satellite, and then train the model according to this corresponding relationship. After training, it will be sent to the corresponding service satellite. Then comes the second part, which is the model transmission and codebook allocation. After the aforementioned training is completed, the service satellite transmits the trained model of the receiving end to the corresponding user, and then synchronizes the corresponding private codebook to the corresponding user. After that, information transmission can be carried out. In the third part, the service satellite can transmit the corresponding signal to the user group.

[0061] The following Figure 1 and Figure 2 are used to illustrate the embodiments of the present application in detail.

[0062] Regarding step S101, before the actual operation after the model training in this embodiment is completed, the model needs to be trained first.

[0063] The specific training process is as follows:

[0064] In some embodiments, the training process of the trained communication model includes: determining the user group that the satellite transmitting end corresponds to receive services; training the communication model to be trained and the codebook based on the user group to obtain the trained communication model and the private codebook; sending the trained communication model to the corresponding user group, and sending the private codebook to the corresponding users in the user group.

[0065] In some embodiments, the determining the user group that the satellite transmitting end corresponds to receive services includes: using the service control center to receive the service requests sent by the ground users who need services; determining the service satellite and the user group that receives services corresponding to the service satellite according to the current utilization situation of satellite resources, the service requests, and the service types of the users.

[0066] In this embodiment, multiple terrestrial users in need of services send service requests to the space-ground network (including satellite nodes and terrestrial gateways). After receiving the requests, the service control center, based on the satellite resource utilization, the user and space-ground network topology, and the user service type, confirms the service satellites and the user groups to receive services, and sends link confirmation information to the satellites and users. According to the user service type and the number of users, the transceiver model and the codebook are trained, and the trained model and codebook are sent to the service satellites.

[0067] The space-ground network consists of satellite nodes, terrestrial gateways, and other communication infrastructures. Multiple terrestrial users in need of services send service requests to the space-ground network through communication links. The service requests mainly include information such as the communication requirements of the users (such as bandwidth, latency requirements, etc.), service types (such as data transmission, video streaming, etc.), and user locations.

[0068] The Service Control Center (SCC), as the core scheduling agency of the space-ground network, is responsible for receiving the service requests of users and processing the requests based on the following key information:

[0069] Satellite resource utilization: including the remaining communication bandwidth of the satellite, available computing resources (such as the ability to process data), the current load of the satellite, etc. SCC needs to ensure reasonable resource allocation to avoid overload or resource waste.

[0070] The topology of the user and the space-ground network: the physical location relationship between the satellites and terrestrial users in the space-ground network (such as satellite coverage, user distribution area, etc.). The topology of the space-ground network (such as which terrestrial gateways are connected to which satellites, whether the satellites support relay communication, etc.) determines the feasibility and efficiency of the service.

[0071] User service type: Different service types have different communication requirements. For example: real-time video streaming requires low latency and high bandwidth; ordinary data transmission has lower latency requirements, but may have higher requirements for data integrity. SCC formulates resource allocation strategies according to the priority of service types.

[0072] SCC comprehensively considers the above factors, confirms the service satellites that can provide services (i.e., satellite nodes that provide communication services to users), determines the user groups that can finally receive services, and records this information.

[0073] Furthermore, link confirmation information is sent to the satellites and users. Link confirmation information: specifies the specific parameters for establishing a communication link between the service satellite and the user (such as allocated frequency bandwidth, time slot, transmit power, etc.). Ensure that a reliable communication link can be established between the user and the satellite.

[0074] Sending process: The SCC sends confirmation information to the relevant satellites and users respectively to ensure that both parties can establish communication according to the link allocation scheme. At the same time, the users and satellites need to adjust the configuration of their communication devices based on the link confirmation information to match the corresponding communication conditions.

[0075] Furthermore, according to the user service types and quantities, the transceiver models and codebooks are trained. Transceiver models: In satellite-ground networks, intelligent models are used in communication to handle semantic information or signal transmission problems. The purpose of training the transceiver models is to optimize key steps in communication, such as data encoding and decoding.

[0076] Codebooks: In communication, a "codebook" is used to efficiently represent semantic information or signal features as corresponding encoded forms, making it more convenient for transmission and decoding. The design of the codebook needs to be optimized according to specific service types and the number of users. For example: when the number of users is large, the codebook needs to allocate resources more efficiently to avoid conflicts. Real-time service types may require a codebook structure that enables faster decoding.

