Data transmission method and device, communication equipment and storage medium

Abstract

The embodiment of the invention discloses a data transmission method and device, communication equipment and a storage medium. The method comprises the following steps: a first node performs semantic analysis on first data based on a local first knowledge base to obtain a plurality of semantic feature vectors and determine a first parameter of each semantic feature vector; the first data is from terminal equipment; the first parameter represents the semantic value of the semantic feature vector; and the first node encodes each semantic feature vector based on the first parameter and the first knowledge base to obtain second data, and sends the second data to a base station.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a data transmission method, apparatus, communication equipment and storage medium. Background Art

[0002] In traditional symbol-based communications, transceivers design their codecs based on the principle of complete and accurate transmission of the original data stream, making it difficult to achieve breakthroughs in communication capacity. In task-centric communications, raw data follows the principle of "understanding first, then transmission." Application intent is first extracted as semantic information, and then this semantic information is encoded and decoded to achieve more efficient and concise communication of intent. In semantic communication, source symbols are mapped to semantic code streams based on artificial intelligence (AI) encoding.

[0003] In traditional semantic communication, the computing and storage capabilities of user equipment (UE) are limited, and often cannot meet the training and deployment requirements of AI-based semantic coding models. Summary of the Invention

[0004] To solve existing technical problems, embodiments of the present invention provide a data transmission method, apparatus, communication device, and storage medium.

[0005] To achieve the above-mentioned purpose, the technical solution of the embodiment of the present invention is implemented as follows:

[0006] In a first aspect, an embodiment of the present invention provides a data transmission method, which is applied to a first node and includes:

[0007] The first node performs semantic analysis on the first data based on a local first knowledge base to obtain a plurality of semantic feature vectors and determine a first parameter of each semantic feature vector; the first data comes from a terminal device; and the first parameter represents a semantic value of the semantic feature vector;

[0008] The first node encodes each semantic feature vector based on the first parameter and the first knowledge base to obtain second data, and sends the second data to the base station.

[0009] In the above solution, determining the first parameter of each semantic feature vector includes:

[0010] Based on the first knowledge base, the first node determines the first type corresponding to the first data, evaluates the semantic value of each semantic feature vector based on the first type, and determines the first parameter of each semantic feature vector; the first parameter represents the semantic value of the semantic feature vector corresponding to the first type; wherein, the first type is a human perception-oriented type or a task accuracy-oriented type.

[0011] In the above solution, the method further includes: the first node sending the first parameter to the base station.

[0012] In the above scheme, the first node encodes each semantic feature vector based on the first parameter and the first knowledge base, including: the first node judges whether it can or cannot provide the first service corresponding to the first data based on the first knowledge base, determines the mapping method based on the judgment result, and encodes each semantic feature vector based on the mapping method and the first parameter.

[0013] In the above solution, the mapping mode includes a first mapping mode, a second mapping mode or a third mapping mode;

[0014] The amounts of the second data encoded by respectively adopting the first mapping mode, the second mapping mode and the third mapping mode increase in sequence;

[0015] And / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

[0016] In the above solution, determining the mapping method based on the judgment result includes:

[0017] If the result of the determination is that the first service corresponding to the first data can be provided, determining a first mapping mode;

[0018] If the judgment result is that the first service corresponding to the first data cannot be provided, a third mapping method is determined.

[0019] In the above solution, encoding each semantic feature vector based on the mapping method and the first parameter includes: the first node encoding each semantic feature vector based on the mapping method, the first parameter and the first neural network model.

[0020] In the above solution, the method also includes: the first node determines or updates the parameters of the first neural network model based on wireless channel environment information and / or wireless resource status information.

[0021] In the above solution, after sending the second data to the base station, the method further includes: the first node receiving first information sent by the base station, where the first information is used to request or instruct re-encoding of the second data.

[0022] In the above solution, when the mapping mode is the first mapping mode, the method further includes:

[0023] The first node re-encodes each semantic feature vector according to the second mapping manner to obtain re-encoded second data, and sends the re-encoded second data to the base station.

