CN116073889B - Satellite communication network architecture based on semantic content - Google Patents

Abstract

The invention relates to a satellite communication network architecture based on semantic content, comprising: the application layer comprises a semantic layer, the semantic layer is used for extracting semantic information and semantic tags of satellite communication data, and the application layer, the transmission layer and the network layer generate semantic content data packets according to the semantic information and the semantic tags; the system comprises at least one virtual anchor point, wherein the virtual anchor point comprises at least one satellite and a resource management node, the resource management node is used for managing each satellite in the virtual anchor point, and each satellite is used for acquiring a semantic content data packet and communicating based on a semantic content data packet protocol. The satellite communication network architecture realizes the functions of distributed caching and collaborative storage among the satellite groups, has high resource utilization rate of the space-based satellite communication network, can improve the transmission efficiency of space-based information, the fault tolerance rate of space-based communication and the service efficiency, and realizes the high-efficiency and high-quality communication of the space-based satellite communication network.

Description

A Semantic Content-Based Satellite Communication Network Architecture

technical field

The invention mainly relates to the technical field of satellite communication network architecture, in particular to a satellite communication network architecture based on semantic content.

Background technique

Space-ground integrated information network is the key research direction of current communication network technology. Based on space-based satellite communication network, space-based communication network and ground-based land communication network, the integration and coordination of space-space-ground information network can be realized, and the space-space-ground integrated information network can be formed to realize the realization of all things. Interconnected global communication network.

Space-based satellite communication network has become an important part of modern communication methods because of its wide coverage and long communication distance. However, with the development of mobile communications, especially 5G+ technology and 6G technology, the demand for communication speed and communication data volume continues to increase, and it is constantly approaching the communication theoretical limit of information theory. Space-based satellite communication networks have limited communication resources and actual utilization of spectrum resources. Problems such as low bandwidth, long user request delay, and difficult inter-satellite management make it difficult for space-based satellite communication networks to meet the growing needs of current mobile communication services. How to build a space-based communication network architecture, improve the utilization rate of space-based satellite communication network resources, realize efficient and high-quality communication in space-based satellite communication networks; realize reasonable allocation and full utilization of space-based resources, reduce user request delays, and reduce inter-satellite communications. Multi-hop transmission, caching and distributing service content on demand; realizing collaborative management and deployment of space-based relay nodes in space-based satellite communication networks is an urgent problem to be solved in the development of space-based satellite communication networks.

The current semantic communication technology is proposed to improve the utilization of communication bandwidth to save communication resources, such as cloud computing and fog computing, which can realize the delay optimization of space-based satellite communication network. The solution to the space-based satellite communication network is mainly based on network reconstruction, and the problem of collaborative management of the space-based satellite network is solved by designing virtualized nodes to implement routing algorithms. Patent CN114615499A provides a semantic optical communication system and method for image transmission. The system provides a change module and a picture coding module. The first change module increases the number of picture channels, cuts the picture, and uses the picture coding module to change the picture. Semantic coding is performed, and after semantic decoding is performed at the receiving end, the image is transmitted through the semantically based image restored by the transformation module. Patent CN113315972A proposes a semantic-based video semantic communication method. By constructing a multi-layer signal perception network and semantic abstraction network, the structured semantic features of the video signal are extracted based on the signal perception network and semantic abstraction network in the hierarchical knowledge base, and then according to the structure Semantic features use the semantic reconstruction network and signal reconstruction network in the hierarchical knowledge base to reconstruct the video signal. By mining the semantic features of different scales and using the structured data structure to represent the semantics, a semantic-based video semantic communication method is realized.

Patent CN109525304A provides a spatial intelligent network architecture integrating perception, computing and storage, compatible with IP (Internet Protocol) networks, realizing the integration of data transmission and content perception, and cross-layer collaborative services between applications and networks. Provide a space router with content awareness, detect content in content storage, realize storage and transmission in the network, reduce transmission delay, resource and energy consumption. Patent CN107343025A provides a distributed satellite cloud network architecture and a time delay optimization method under energy consumption constraints. The distributed satellite cloud network architecture includes a satellite fog layer, an access layer and a cloud computing layer. Under the energy consumption constraint, by constructing the undirected graph of the distributed satellite cloud network architecture, constructing the delay optimization strategy model under the energy consumption constraint, and determining the delay optimization strategy under the energy consumption constraint.

Patent CN110012558A provides a satellite network architecture with network reconfiguration capability, including space-based network and terrestrial backbone network. In the space-based network, geosynchronous orbit satellites form the space-based backbone network, low-orbit satellite constellations form the space-based access network, and remote sensing satellites, meteorological satellites, and navigation satellites form the sky-sensing network, which solves the problem of joint high-, medium-, and low-orbit satellites. Network integration and utilization, communication, navigation, and remote sensing satellites work together to solve the problem of space-based information network architecture design, realize the integration and utilization of various satellite resources in the existing space, and build a hybrid heterogeneous and flexible satellite network reconstruction Functional satellite network architecture. Patent CN113543173A provides a network element deployment architecture and network element deployment method for a satellite 5G converged network, including access network centralized units deployed on high-orbit satellite nodes and access network distributed units deployed on low-orbit satellite nodes. Service collaboration and migration are realized between satellite nodes through laser or microwave links. Through the high-low rail collaborative networking architecture, the efficient use of network resources is realized and the network openness capability is provided.

The satellite communication network architecture in the prior art still has the problems of low space-based information transmission efficiency and low quality of space-based satellite communication.

Contents of the invention

The technical problem to be solved in this application is to provide a satellite communication network architecture based on semantic content. In this network architecture, the utilization rate of space-based satellite communication network resources is high, which can improve the efficiency of space-based information transmission, and realize the efficient and efficient space-based satellite communication network. High quality communication.

The technical solution adopted by this application to solve the above technical problems is a satellite communication network architecture based on semantic content, including: application layer, transport layer and network layer, the application layer includes a semantic layer, and the semantic layer is used to extract satellite communication data Semantic information and semantic tags, the application layer, transport layer and network layer generate semantic content data packets according to the semantic information and semantic tags; at least one virtual anchor point, the virtual anchor point includes at least one satellite and selects a satellite as a resource management node , the resource management node is used to manage each satellite in the virtual anchor point, and each satellite is used to obtain a semantic content data packet and communicate based on a semantic content data packet protocol.

In an embodiment of the present application, the semantic layer includes: a semantic information knowledge base, the semantic layer extracts the semantic information of the satellite communication data according to the semantic information knowledge base; the semantic feature knowledge base is used to extract the semantic features of the satellite communication data; self-attention A force mechanism model for compressing semantic features to generate semantic labels.

In an embodiment of the present application, the satellite communication data includes images, and the semantic layer extracts the semantic information and semantic tags of the satellite communication data, including: a semantic information extraction step: using a neural network model to divide the image to obtain multiple regions, and Calculate the semantic feature importance of each region. If the semantic feature importance is greater than or equal to the preset threshold, all the semantic information of the region will be retained; if the semantic feature importance is less than the preset threshold, the semantic information of the region will be partially retained; semantic label extraction Steps: Use the semantic feature knowledge base to generate the semantic features of the image. The semantic features include the descriptive language for the image; the self-attention mechanism model compresses the semantic features to generate semantic labels. The steps include: use the following formula to calculate each word x in the semantic features _i 's attention value Attention(xi ₎ :

Among them, N is the length of the descriptive language, is the logarithmic representation of the word x _i , and Value _i represents the weight index of the word x _i ; if the attention value Attention(xi ₎ is greater than or equal to the preset value, the word x _i is reserved as a semantic label.