[0077] Training method: According to the user service types (such as real-time streaming media, file transfer, etc.) and the number of users, relevant communication data is used to train the transceiver models. Through deep learning or semantic communication technologies, the best transceiver models and private codebooks that can adapt to the current communication requirements are generated. Then, the private codebooks are sent to the corresponding users in the user group, and the trained receiver models are broadcast to the corresponding user group.

[0078] In this embodiment, when the service satellite provides communication services to ground users, it is necessary to ensure that the ground users can correctly decode the data transmitted from the satellite side. For this purpose, the service satellite broadcasts the trained receiver models to all users receiving services. These models include: Index recovery model: used to extract the index information of the transmitted data from the received signal. The index is the corresponding position of the data in the codebook, similar to the page number of a word in a dictionary. This index mechanism can quickly locate the representation of the data in the codebook.

[0079] Semantic decoder model: used to restore the received semantic information to the original data that can be understood by users. The role of the semantic decoder model is to improve the accuracy and efficiency of decoding based on the principle of semantic communication by decoding at the semantic level of the data.

[0080] The service satellite sends the above two models to all ground users receiving services simultaneously in a broadcast form. The characteristic of broadcasting is that it can cover all users with a single transmission, avoiding redundant operations of sending to each user individually, thus saving communication bandwidth and time.

[0081] Since the receiver model is trained based on the satellite and is designed for the entire user group, all users can use the same model for decoding. The benefits of this design are: Reduced complexity: User terminals do not need to deploy different models separately. Improved compatibility: All users use the same standard receiver model to ensure consistency in the decoding process.

[0082] Furthermore, a one-to-one private codebook is allocated and sent to each user.

[0083] Private codebook: A private codebook is a unique coding rule designed for each user. Each user's private codebook is different. This mechanism ensures the security and distinctiveness of communication:

[0084] Since each user's codebook is private, even if other users intercept the data, they cannot decode it correctly. In addition, the private codebook can avoid signal confusion between users and ensure that each user only receives the data information they need.

[0085] During communication, the service satellite allocates a private codebook to each user on demand to ensure that each user's data stream can be accurately distinguished during transmission. When the user receives the data, it will use its own private codebook to decode it and restore the encoded data to understandable information.

[0086] After broadcasting the receiving end model, the service satellite will allocate a private codebook to each user and send the corresponding codebook to the user. Each user's codebook is transmitted to the corresponding user through point-to-point transmission, rather than through broadcasting. The reason for this is that the private codebook is exclusive and only applies to a specific user, so it needs to be sent in a point-to-point manner. This can prevent other users from receiving codebooks that do not belong to them, thereby ensuring the privacy and security of communication.

[0087] After the model training is completed, with respect to step S101, the model can be officially put into use.

[0088] refer to Figure 3 , which is a schematic diagram of the information transmission process of an embodiment of the present application.

[0089] In some embodiments, the method of using a satellite transmitter to input at least one first service required by a user into a trained communication model to obtain a semantic feature vector corresponding to each of the first services and a total code book corresponding to all of the first services includes: using a satellite transmitter to input at least one first service required by a user into a trained communication model, encoding the first services using a semantic encoder to obtain a semantic feature vector corresponding to each of the first services and a private code book corresponding to each of the first services; and merging all of the private code books to obtain the total code book.

[0090] In some embodiments, the number of codebook vectors of each of the private codebooks is the same; the number of codebook vectors of the total codebook is the same as the number of codebook vectors of the private codebooks.

[0091] In this embodiment, in a specific embodiment, assume that two users need to obtain different remote sensing image data from a satellite. The requirement of the first user is the topographic information of a certain area, and the requirement of the second user is the vegetation coverage information of the same area. To efficiently transmit this data, the satellite transmitter will use a trained communication model, including a semantic encoder, a decoder, and a codebook, to process the user's requirement data into semantic features and perform quantization processing.

[0092] Specifically, the satellite transmitter first receives the user's task requests, including the topographic service data of the first user and the vegetation coverage service data of the second user. Subsequently, these service data are sequentially input into the trained semantic encoder. The semantic encoder will extract the semantics of the service data of each user and generate corresponding semantic feature vectors. For example, after encoding the topographic data of the first user, a set of semantic feature vectors describing the topographic features is generated; and after encoding the vegetation coverage data of the second user, a set of semantic feature vectors describing the vegetation features is generated.

[0093] Next, while generating the semantic feature vectors, the semantic encoder also generates a private codebook for each user. The private codebook consists of a set of codebook vectors and is used to represent the semantic feature space of the user. For example, the private codebook of the first user may contain 128 codebook vectors, and the dimension of each codebook vector is 64; the private codebook of the second user also contains 128 codebook vectors, and the dimension is also 64. These private codebooks are trained to compactly represent the diversity of user semantic features.

[0094] Then, the satellite transmitter combines the private codebooks of the first user and the second user to generate a total codebook. The combination method is to stack multiple private codebooks in the vector dimension to form a unified codebook that is compatible with the needs of all users. For example, the combined total codebook still contains 128 codebook vectors, but the dimension of each codebook vector rises to 128 to match the complete dimension of the semantic feature vectors.

[0095] In some embodiments, for step S102, the satellite transmitter determines the index of each of the semantic feature vectors based on the total codebook.

[0096] In some embodiments, the satellite transmitter determines the index of each semantic feature vector based on the total codebook, including: using the satellite transmitter to perform vector quantization processing on each semantic feature vector based on the total codebook to determine the index of each semantic feature vector.

[0097] In some embodiments, the performing vector quantization processing on each semantic feature vector based on the total codebook to determine the index of each semantic feature vector includes: calculating the Euclidean distance between each semantic feature vector and each codebook vector of the total codebook; using the index of the codebook vector of the total codebook corresponding to the minimum Euclidean distance as the index of the corresponding semantic feature vector.

[0098] In a feasible embodiment, taking remote sensing image data as an example, the data that the satellite transmitter needs to transmit is the image information required by N users. Through a semantic encoder with learnable parameters, the original image is mapped from the original color space to a semantic latent space, represented as several semantic feature vectors, and the dimension of the semantic feature vector is M.

[0099] To make full use of background knowledge and save bandwidth, vector quantization technology based on a codebook is used to quantize semantic feature vectors, and multiple pre-trained private codebooks are required. The codebook is represented as an ordered sequence of multiple determined codebook vectors. The number of codebook vectors of each private codebook is the same, and the dimension of the codebook vector is 1 / N of the dimension of the semantic feature vector. To ensure the stable transmission of information of multiple users under the same communication resource, multiple private codebooks are merged in the vector dimension. The number of codebook vectors of the merged total codebook is the same as that of the private codebook, and the dimension of the codebook vector is the same as the dimension of the semantic feature vector.

[0100] After forming the total codebook, at the satellite transmitter, calculate the Euclidean distance between each semantic feature vector and the total codebook vector, and use the index of the total codebook vector with the closest Euclidean distance to each semantic feature vector as the signal to be modulated.

[0101] Based on the index sequence corresponding to the semantic feature, perform modulation and use it as the channel input.

[0102] In this embodiment, after having the total codebook, the satellite transmitter performs vector quantization processing on the semantic feature vectors of all users. The specific operation is to calculate the Euclidean distance between each semantic feature vector and all codebook vectors in the total codebook, select the codebook vector with the closest distance as the quantization result, and record its corresponding index at the same time. For example, a set of terrain semantic feature vectors of the first user may be quantized into the index sequence [5, 23, 78], while the vegetation feature vectors of the second user may be quantized into the index sequence [12, 45, 67].

[0103] Further, for steps S103 and S104, in some embodiments, the modulating the index by the satellite transmitter and sending the modulated index to the user of the ground receiver includes: modulating the index to obtain the modulated index; and sending the modulated index and the private codebook corresponding to the first service to the user of the ground receiver.

[0104] In some embodiments, the using the ground receiver to recover the first service based on the received modulated index includes: inputting the received modulated index into an index recovery device to obtain a recovered index; retrieving in the private codebook based on the recovered index, and sequentially combining to obtain the semantic feature vector of the first service corresponding to the user; and using a semantic decoder to recover the semantic feature vector to obtain the first service.