In an embodiment of the present application, the step of generating the semantic content data packet according to the semantic information and the semantic label by the application layer, the transport layer and the network layer includes: the application layer obtains the semantic information and the semantic label extracted by the semantic layer; As data bits are packaged, the application layer header and the transport layer header are added; the network layer packs the semantic label, and the network layer header is added to complete the packaging of the semantic content data packet.

In an embodiment of the present application, the steps for each satellite to communicate based on the semantic content data packet protocol include: the satellite receives the semantic content data packet, unpacks the semantic content data packet at the network layer, and obtains the semantic label; continues to unpack the semantic content data packet at the transport layer; Package semantic content data package to obtain semantic information; cache semantic information at the application layer.

In an embodiment of the present application, the network architecture further includes a satellite communication management module and a client, and the satellite communication management module manages the satellite communication step including a request matching step, and the request matching step includes: when the semantic content data packet requested by the client is routed When arriving at a certain satellite, the satellite forwards the semantic content data packet to the resource management node in the current virtual anchor; the resource management node unpacks the semantic content data packet at the network layer to obtain the semantic label, and the resource management node judges the current virtual anchor according to the semantic label Whether the point can provide services; if the current virtual anchor point cannot provide services, the resource management node transmits the semantic content data packet to another virtual anchor point; if all virtual anchor points cannot provide services, the semantic content data packet is transmitted to the ground Service content data center.

In an embodiment of the present application, the step of the resource management node judging whether the current virtual anchor can provide services according to the semantic tags includes: the resource management node obtains the user request semantic features according to all user request data in the current virtual anchor service area, And generate the common request label Comm_Tag; the resource management node uses the following formula to calculate the user request semantic label User_Req=(RTag ₁ , RTag ₂ ,...,RTag _M ) and the common request label Comm_Tag=(CTag ₁ , CTag ₂ , .. ., CTag _N ) matching value Match_Lev1:

Match_Lev1 = Matching(Comm_Tag, User_Req)

Among them, RTag _i represents the content of semantic tags requested by users, M represents the number of semantic tags requested by users, CTag _i represents the content of common request tags, and N represents the number of common request tags; the resource management node judges whether the matching degree value Match_Lev1 is greater than or equal to the preset Set the matching degree baseline value Match_BaselL1; if the matching degree value Match_Lev1 is greater than or equal to the preset matching degree baseline value Match_BaseL1, perform data matching with the corresponding service content in the current virtual anchor point, and calculate the matching degree value Match_Lev2 for the service content; the resource management node uses _The following formula calculates the semantic tag Serv_Tag= _{( STag 1} , STag ₂ ,..., STag _G ) matching value Match_Lev2:

Match_Lev2 = Matching(User_Req, Serv_Tag)

Among them, RTag _i represents the content of the semantic tag requested by the user, H represents the number of semantic tags requested by the user, STag _i represents the content of the semantic tag of the service content, G represents the number of semantic tags of the current service content; the resource management node judges the content of the service content Whether the matching degree value Match_Lev2 is greater than or equal to the preset matching degree baseline value Match_BaseL2; if the matching degree value Match_Lev2 is greater than or equal to the preset matching degree baseline value Match_BaseL2, the current virtual anchor will provide services; if the matching degree value Match_Lev1 is less than the preset matching degree If the baseline value is Match_BaseL1, the current virtual anchor cannot provide services, and user request data needs to be routed to subsequent virtual anchors.

In an embodiment of the present application, the step of managing satellite communication by the satellite communication management module further includes a demand matching step, and the demand matching step includes: in the process of returning the semantic content data packet for providing the service to the client, when used to provide When the semantic content data packet of the service is routed to a satellite, the satellite forwards the semantic content data packet used to provide the service to the resource management node in the current virtual anchor; the resource management node unpacks the semantic content used to provide the service at the network layer The data packet is used to obtain the service content semantic label used to provide services, and the resource management node judges whether it is necessary to cache the semantic content data package used to provide services according to the service content semantic label; if caching is required, the resource management node will use A copy of the semantic content package is cached to any satellite in the current virtual anchor.

In an embodiment of the present application, the resource management node judges whether it is necessary to cache the semantic content data packets used to provide the service according to the semantic label of the service content, including: judging whether the semantic label of the service content is the current The anchor point has requested, and judge whether the relevant semantic tags have been cached and have certain content in the current virtual anchor point. If there is no cache, then directly cache the semantic content data packet used to provide services; and the resource management node serves All users in the region request data, get the semantic features of user requests, and generate the common request tag Comm_Tag; the resource management node uses the following formula to calculate the matching value Match_Lev3 between the service content semantic tag Serv_Tag and the common request tag Comm_Tag:

Match_Lev3 = Matching(Comm_Tag, Serv_Tag)

The resource management node judges whether the matching degree value Match_Lev3 is greater than or equal to the preset matching degree baseline value Match_BaseL3; if the matching degree value Match_Lev3 is greater than or equal to the preset matching degree baseline value Match_BaseL3, it needs to cache the semantic content data packets used to provide services; if the matching degree If the value Match_Lev3 is smaller than the preset matching degree baseline value Match_BaseL3, then there is no need to cache the semantic content data packets for providing services.

In an embodiment of the present application, a high-orbit satellite is set as a resource management node in the virtual anchor point.

In an embodiment of the present application, any low-orbit satellites are preset to form a resource management node queue, and the low-orbit satellites in the resource management node queue can move synchronously. During the movement of the low-orbit satellites to another virtual anchor point, The low-orbit satellites in the resource management node queue are always in one-to-one correspondence with the virtual anchor points, and the low-orbit satellites in one-to-one correspondence with the virtual anchor points are used as resource management nodes.

In an embodiment of the present application, the satellite communication network architecture further includes a satellite movement management module, and the steps of the satellite movement management module managing the satellite movement include: when a certain satellite leaves the current virtual anchor point and enters another virtual anchor point, resource management The node distributes the cached content of the satellite to at least one satellite in the current virtual anchor point; and the resource management node of another virtual anchor point judges the matching degree between the cached content and the virtual anchor point; if the matching degree is high, the satellite retains its own To cache content, the resource management node of another virtual anchor distributes service content based on the cache space of the satellite; if the matching degree is low, the satellite deletes its own cache content, and the resource management node of another virtual anchor distributes a copy of the corresponding service content to satellite.

The technical solution of this application builds a satellite communication network architecture including application layer, transport layer and network layer, and embeds a semantic layer based on semantics in the application layer, and the semantic layer implements semantic information extraction and restoration of satellite communication data in the application layer and the semantic tag extraction function; this application generates semantic content data packets based on the extracted semantic information and semantic tags in the application layer, transport layer and network layer, and designs a semantic content data packet protocol for satellite communications, and implements the protocol The lightweight design and transparent transmission can save the amount of space-based communication data and improve the utilization rate of communication bandwidth; this application designs a virtual anchor point based on the service area, and the satellites in the virtual anchor point form a satellite cluster to provide collaborative services for the virtual anchor point , the resource management node is introduced into the virtual anchor point to manage the satellite cluster. The resource management node can match the semantic content data packets between the satellites as needed, realize the distributed cache and collaborative storage function between the constellations, and save the cache resources of the space-based communication network. Improve the collaborative processing capability between constellations, reduce service content distribution delay, and reduce inter-satellite routing hops; the satellite communication network architecture of this application can improve the overall efficiency of space-based information transmission, save space-based communication and cache resources, and improve space-based communication. Fault tolerance rate and service efficiency.

Description of drawings

In order to make the above-mentioned purposes, features and advantages of the present application more obvious and understandable, the specific implementation methods of the present application will be described in detail below in conjunction with the accompanying drawings, wherein:

Fig. 1 is an exemplary architecture diagram of a satellite communication network architecture based on semantic content according to an embodiment of the present application;

FIG. 2 is an exemplary architecture diagram of a semantic content-based satellite communication network architecture according to another embodiment of the present application;

Fig. 3 is an exemplary flow chart of generating semantic content data packets and satellite-based semantic content data packet protocol communication in an embodiment of the present application;

Fig. 4 is an exemplary flow chart of request matching for semantic content data packets in an embodiment of the present application;

Fig. 5 is an exemplary flowchart of performing requirement matching on semantic content data packets in an embodiment of the present application;

Fig. 6 is an exemplary flow chart of satellite movement in different virtual anchor points and movement of satellite cache content in an embodiment of the present application.

Detailed ways

In order to make the above objects, features and advantages of the present application more comprehensible, the specific implementation manners of the present application will be described in detail below in conjunction with the accompanying drawings.

Many specific details are set forth in the following description to facilitate a full understanding of the present application, but the present application can also be implemented in other ways than those described here, so the present application is not limited by the specific embodiments disclosed below.

As indicated in this application and claims, the terms "a", "an", "an" and/or "the" do not refer to the singular and may include the plural unless the context clearly indicates an exception. Generally speaking, the terms "comprising" and "comprising" only suggest the inclusion of clearly identified steps and elements, and these steps and elements do not constitute an exclusive list, and the method or device may also contain other steps or elements.

The flow chart is used in this application to illustrate the operations performed by the system according to the embodiment of this application. It should be understood that the preceding or following operations are not necessarily performed in an exact order. Instead, various steps may be processed in reverse order or concurrently. At the same time, other operations are either added to these procedures, or a certain step or steps are removed from these procedures.

This application proposes a satellite communication network architecture based on semantic content, which can be applied to communication scenarios between space-based satellites, and can also be applied to communication scenarios between space-based satellites and ground users.

Fig. 1 is an exemplary architecture diagram of a satellite communication network architecture based on semantic content according to an embodiment of the present application. Referring to Fig. 1, the satellite communication network architecture based on semantic content includes: application layer 301, transport layer 302 and network layer 303, the application layer 301 includes a semantic layer 3011, the semantic layer 3011 is used to extract the semantic information and semantic tags of the satellite communication data, the application layer 301, the transport layer 302 and the network layer 303 generate a semantic content data packet 304 according to the semantic information and semantic tags ; At least one virtual anchor point (such as virtual anchor point 100), the virtual anchor point 100 includes at least one satellite (such as satellite 1 ~ satellite N) and selects a satellite as the resource management node 1001, the resource management node 1001 is used to manage the virtual Each satellite in the anchor point 100, each satellite is used to obtain the semantic content data packet 304, and communicate based on the semantic content data packet protocol. Exemplarily, when only one satellite is included in the virtual anchor point, the satellite needs to have the function of a resource management node, and this application does not limit the number of satellites included in the virtual anchor point.

For example, the space-based satellite communication network covers a large service area and involves a large number of space-based satellites. Therefore, this application proposes to divide the service area and design a virtual anchor point scheme based on the service area. Each virtual anchor point serves a certain area. Because the low-orbit satellites will move at high speed relative to the ground, the addressing of the satellite communication network of the present application is realized based on virtual anchor points. The minimum number of satellites in a cluster is one). In the satellite cluster corresponding to each virtual anchor point, there is a satellite node as the resource management node of the virtual anchor point. The resource management node is responsible for storing the content of the satellite node in the virtual anchor point and entering the satellite cluster during the inter-satellite transmission relay process. The content of the node and the needs of the region corresponding to the virtual anchor point are managed. The introduction of virtual anchor resource management nodes effectively solves the problem of distributed cache and management of virtual anchor satellite clusters, and further improves the utilization rate of space-based resources.

Exemplarily, when the semantic content data packet 304 is routed to a virtual anchor point in the space-based communication network architecture of the present application, the space-based satellite in the virtual anchor point forwards the semantic content data packet 304 to the current virtual anchor point The current resource management node, after unpacking the semantic content data packet at the network layer 303 at the network layer, the current resource management node realizes requirement matching based on the semantic tag, and implements a copy cache function for the service content based on the requirement matching result and the caching strategy. When performing the copy caching function, the resource management node of the virtual anchor caches the current copy to any satellite in the virtual anchor satellite cluster. The steps for matching requirements will be described later.

For example, due to the high dynamic characteristics of space-based satellites, when a space-based satellite leaves the current virtual anchor point and enters the next virtual anchor point, the resource management node distributes the cached content of the space-based satellite to other satellite clusters. in the satellite. The data management node of the next virtual anchor point simultaneously allocates a copy of the relevant service content and caches it in the space-based satellite that moves across the virtual anchor point. The motion management of satellites across virtual anchors will be introduced later.

The technical solution of the present application constructs a satellite communication network architecture including an application layer 301, a transport layer 302, and a network layer 303, and embeds a semantic layer 3011 based on semantics in the application layer 301, and the semantic layer 3011 communicates satellite communications in the application layer 301. The data implements semantic information extraction, recovery and semantic label extraction functions; the application generates semantic content data package 304 based on the extracted semantic information and semantic labels in the application layer 301, transport layer 302 and network layer 303, and designs a satellite The semantic content data packet protocol for communication realizes the lightweight design and transparent transmission of the protocol, which can save space-based communication data volume and improve the utilization rate of communication bandwidth; this application designs virtual anchor points 100 and virtual anchor points 100 based on the service area The satellites 1 to N in the network form a satellite cluster to provide collaborative services to the virtual anchor point 100. A resource management node 1001 is introduced into the virtual anchor point 100 to manage the satellite cluster. The resource management node 1001 can match semantic content data packets between satellites as needed 304. Realize the function of distributed caching and cooperative storage among constellations, save space-based communication network caching resources, improve the collaborative processing capability among constellations, reduce service content distribution delay, and reduce the number of inter-satellite routing hops; the satellite communication network of this application The architecture can improve the efficiency of space-based information transmission as a whole, save space-based communication and cache resources, and improve space-based communication fault tolerance and service efficiency.

Fig. 2 is an exemplary architecture diagram of a semantic content-based satellite communication network architecture according to another embodiment of the present application. The satellite communication network architecture of the embodiment in FIG. 2 is firstly introduced here. Referring to FIG. 2, the entire network architecture includes a virtual anchor point A and a virtual anchor point B (the virtual anchor point can be extended to multiple as required). Virtual anchor point A contains a satellite cluster composed of several space-based satellites (such as satellite A1 ~ satellite N1) and a space-based satellite as a resource management node A1001; virtual anchor point B contains several space-based satellites (such as satellite A satellite cluster composed of B1～satellite N2) and a space-based satellite as a resource management node B1001; each satellite node in the virtual anchor point can perform request matching and demand matching, and virtual anchor point A and virtual anchor point B can Perform inter-satellite routing and data migration.