[0105] In this embodiment, during the transmission process, the satellite only needs to send the quantized index sequence to the ground user, and at the same time transmit the private codebook corresponding to each user to the user side. In this way, when the ground user receives the index sequence, they can use their own private codebook to retrieve the corresponding codebook vector according to the index, and recover the complete semantic feature vector by combining these codebook vectors. Finally, the ground user decodes the recovered semantic feature vector into the original service data through a semantic decoder. For example, the first user finally obtains a high-precision terrain image, while the second user obtains an accurate vegetation coverage image.

[0106] This process reflects the efficiency of semantic communication. Through the semantic encoding and quantization of service data, only the index sequence and the codebook need to be transmitted, greatly reducing the amount of data. At the same time, using the privacy and multi-user discrimination ability of the codebook, secure and reliable multi-user separation transmission is achieved.

[0107] In another feasible embodiment, in a satellite remote sensing image transmission task, the satellite needs to transmit large-scale and high-resolution remote sensing image data to multiple ground users. Traditional remote sensing image transmission methods are usually limited by problems such as the bandwidth of the satellite channel, path loss, and multi-user interference, often resulting in large data transmission delays and error rates. This solution significantly improves the transmission efficiency and quality of remote sensing images through the following steps:

[0108] Satellite transmitter processing:

[0109] The semantic encoder on the remote sensing satellite maps the original image data from the pixel space to the semantic latent space, thereby extracting high-level semantic features (such as target shape, material information, etc.). Through the pre-trained total codebook, the semantic features are vector quantized to reduce data redundancy and generate a codebook index sequence. After generating a unified index sequence applicable to all users, the index sequence is channel-encoded and transmitted through channel modulation technology.

[0110] Ground receiving end processing:

[0111] After the ground user receives the modulated signal transmitted by the satellite, the codebook index sequence is recovered from the signal through the index recovery device. Each user uses its private codebook to restore the index sequence to the corresponding semantic feature vector and reconstructs the remote sensing image data through the semantic decoder.

[0112] Through this solution, the transmission efficiency of remote sensing image data is significantly improved. Multiple users are effectively isolated through independent private codebooks, avoiding data cross-interference and ensuring the reliability and security of transmission. It supports satellite channel environments with complex climate conditions and multi-user simultaneous transmission.

[0113] In another feasible embodiment, in the global coverage satellite voice communication service, due to the characteristics of long delay, high attenuation, and Doppler frequency shift in the satellite channel, the traditional waveform- and symbol-based voice communication methods are insufficient in terms of efficiency and anti-interference ability. This solution significantly improves the voice transmission efficiency and anti-interference ability through semantic communication technology. The ground user extracts the key semantic features (such as intonation, lexical semantics, etc.) in the voice data through the semantic encoder, removes the redundant information irrelevant to the voice content, and generates a semantic feature vector. Based on the user-assigned private codebook and the combined total codebook, the semantic features are quantized and the corresponding index sequence is generated for further transmission. The satellite system transmits the index sequence to the target receiving user through channel modulation technology. The receiving end analyzes the index sequence in the signal through the index recovery device and restores the original semantic features using the private codebook. Subsequently, the semantic decoder restores the voice data according to the semantic features to ensure the accurate reconstruction of the voice quality.

[0114] Through this solution, the amount of transmitted data can be significantly reduced, thereby reducing the burden on the satellite channel. It is more robust to channel noise and interference, ensuring high-quality transmission of voice signals. Through the isolation mechanism of the user private codebook, the problem of mutual interference in multi-user communication is avoided.

[0115] In another feasible embodiment, in the ocean resource monitoring scenario, a large number of distributed sensors transmit monitoring data back to the ground control center through the satellite network. The traditional backhaul method is easily affected by satellite bandwidth limitations and sensor data conflicts, making it difficult to meet the requirements of real-time and reliability.

[0116] This solution optimizes the data feedback process of the sensor network through semantic communication and mode division multiple access technology. The environmental data (such as temperature, marine pollutant concentration, etc.) collected by each sensor received by the satellite is first subjected to feature extraction through a semantic encoder, mapping the data into a high-dimensional semantic feature representation. Combining the private codebook assigned to each sensor and the merged total codebook, the semantic features are quantized into a codebook index sequence, compressing the data volume. The receiving end performs semantic decoding according to the codebooks of different sensors, restoring the original data collected by the sensors one by one.