Continuing to refer to Figure 2, each space-based satellite includes a semantic layer 3011 at the application layer 301, and includes a semantic information knowledge base (not shown), a semantic feature knowledge base (not shown) and a self-attention mechanism at the semantic layer 3011 The model (not shown in the figure) performs semantic information and semantic feature extraction on the service content in the semantic layer 3011, and the application layer 301 can cache the service content. The transport layer encapsulation of the semantic content data packet is carried out at the transport layer 302 , and the semantic content data packet transmission layer encapsulation is carried out at the network layer 303 to perform requirement matching and request matching. The ground user terminal 200 requests 101 from the uplink user of the space-based satellite communication network; the service content data 1011 of the space-based satellite communication network comes from the ground service content data center 400 and the service content data 1011 generated by the space-based satellite itself.

In some embodiments, as shown in FIG. 2, the semantic layer 3011 includes: a semantic information knowledge base, the semantic layer 3011 extracts the semantic information of the satellite communication data according to the semantic information knowledge base; the semantic feature knowledge base is used to extract the satellite communication data Semantic features; self-attention mechanism model, used to compress semantic features to generate semantic labels. Exemplarily, the semantic information knowledge base and the semantic feature knowledge base respectively include a neural network model, and the neural network model is trained by the ground terminal and cached in advance in each satellite. In the semantic layer 3011, based on the semantic information knowledge base and the semantic feature knowledge base, the semantic information extraction and semantic feature extraction of the satellite communication data service content are realized respectively. For the extracted semantic features, based on the self-attention (Self-Attention) mechanism model, the semantic features are compressed and extracted, and the specified length features are obtained as the semantic tags of the service content.

In some embodiments, the satellite communication data includes images, and the step of extracting semantic information and semantic labels of the satellite communication data by the semantic layer includes:

Step Sa1, semantic information extraction step: use the neural network model to divide the image to obtain multiple regions, and calculate the semantic feature importance of each region, if the semantic feature importance is greater than or equal to the preset threshold, all the semantic information of the region is retained ; If the semantic feature importance is less than the preset threshold, then partially retain the semantic information of the region;

Step Sa2, semantic label extraction step: use the semantic feature knowledge base to generate the semantic features of the image, the semantic features include the descriptive language of the image; the self-attention mechanism model compresses the semantic features to generate the semantic label. The steps include: using the following formula (1 )～(2) Calculate the attention value Attention( _xi ) of each word x _i in the semantic features:

Among them, N is the length of the descriptive language, is the logarithmic representation of the word _xi , and Value _i represents the weight index of the word _xi ; if the attention value Attention(xi) is greater than or equal to the preset value, the word _xi is reserved as a semantic label.

Exemplarily, since space-based satellites have rich application scenarios and service content media formats, and different application scenarios target different service content forms, the satellite communication network architecture of the present application supports service content in various formats. Here, taking the service content in image format as an example, the semantic information extraction and restoration of images and the extraction of semantic tags are introduced.

In the previous step Sa1, for the semantic information extraction of the image, the image is first divided into several N*N size squares, and for each square, the semantic feature importance of the square is calculated based on the neural network model. For a block with low importance, all its information is retained. For a block with low importance, the largest pixel in the block is calculated, and only the content of this pixel is retained. The square part that retains important information is the semantic information of the image. By retaining the important information and blurring the unimportant information, the semantic information extraction of the image is realized. In the subsequent recovery process, for the important semantic information part of the retained service content, it is only necessary to restore the service content based on the semantic information; for the low-importance information, it is sufficient to restore the reserved part, thereby realizing the semantic information of the image Extract and restore functions.

In the previous step Sa2, for the semantic label extraction of the image, a descriptive language is generated for the image based on the image identification scheme, that is, the semantic feature of the image is generated based on the semantic feature knowledge base. For example, for an image, the generated semantic feature may be "a gray rabbit is jumping and grazing on the grass in the wild". For the generated semantic features, based on the self-attention mechanism, multiple semantic label extractions are performed on the semantic features, and the semantic features generated by the semantic feature knowledge base are used as the input of the self-attention mechanism model. Perform word segmentation to obtain the word sequence Source_Word[x ₁ , x ₂ , x ₃ , ..., x _N ], calculate the Softmax normalization function of each word x _i , and continue to calculate the attention of word x _i after obtaining the intermediate result a _i Attention value, that is, the semantic feature importance of the word x _i . Based on the calculated attention value Attention(xi ₎ , the semantic importance of each word is determined, thereby extracting the semantic label. Since semantic features are often long in practical applications, in order to realize the lightweight design of semantic tags, this application performs multiple self-attention training on semantic tags to obtain semantic tags of a specified length.

By extracting semantic information from satellite communication data service content, this application can realize the compression of space-based communication network transmission data volume, save communication resources, and improve communication bandwidth utilization; by extracting semantic tags from satellite communication data service content, semantic-based Tag request matching and demand matching functions reduce user request delay and reduce inter-satellite multi-hop transmission.

In some embodiments, as shown in FIG. 1 and FIG. 2, the steps of the application layer 301, the transport layer 302, and the network layer 303 generating a semantic content data packet according to the semantic information and semantic tags include: the application layer 301 obtains the semantic content extracted by the semantic layer 3011 Semantic information and semantic tags; transport layer 302 packs semantic information as data bits, adds application layer header and transport layer header; network layer 303 packs semantic tags, adds network layer header, and completes the packaging of semantic content data packets. Exemplarily, after the space-based semantic layer 3011 extracts the semantic information and semantic tags from the service content, the transport layer 302 packs the semantic information as data bits, and after adding the application layer header and the transport layer header, the network layer 303 performs semantic The label is packaged and the network layer header is added, thus completing the package processing of the semantic content data package. In this application, the semantic information extracted by the semantic layer 3011 is packaged in the transport layer 302, and the semantic label is packaged in the network layer 303, thereby realizing the lightweight and transparent design of the semantic content data package, and the space-based satellite communication network In the network layer 303, the demand matching and request matching functions based on semantic tags are realized to save space-based computing resources.

In some embodiments, as shown in FIG. 1 and FIG. 2 , the steps for each satellite to communicate based on the semantic content packet protocol include: the satellite receives the semantic content packet, unpacks the semantic content packet at the network layer 303, and obtains the semantic content packet. label; continue to unpack the semantic content data packet at the transport layer 302 to obtain semantic information; cache the semantic information at the application layer 301. Exemplarily, the semantic information knowledge base of the semantic layer 3011 supports restoration of service content based on the semantic information. When Tianji receives the semantic content data packet, it decides to directly cache or restore the service content according to the actual needs. Cache to save cache resources; if the service content needs to be restored, after unpacking the semantic content data packet to obtain the semantic information, restore it according to the semantic information knowledge base to obtain the original service content.

Exemplarily, this application designs a semantic content data packet protocol based on semantic content. Based on the semantic information knowledge base and semantic feature knowledge base of the semantic layer 3011, the semantic information and semantic tags of the service content are obtained and packaged. The semantic information is encapsulated in the application layer and the transport layer in sequence at the application layer 301 and the transport layer 302 to obtain a transport layer data packet with the semantic information as data bits. Before the semantic label is added to the transport layer data packet in the network layer 303, a network layer header is added to realize the network layer packet, and the semantic content data packet is completed. The semantic information of the service content realizes the compression of the data volume of the service content, and the semantic information is used as a data bit to realize the lightweight design of the semantic content data package. Semantic tags are packaged at the network layer 303. After the space-based satellite receives the semantic content data packet, it directly obtains the semantic tag when unpacking at the network layer 303, and the space-based satellite directly implements semantic tag-based request matching and demand matching at the network layer 303. Function, no need to continue unpacking at the transport layer 302 and application layer 301, thereby reducing the routing delay of semantic content data packets, saving space-based cache and computing resources, and realizing the transparent design of semantic content data packets.

The semantic content packet protocol of this application is introduced here. Fig. 3 is an exemplary flow chart of generating a semantic content data packet and satellite-based semantic content data packet protocol communication in an embodiment of the present application. Referring to Fig. 3 , in step S310, the first satellite generates a semantic tag for the service content at the application layer and semantic information; in step S320, the first satellite encapsulates semantic information at the transport layer; in step S330, the first satellite encapsulates semantic tags at the network layer; in step S340, the first satellite generates semantic content data packets through inter-satellite The link is transmitted to the resource management node; in step S350 the resource management node unpacks the semantic content data packet at the network layer to obtain the semantic label; in step S360 it is judged whether the request matching is successful, if it is judged as unsuccessful, then in step S361 the resource management node Transmit the semantic content data packet to another virtual anchor point; if judged as successful, then in step S370 the resource management node unpacks the semantic content data packet in the transport layer to obtain semantic information; in step S380 the resource management node caches the semantic information in the application layer .

In some embodiments, as shown in FIG. 2, the network architecture also includes a satellite communication management module and a user terminal 200, and the satellite communication management module manages the satellite communication. The step includes a request matching step, and the request matching step includes:

Step Sb1, when the semantic content data packet requested by the client 200 is routed to a certain satellite, the satellite forwards the semantic content data packet to the resource management node in the current virtual anchor point;

Step Sb2, the resource management node unpacks the semantic content data packet at the network layer to obtain the semantic label, and the resource management node judges whether the current virtual anchor can provide services according to the semantic label;

Step Sb3, if the current virtual anchor point cannot provide services, the resource management node transmits the semantic content data packet to another virtual anchor point; if all virtual anchor points cannot provide services, the semantic content data packet is transmitted to the ground service content data Center 400.

Exemplarily, when the semantic content data packet requested by the client 200 is routed to a virtual anchor, the resource management node of the virtual anchor unpacks it, and judges that the service content requested by the user matches the service content cached by the current virtual anchor If the current virtual anchor satellite cluster caches the service content or similar content, the space-based satellites in the current virtual anchor satellite cluster that cache the service content will provide the service content to the user, and there is no need to continue the user request to the subsequent satellite cluster. Inter-node routing can realize on-demand distribution of service content, reduce user request delay, and reduce multi-hop transmission of inter-satellite routes.

The request matching function of this application is introduced here. Fig. 4 is an exemplary flow chart of request matching to the semantic content data packet in an embodiment of the present application, with reference to shown in Fig. 4, the semantic content data packet requested by the ground user terminal in step S410 is routed to the current satellite; in step S420 the current satellite The satellite routes the semantic content data packet to the current resource management node in the current virtual anchor point; in step S430, the current resource management node unpacks the semantic content data packet; in step S440, the current resource management node calculates the matching degree, and judges whether the request matching is successful, If it is judged as successful, then the current virtual anchor directly provides service content in step S450; if it is judged as unsuccessful, then in step S441, the current resource management node transmits the semantic content data packet to other virtual anchors, if all virtual anchors cannot To provide services, the semantic content data packets are finally transmitted to the ground service content data center.

In some embodiments, in step Sb2 described above, the step of the resource management node judging whether the current virtual anchor can provide services according to the semantic label includes:

Step Sc1, the resource management node obtains the semantic features of user requests according to all user request data in the current virtual anchor service area, and generates a common request tag Comm_Tag;

Step Sc2, the resource management node uses the following formula (3) to calculate the user request semantic tag User_Req=(RTag ₁ , RTag ₂ , . . . , RTag _M ) and the common request tag Comm_Tag=(CTag ₁ , CTag ₂ , ..., CTag _N ) matching value Match_Lev1:

Match_Lev1 = Matching(Comm_Tag, User_Req) (3)

Among them, RTag _i represents the content of the semantic tag requested by the user, M represents the number of semantic tags requested by the user, CTag _i represents the content of common request tags, and N represents the number of common request tags;

Step Sc3, the resource management node judges whether the matching degree value Match_Lev1 is greater than or equal to the preset matching degree baseline value Match_BaseL1;

Step Sc4, if the matching degree value Match_Lev1 is greater than or equal to the preset matching degree baseline value Match_BaseL1 (that is, indicating a high matching degree), perform data matching with the corresponding service content in the current virtual anchor point, and calculate the matching degree value Match_Lev2 for the service content;

Step Sc5, the resource management node uses the following formula (4) to calculate the relationship between the user request semantic tag User_Req=(RTag ₁ , RTag ₂ ,..., RTag _H ) and the service content involved in the current virtual anchor point after preliminary matching Semantic tag Serv_Tag=(STag ₁ , STag ₂ , . . . , STag _G ) matching value Match_Lev2:

Match_Lev2 = Matching (User_Req, Serv_Tag) (4)

Among them, RTag _i represents the content of the semantic tag requested by the user, H represents the number of semantic tags requested by the user, STag _i represents the content of the semantic tag of the service content, and G represents the number of semantic tags of the current service content;

Step Sc6, the resource management node judges whether the matching degree value Match_Lev2 of the service content is greater than or equal to the preset matching degree baseline value Match_BaseL2;

Step Sc7, if the matching degree value Match_Lev2 is greater than or equal to the preset matching degree baseline value Match_BaseL2 (that is, the matching degree is high), then the current virtual anchor point provides services, that is, the satellite cluster to which the current virtual anchor point belongs provides service content for the user request , the data management node coordinates space-based satellites to provide service content for the user request based on the user request;

Step Sc8, if in the aforementioned step Sc3, the matching degree value Match_Lev1 is smaller than the preset matching degree baseline value Match_BaseL1 (that is, the matching degree is low), then the current virtual anchor point cannot provide services, and the user request data needs to be sent to the subsequent virtual anchor point. routing.

The request matching function of this application realizes the on-demand distribution function for user requests by space-based satellites. Usually, the semantic content data packets requested by users do not need to be routed to the ground service content data center with the most complete data volume and a long distance, thereby reducing user requests. delay, and reduce inter-satellite multi-hop transmission, saving space-based satellite communication network communication resources.

In some embodiments, the step of managing satellite communication by the satellite communication management module further includes a demand matching step, and the demand matching step includes:

Step Sd1, during the process of the space-based satellite communication network returning the semantic content data packet for providing services to the client, when the semantic content data packets for providing services are routed to a satellite of the virtual anchor point, the satellite forwards with The semantic content data packet for providing services to the resource management node in the current virtual anchor;

Step Sd2, the resource management node unpacks the semantic content data packet used to provide services at the network layer, and obtains the service content semantic labels used to provide services, and the resource management node judges whether it is necessary to cache the semantic content used to provide services according to the service content semantic labels content data package;

Step Sd3, if caching is required, the resource management node caches a copy of the semantic content data packet used to provide the service to any satellite in the current virtual anchor point.