[0117] Through this solution, the data feedback efficiency of the sensors can be significantly improved, supporting the simultaneous online operation of a large number of sensors. By extracting semantic features, the data redundancy is reduced, and the resources occupied by the satellite channel are decreased. The accuracy and real-time performance of data feedback are ensured, adapting to dynamic environmental changes.

[0118] As can be seen from the above embodiments, in the user communication method described in the embodiments of the present application, at least one first service required by a user is input into a trained communication model by a satellite transmitter to obtain a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all of the first services; the satellite transmitter determines an index of each of the semantic feature vectors based on the total codebook; the satellite transmitter modulates the index and sends the modulated index to a user at a ground receiving end; and the ground receiving end restores the first service based on the received modulated index. In the embodiments of the present application, an original service data (such as text, image, or other form of data) is extracted as a semantic feature vector by a semantic encoder, only the core meaning of the data is retained, and redundant information is removed. This semantic compression mechanism greatly reduces the amount of data, improves the transmission efficiency, and reduces the requirement for communication bandwidth. Vector quantization is performed on the semantic feature vectors using the total codebook to represent high-dimensional continuous data as discrete indexes. By transmitting compact indexes instead of complete feature vectors, the amount of data is further compressed and the communication efficiency is optimized. Each user has its own private codebook, and the characteristics of the codebook ensure that data transmission between different users does not interfere with each other. During the transmission process, the codebook mechanism can ensure that the data of each user corresponds one-to-one with its private codebook through the quantized indexes, so as to effectively distinguish between multiple users. Through a pre-trained semantic encoder, decoder, and codebook communication model, which are distributed to the satellite and ground user terminals in advance, the computational overhead during real-time transmission is significantly reduced. This mechanism enables the update of the model and the codebook to be independent of the communication process, enhancing the flexibility of the system. The total codebook is generated by aggregating the private codebooks of multiple services and can be compatible with and support multiple different types of service requirements. This design meets the requirements of multi-service scenarios, serves multiple users under one communication resource, and has good scalability. The private codebook of each user is only distributed to the corresponding user, and other users cannot access it, thereby enhancing the privacy and security of the data. Even if the index data is intercepted, the semantic features cannot be restored without the corresponding private codebook. The private codebook, as a decoding tool exclusive to the user, effectively prevents eavesdropping or interference between different users and improves the communication security of the system.

[0119] It should be noted that the method in the embodiments of the present application can be executed by a single device, such as a computer or a server. The method in this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only execute one or more steps in the method in the embodiments of the present application, and these multiple devices will interact with each other to complete the described method.

[0120] It should be noted that some embodiments of the present application are described above. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the above embodiments and still achieve the desired result. Additionally, the processes depicted in the figures do not necessarily require the particular order or sequential order shown to achieve the desired result. In certain embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0121] Based on the same inventive concept, corresponding to any of the above method embodiments, the present application further provides a user communication system

[0122] Reference Figure 4 , the user communication system includes:

[0123] A satellite transmitting end 41, configured to input at least one first service required by a user into a trained communication model, obtain a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all of the first services; based on the total codebook, determine an index of each of the semantic feature vectors; modulate the index, and send the modulated index to a user at a ground receiving end;

[0124] A ground receiving end 42, configured to recover the first service based on the received modulated index.

[0125] The system of the above embodiment is used to implement the corresponding user communication method in any of the foregoing embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated herein.

[0126] Those of ordinary skill in the art should understand that: the discussion of any of the above embodiments is only exemplary, and is not intended to imply that the scope of the present application (including the claims) is limited to these examples; under the idea of the present application, the technical features in the above embodiments or different embodiments can also be combined, the steps can be implemented in any order, and there are many other variations in different aspects of the embodiments of the present application as described above, and they are not provided in detail for the sake of brevity.