Exemplarily, after the resource management node distributes the copy of the semantic content data packet to the virtual anchor satellite cluster for caching, the semantic content data packet continues to be routed backwards, and the semantic content data packet for providing services is finally transmitted to ground client. The demand matching function designed in this application realizes the on-demand storage of data in the space-based satellite communication network, and caches relevant content in the virtual anchor point according to common requests in the current service area. When the user initiates a content request, the current virtual anchor point can directly serve the Users distribute and provide service content on demand, thereby reducing user request delay and saving communication resources of space-based satellite communication network.

Here we introduce the requirement matching function of this application. Fig. 5 is an exemplary flow chart of carrying out requirement matching to semantic content data packet in one embodiment of the present application, with reference to Fig. 5, in step S510, the semantic content data packet used to provide services is routed to the current satellite; in step S520 the current The satellite routes the semantic content data packet for providing services to the current resource management node in the current virtual anchor point; in step S530, the current resource management node unpacks the semantic content data packet for providing services; in step S540, the current resource management node The node calculates the matching degree, and judges whether the demand matching is successful. If it is judged to be successful, then in step S550, the current resource management node caches the copy of the semantic content data packet used to provide the service to any satellite in the current virtual anchor point; if If it is judged unsuccessful, then in step S541, the current resource management node transmits the semantic content data packet for providing services to other virtual anchor points, and the semantic content data packets for providing services are finally transmitted to the ground user terminal.

In some embodiments, in step Sd2 mentioned above, the resource management node determines whether it is necessary to cache the semantic content data package for providing the service according to the semantic label of the service content, including:

Step Se1, according to historical requests within a certain period of time, judge whether the service content semantic tag has been requested by the current anchor point, and judge whether the relevant semantic tag has been cached and has certain content in the current virtual anchor point, if there is no cache, then directly cache Semantic content packages for serving; and

Step Se2, the resource management node obtains the semantic features of user requests according to all user request data in the current virtual anchor service area, and generates a common request tag Comm_Tag;

In step Se3, the resource management node uses the following formula (5) to calculate the match value Match_Lev3 between the service content semantic tag Serv_Tag and the common request tag Comm_Tag based on the matching algorithm:

Match_Lev3 = Matching(Comm_Tag, Serv_Tag) (5)

Step Se4, the resource management node judges whether the matching degree value Match_Lev3 is greater than or equal to the preset matching degree baseline value Match_BaseL3;

Step Se5, if the matching degree value Match_Lev3 is greater than or equal to the preset matching degree baseline value Match_BaseL3 (that is, indicating a high matching degree), then it is necessary to cache the semantic content data packets used to provide services;

Step Se6, if the matching degree value Match_Lev3 is smaller than the preset matching degree baseline value Match_BaseL3 (that is, the matching degree is low), then there is no need to cache the semantic content data packets for providing services.

Exemplarily, the resource management node of the virtual anchor caches the common service request tags of the current virtual anchor for demand matching and request matching functions, and the resource management node of the virtual anchor also caches the services to the current virtual anchor satellite cluster Content copies are negotiated and cached. In view of the limited cache resources of the satellite communication network, when resource management nodes distribute service content to space-based satellites for caching, there may be a problem of insufficient space-based satellite cache resources. This application designs resource management nodes to negotiate and cache for this situation. Based on the caching strategy, delete or relocate some duplicate or highly similar service content, and only keep the minimum capacity content that can meet the quality of service (Quality of Service, QoS) requirements for storage. The application realizes centralized management of the space-based satellite communication network by setting virtual anchor points and resource management nodes, and improves the resource utilization rate of the space-based satellite communication network and the collaborative processing capability among constellations.

In some embodiments, high-orbiting satellites are set as resource management nodes in virtual anchors. For example, high-orbit satellites have a wide coverage area, and high-orbit satellites are preferred as resource management nodes for virtual anchor points. Compared with the high dynamics of low-orbit satellites, high-orbit satellites have poor dynamics and can maintain a relatively long space over the virtual anchor point. time, avoid frequent replacement of resource management nodes, and save communication resources.

In some embodiments, any low-orbit satellites are preset to form a resource management node queue, and the low-orbit satellites in the resource management node queue can move synchronously. During the movement of the low-orbit satellite to another virtual anchor point, the resource management node The low-orbit satellites in the queue are always in one-to-one correspondence with the virtual anchor points, and the low-orbit satellites in one-to-one correspondence with the virtual anchor points are used as resource management nodes. Exemplarily, any low-orbit satellites are pre-designated to form a queue of resource management nodes. For example, at first, virtual anchor points A, B, and C correspond to resource management nodes a, b, and c respectively. Later, resource management nodes a, b, and c will press Synchronous movement in a predetermined order, resource management node a moves to virtual anchor point B, resource management node b moves to virtual anchor point C, resource management node c moves to virtual anchor point A, and the low-orbit satellites a, b, and c are always used as Resource management node for the current region.

This application selects several low-orbit satellites from the overall low-orbit satellite constellation as the resource management nodes of the virtual anchor point to ensure that there is always a low-orbit satellite in each virtual anchor point as the virtual anchor during the high dynamic period of the low-orbit satellite. Point resource management nodes, thus solving the problem of high-speed changes of satellite management nodes brought about by the high dynamics of low-orbit satellites.

In some embodiments, the satellite communication network architecture further includes a satellite movement management module, and the steps of the satellite movement management module managing the satellite movement include:

Step Sf1, when a certain satellite leaves the current virtual anchor point and enters another virtual anchor point, the resource management node distributes the cached content of the satellite to at least one satellite in the current virtual anchor point;

Step Sf2, the resource management node of another virtual anchor judges the matching degree between the cache content and this virtual anchor;

Step Sf3, if the matching degree is high, the satellite retains its own cache content, and the resource management node of another virtual anchor distributes service content based on the cache space of the satellite;

Step Sf4, if the matching degree is low, the satellite deletes its cache content, and the resource management node of another virtual anchor distributes a copy of the corresponding service content to the satellite.

For example, due to the high dynamics of the satellite communication network, the satellite cluster managed by the virtual anchor point often has the situation that the space-based satellite leaves the current virtual anchor point and enters the next virtual anchor point, so it is necessary to solve the data of the satellite cache content Relocation problem. To solve this problem, this application designs that when the space-based satellite leaves the current virtual anchor point, it sends its own cached service content to the resource management node of the virtual anchor point, and the resource management node deletes or distributes this part of the service content based on the cache strategy to the cached in other space-based satellites. After the space-based satellite enters the next virtual anchor point, it matches its own service content with the resource management node of the next virtual anchor point. Space-based satellite-based cache space distribution service content; if the matching degree is low, the current cache service content will be deleted directly, and the service content distributed by the resource management node of the next virtual anchor point will be received and cached. In addition, due to the shortage of space-based satellite cache resources, there may be a problem of insufficient cache space when the space-based satellite caches service content copies. Based on this problem, when caching copy content, if the space-based satellite cache space is insufficient, the resource management node will Scheduling, realizing space-based negotiation and caching function among virtual anchor satellite clusters.

Here is an introduction to the process of moving data from space-based satellites in this application. Fig. 6 is an exemplary flow chart of the movement of satellites in different virtual anchor points and the movement of satellite cache content in an embodiment of the present application. Referring to Fig. 2 and Fig. 6, in step S610 satellite A1 leaves the current virtual anchor point A and enters another A virtual anchor point B; in step S621, the satellite A1 returns its cached content to the resource management node A1001; in step S631, the resource management node A1001 distributes the service content cached by the satellite A1 to at least one satellite in the virtual anchor point A; in step S631 S622 satellite A1 sends its own cached content to resource management node B1001; in step S632 the resource management node B1001 of the virtual anchor point B judges the degree of matching between the cached content and the current virtual anchor point, if the degree of matching is high, then in step S640 satellite A1 Reserve its own cache content, the resource management node B1001 of the virtual anchor point B distributes the service content based on the cache space of the satellite A1; B1001 distributes corresponding service content to satellite A1.

The satellite communication network architecture mentioned above in this application is the semantic content-based space-based highly dynamic memory-transfer integrated network architecture, which can improve the utilization rate of space-based communication resources, save space-based cache resources, and reduce space-based communication delay. This application mainly includes the following designs: (1) The semantic layer scheme is designed to be embedded in the space-based application layer, and semantic features, semantic information extraction and semantic label generation based on semantics are realized for service content, which can save space-based communication resources and improve communication bandwidth utilization efficiency; design a semantic content data packet protocol based on semantic information and semantic tags, and realize lightweight design and efficient distribution of space-based communication data packets. (2) Design a solution based on semantic content to realize user request matching and virtual anchor point matching in space-based communication networks, realize on-demand storage and on-demand distribution, and reduce user request delays in space-based satellite communication networks and reduce satellite traffic. Inter-multi-hop transmission, saving space-based communication resources. (3) Design the virtual anchor point scheme of the space-based satellite communication network, divide the service area and build the virtual anchor point, each virtual anchor point includes a satellite cluster composed of at least one space-based satellite, and set the constellation of the satellite cluster in each satellite cluster Resource management nodes, manage and coordinate space-based relay nodes in the current virtual anchor satellite cluster, realize distributed caching and collaborative caching of space-based relay nodes, realize unified resource management among constellations, and improve resource utilization of space-based satellite communication networks .

The basic concept has been described above, obviously, for those skilled in the art, the above disclosure of the invention is only an example, and does not constitute a limitation to the present application. Although not expressly stated here, various modifications, improvements and amendments to this application may be made by those skilled in the art. Such modifications, improvements, and amendments are suggested in this application, so such modifications, improvements, and amendments still belong to the spirit and scope of the exemplary embodiments of this application.

Meanwhile, the present application uses specific words to describe the embodiments of the present application. For example, "one embodiment", "an embodiment", and/or "some embodiments" refer to a certain feature, structure or characteristic related to at least one embodiment of the present application. Therefore, it should be emphasized and noted that two or more references to "an embodiment" or "an embodiment" or "an alternative embodiment" in different places in this specification do not necessarily refer to the same embodiment . In addition, certain features, structures or characteristics of one or more embodiments of the present application may be properly combined.

Some aspects of the present application may be entirely implemented by hardware, may be entirely implemented by software (including firmware, resident software, microcode, etc.), or may be implemented by a combination of hardware and software. The above hardware or software may be referred to as "block", "module", "engine", "unit", "component" or "system". The processor can be one or more Application Specific Integrated Circuits (ASICs), Digital Signal Processors (DSPs), Digital Signal Processing Devices (DAPDs), Programmable Logic Devices (PLDs), Field Programmable Gate Arrays (FPGAs), processors , a controller, a microcontroller, a microprocessor, or a combination thereof. Additionally, aspects of the present application may be embodied as a computer product comprising computer readable program code on one or more computer readable media. For example, computer-readable media may include, but are not limited to, magnetic storage devices (e.g., hard disk, floppy disk, magnetic tape...), optical disks (e.g., compact disk CD, digital versatile disk DVD...), smart cards, and flash memory devices ( For example, cards, sticks, key drives...).

A computer readable medium may contain a propagated data signal embodying a computer program code, for example, in baseband or as part of a carrier wave. The propagated signal may take many forms, including electromagnetic, optical, etc., or a suitable combination. The computer-readable medium can be any computer-readable medium, except computer-readable storage media, that can communicate, propagate, or transfer the program for use by being coupled to an instruction execution system, apparatus, or device. Program code on a computer readable medium may be transmitted over any suitable medium, including radio, electrical cables, fiber optic cables, radio frequency signals, or the like, or combinations of any of the foregoing.

In the same way, it should be noted that in order to simplify the expression disclosed in the present application and help the understanding of one or more embodiments of the invention, in the foregoing description of the embodiments of the present application, sometimes multiple features are combined into one embodiment, drawings or descriptions thereof. This method of disclosure does not, however, imply that the subject matter of the application requires more features than are recited in the claims. Indeed, embodiment features are less than all features of a single foregoing disclosed embodiment.

In some embodiments, numbers describing the quantity of components and attributes are used. It should be understood that such numbers used in the description of the embodiments use the modifiers "about", "approximately" or "substantially" in some examples. grooming. Unless otherwise stated, "about", "approximately" or "substantially" indicates that the stated figure allows for a variation of ±20%. Accordingly, in some embodiments, the numerical parameters used in the specification and claims are approximations that can vary depending upon the desired characteristics of individual embodiments. In some embodiments, numerical parameters should take into account the specified significant digits and adopt the general digit reservation method. Although the numerical ranges and parameters used in some embodiments of the present application to confirm the breadth of the scope are approximate values, in specific embodiments, such numerical values are set as precisely as practicable.

Although the present application has been described with reference to the current specific embodiments, those of ordinary skill in the art should recognize that the above embodiments are only used to illustrate the present application, and can also be made without departing from the spirit of the present application. Various equivalent changes or substitutions, therefore, as long as the changes and modifications to the above-mentioned embodiments are within the spirit of the present application, they will all fall within the scope of the claims of the present application.

Claims (11)

1. A semantic content based satellite communications network architecture comprising:

the application layer comprises a semantic layer, the semantic layer is used for extracting semantic information and semantic tags of satellite communication data, and the application layer, the transmission layer and the network layer generate semantic content data packets according to the semantic information and the semantic tags, wherein the semantic layer comprises: the semantic layer extracts the semantic information of the satellite communication data according to the semantic information knowledge base; the semantic feature knowledge base is used for extracting semantic features of the satellite communication data; a self-attention mechanism model for compressing the semantic features to generate the semantic tags;

the system comprises at least one virtual anchor point, wherein the virtual anchor point comprises at least one satellite and a resource management node, the resource management node is used for managing each satellite in the virtual anchor point, and each satellite is used for acquiring the semantic content data packet and communicating based on the semantic content data packet protocol.