[0127] In addition, for simplicity of explanation and discussion, and so as not to make the embodiments of the present application difficult to understand, well-known power / ground connections to integrated circuit (IC) chips and other components may or may not be shown in the provided drawings. Further, the devices may be shown in block diagram form in order to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that details regarding the implementation of these block diagram devices are highly dependent on the platform on which the embodiments of the present application are to be implemented (i.e., these details should be fully within the understanding of those skilled in the art). In cases where specific details (such as circuits) are set forth to describe exemplary embodiments of the present application, it will be apparent to those skilled in the art that the embodiments of the present application may be practiced without these specific details or with variations of these specific details. Accordingly, these descriptions should be considered illustrative rather than restrictive.

[0128] Although the present application has been described in connection with specific embodiments thereof, many alternatives, modifications, and variations of these embodiments will be apparent to those of ordinary skill in the art based on the foregoing description. For example, other memory architectures (such as dynamic RAM (DRAM)) may be used with the embodiments discussed.

[0129] Embodiments of the present application are intended to cover all such alternatives, modifications, and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the embodiments of the present application shall be included within the protection scope of the present application.

Claims

1. A user communication method, characterized in that: The method comprises: Using a satellite transmitting end, inputting at least one first service required by a user into a trained communication model to obtain a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all of the first services; Determine the index of each of the semantic feature vectors based on the total codebook using a satellite transmitting end; Modulating the index using a satellite transmitter, and sending the modulated index to a user at a ground receiving end; Using the ground receiving end, the first service is recovered based on the received modulated index.

2. The method according to claim 1, characterized in that The step of inputting at least one first service required by a user into a trained communication model using a satellite transmitting end to obtain a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all the first services includes: Using a satellite transmitting end, inputting at least one first service required by a user into a trained communication model, encoding the first service using a semantic encoder, and obtaining a semantic feature vector corresponding to each first service and a private codebook corresponding to each first service; All the private codebooks are combined to obtain the total codebook.

3. The method according to claim 2, characterized in that The number of codebook vectors in each of the private codebooks is the same; the number of codebook vectors in the total codebook is the same as the number of codebook vectors in the private codebook.

4. The method according to claim 1, characterized in that: The using the satellite transmitting end to determine the index of each of the semantic feature vectors based on the total codebook includes: Using the satellite transmitting end, vector quantization processing is performed on each of the semantic feature vectors based on the total codebook to determine an index of each of the semantic feature vectors.

5. The method according to claim 4, characterized in that The performing vector quantization processing on each of the semantic feature vectors based on the total codebook to determine the index of each of the semantic feature vectors includes: Calculating the Euclidean distance between each of the semantic feature vectors and each codebook vector of the total codebook; The index of the codebook vector of the total codebook corresponding to the minimum Euclidean distance is used as the index of the corresponding semantic feature vector.

6. The method according to claim 2, characterized in that The step of modulating the index by using a satellite transmitting end and sending the modulated index to a user at a ground receiving end includes: Modulating the index to obtain the modulated index; The modulated index and the private codebook corresponding to the first service are sent to a user at a ground receiving end.

7. The method according to claim 6, characterized in that The method of using the ground receiving end to recover the first service based on the received modulated index includes: Inputting the received modulated index into an index restorer to obtain a restored index; Based on the restored index, searching in the private codebook, and combining in order to obtain a semantic feature vector of the first service corresponding to the user; The semantic feature vector is restored by using a semantic decoder to obtain the first service.

8. The method according to claim 1, characterized in that: The training process of the trained communication model includes: Determine the user group that receives the service from the satellite transmitter; Training the communication model and the codebook to be trained based on the user group to obtain the trained communication model and the private codebook; The trained communication model is sent to a corresponding user group, and the private codebook is sent to corresponding users in the user group.

9. The method according to claim 8, characterized in that Determining the user group that the satellite transmitting end corresponds to and receives the service includes: Utilize the service control center to receive service requests sent by ground users who need services; According to the current utilization of satellite resources, the service request and the service type of the user, a service satellite and the user group receiving the service corresponding to the service satellite are determined.

10. A user communication system, characterized in that: include: The satellite transmitting end is configured to input at least one first service required by a user into a trained communication model, obtain a semantic feature vector corresponding to each of the first services and a total codebook corresponding to all of the first services; and determine an index of each of the semantic feature vectors based on the total codebook; Modulating the index, and sending the modulated index to a user at a ground receiving end; The ground receiving end is configured to recover the first service based on the received modulated index.

Images