2. The satellite communication network architecture of claim 1, wherein the satellite communication data comprises an image, and the step of extracting semantic information and semantic tags of the satellite communication data by the semantic layer comprises:

Semantic information extraction: dividing the image by using a neural network model to obtain a plurality of areas, calculating the importance degree of the semantic features of each area, and if the importance degree of the semantic features is greater than or equal to a preset threshold value, completely reserving the semantic information of the areas; if the importance of the semantic features is smaller than the preset threshold, partially reserving the semantic information of the region, wherein partially reserving the semantic information of the region comprises: calculating the maximum pixel of the duty ratio in the region, and reserving the content of the pixel;

semantic tag extraction: generating semantic features of the image using the semantic feature knowledge base, the semantic features comprising a descriptive language for the image; the step of the self-attention mechanism model compressing the semantic features to generate the semantic tags includes: calculating each word x in the semantic feature using the following formula _i Attention value Attention (x) _i )：

Wherein N is the length of the descriptive language,for word x _i Logarithmic representation of Value _i Representing word x _i Weight index of (2);

if the Attention value Attention (x _i ) If the number is larger than or equal to a preset number, the word x is reserved _i As the semantic tags.

3. The satellite communications network architecture of claim 1, wherein the steps of the application layer, the transport layer, and the network layer generating semantic content data packets from the semantic information and the semantic tags comprise:

the application layer obtains the semantic information and the semantic tag extracted by the semantic layer;

the transmission layer packs the semantic information as data bits and adds an application layer packet header and a transmission layer packet header;

and the network layer packages the semantic tags, adds a network layer packet header and completes packaging the semantic content data packet.

4. A satellite communications network architecture according to claim 3, wherein the step of each satellite communicating based on the semantic content data packet protocol comprises:

the satellite receives the semantic content data packet, unpacks the semantic content data packet at the network layer, and obtains the semantic tag;

continuing unpacking the semantic content data packet at the transmission layer to obtain the semantic information;

and caching the semantic information at the application layer.

5. The satellite communication network architecture of claim 1, wherein the network architecture further comprises a satellite communication management module and a user side, the step of managing satellite communications by the satellite communication management module comprising a request matching step comprising:

When the semantic content data packet requested by the user side is routed to a certain satellite, the satellite forwards the semantic content data packet to the resource management node in the current virtual anchor point;

the resource management node unpacks the semantic content data packet at the network layer to obtain the semantic tag, and the resource management node judges whether the current virtual anchor point can provide service according to the semantic tag;

if the current virtual anchor point can not provide service, the resource management node transmits the semantic content data packet to another virtual anchor point;

and if all the virtual anchors can not provide service, the semantic content data packet is transmitted to a ground service content data center.

6. The satellite communications network architecture of claim 5, wherein the step of the resource management node determining whether the current virtual anchor point can provide service based on the semantic tag comprises:

the resource management node obtains user request semantic features according to all user request data of the current virtual anchor point service area, and generates a common request Tag Cimm_tag;

the resource management node calculates a User request semantic tag user_req= (RTag) using the following formula ₁ ，RTag ₂ ，...，RTag _M ) Comm_tag= (CTag) with the common request Tag ₁ ，CTag ₂ ，...，CTag _N ) Match_lev1:

Match_Lev1＝Matching(Comm_Tag，User_Req)

wherein RTag _i Representing the content of the user request semantic tags, M representing the number of the user request semantic tags, CTag _i Representing the content of the common request tags, and N represents the number of the common request tags;

the resource management node judges whether the matching degree value Match_Lev1 is more than or equal to a preset matching degree base line value Match_BaseL1;

if the matching degree value Match_Lev1 is larger than or equal to the preset matching degree baseline value Match_BaseL1, performing data matching with corresponding service contents in the current virtual anchor point, and calculating a matching degree value Match_Lev2 related to the service contents;

the resource management node calculates the User request semantic tag user_req= (RTag) using the following formula ₁ ，RTag ₂ ，...，RTag _H ) Semantic Tag serv_tag= (STag) of service content involved after preliminary matching with current virtual anchor point ₁ ，STag ₂ ，...，STag _G ) Match_lev2:

Match_Lev2＝Matching(User_Req，Serv_Tag)

wherein RTag _i Representing the content of the user request semantic tags, H representing the number of the user request semantic tags, STag _i Content representing semantic tags of the service content, G representing the number of semantic tags of the current service content;

The resource management node judges whether the Match degree value Match_Lev2 about the service content is larger than or equal to a preset Match degree base line value Match_BaseL2;

if the matching degree value Match_Lev2 is larger than or equal to the preset matching degree baseline value Match_BaseL2, providing service by the current virtual anchor point;

if the Match degree value match_lev1 is smaller than the preset Match degree baseline value match_basel1, the current virtual anchor point cannot provide service, and the user request data is required to be routed to a subsequent virtual anchor point continuously.

7. The satellite communication network architecture of claim 5, wherein the step of the satellite communication management module managing satellite communications further comprises a demand matching step comprising:

in the process of returning the semantic content data packet for providing the service to the user side, when the semantic content data packet for providing the service is routed to a certain satellite, the satellite forwards the semantic content data packet for providing the service to the resource management node in the current virtual anchor point;

the resource management node unpacks the semantic content data packet for providing the service at the network layer to obtain a service content semantic tag for providing the service, and the resource management node judges whether the semantic content data packet for providing the service needs to be cached according to the service content semantic tag;

And if the cache is needed, the resource management node caches the copy of the semantic content data packet for providing the service to any satellite in the current virtual anchor point.

8. The satellite communication network architecture of claim 7, wherein the step of the resource management node determining whether buffering of the semantic content data packets for providing services is required based on the service content semantic tags comprises:

judging whether the service content semantic tags are requested by the current anchor point according to the history request within a certain time, judging whether the related semantic tags have cached certain content in the current virtual anchor point, and if not, directly caching the semantic content data packet for providing service; and

the resource management node obtains user request semantic features according to all user request data of the current virtual anchor point service area, and generates a common request Tag Comm_Tag;

the resource management node calculates a matching degree value Match_Lev3 of the service content semantic Tag Serv_Tag and the common request Tag Comm_Tag by using the following formula:

Match_Lev3＝Matching(Comm_Tag，Serv_Tag)

the resource management node judges whether the matching degree value Match_Lev3 is more than or equal to a preset matching degree base line value Match_BaseL3;

If the matching degree value Match_Lev3 is larger than or equal to the preset matching degree baseline value Match_BaseL3, the semantic content data packet for providing service needs to be cached;

if the matching degree value Match_Lev3 is smaller than the preset matching degree baseline value Match_BaseL3, the semantic content data packet for providing service does not need to be cached.

9. The satellite communication network architecture of claim 1, wherein a high orbit satellite is set in the virtual anchor as the resource management node.

10. The satellite communication network architecture of claim 1, wherein a resource management node queue is formed by presetting any low-orbit satellite, the low-orbit satellites in the resource management node queue can synchronously move, and in the process that the low-orbit satellites move to another virtual anchor point, the low-orbit satellites in the resource management node queue are always in one-to-one correspondence with the virtual anchor point, and the low-orbit satellites in one-to-one correspondence with the virtual anchor point serve as the resource management nodes.

11. The satellite communications network architecture of claim 1, further comprising a satellite motion management module, the satellite motion management module managing satellite motion comprising: when a certain satellite leaves a current virtual anchor point and enters another virtual anchor point, the resource management node distributes the cache content of the satellite to at least one satellite in the current virtual anchor point; and

The resource management node of the other virtual anchor point judges the matching degree of the cache content and the virtual anchor point;

if the matching degree is high, the satellite reserves the self cache content, and the resource management node of the other virtual anchor point distributes service content based on the cache space of the satellite;

and if the matching degree is low, deleting the self cache content by the satellite, and distributing corresponding service content copies to the satellite by the resource management node of the other virtual anchor point.