[0024] In the above solution, when the mapping mode is the second mapping mode, the method further includes:

[0025] The first node re-encodes each semantic feature vector according to the third mapping manner to obtain re-encoded second data, and sends the re-encoded second data to the base station.

[0026] In the above solution, the first node receiving the first information sent by the base station includes:

[0027] The first node receives a first medium access control protocol data unit (MAC PDU) sent by the base station, where the first MAC PDU includes a first medium access control control element (MAC CE), and the first MAC CE includes the first information.

[0028] In the above solution, the first MAC CE is located at the beginning of the first MAC PDU.

[0029] In the above solution, the second data is carried by the second MAC PDU.

[0030] In the above solution, the second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

[0031] In the above solution, the second MAC CE is located at the end of the second MAC PDU.

[0032] In the above solution, when the second data is encoded according to the first mapping mode or the second mapping mode, the second data is carried in the second MAC CE;

[0033] When the second data is obtained by encoding according to the third mapping manner, the second data is carried in other fields of the second MAC PDU except the second MAC CE.

[0034] In a second aspect, an embodiment of the present invention further provides a data transmission method, which is applied to a base station and includes:

[0035] The base station receives the second data sent by the first node, decodes the second data based on a local second knowledge base, and obtains a plurality of semantic feature vectors when the decoding is successful;

[0036] The base station drives related tasks based on the multiple semantic feature vectors, or converts the multiple semantic feature vectors based on the second knowledge base to obtain first data.

[0037] In the above scheme, the multiple semantic feature vectors are converted based on the second knowledge base to obtain the first data, including: the base station corrects the errors of the multiple semantic feature vectors based on the second knowledge base, and converts the multiple semantic feature vectors after error correction to obtain the first data.

[0038] In the above solution, the method further includes: the base station receiving a first parameter sent by the first node, where the first parameter represents the semantic value of the semantic feature vector.

[0039] In the above solution, decoding the second data based on the local second knowledge base includes:

[0040] The base station decodes the second data based on the second knowledge base and the first parameter.

[0041] In the above solution, decoding the second data based on the local second knowledge base includes:

[0042] The base station decodes the second data based on a second neural network model in the second knowledge base.

[0043] In the above solution, the method also includes: the base station determines or updates the parameters of the second neural network model based on wireless channel environment information and / or wireless resource status information.

[0044] In the above solution, the method further includes: when decoding is unsuccessful, the base station sends the second data to other base stations, and receives processing results corresponding to the second data sent by the other base stations.

[0045] In the above scheme, the method also includes: when the processing result is that the other base station cannot successfully decode the second data, the base station sends first information to the first node, and the first information is used to request or instruct re-encoding of the second data.

[0046] In the above solution, the method further includes: the base station receiving the re-encoded second data sent by the first node.

[0047] In the above solution, the base station sending the first information to the first node includes:

[0048] The base station receives a first MAC PDU sent by the first node, where the first MAC PDU includes a first MAC CE, and the first MAC CE includes the first information.

[0049] In the above solution, the first MAC CE is located at the beginning of the first MAC PDU.

[0050] In the above solution, the second data is carried by the second MAC PDU.

[0051] In the above solution, the second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

[0052] In the above solution, the second MAC CE is located at the end of the second MAC PDU.

[0053] In the above solution, when the second data is encoded according to the first mapping mode or the second mapping mode, the second data is carried in the second MAC CE;

[0054] When the second data is obtained by encoding according to the third mapping manner, the second data is carried in other fields except the second MAC CE in the second MAC PDU;

[0055] The amounts of the second data encoded by respectively adopting the first mapping mode, the second mapping mode and the third mapping mode increase in sequence;

[0056] And / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

[0057] In a third aspect, an embodiment of the present invention further provides a data transmission device, which is applied to a first node and includes: a first processing unit and a first communication unit; wherein,

[0058] The first processing unit is configured to perform semantic analysis on the first data based on a local first knowledge base to obtain a plurality of semantic feature vectors and determine a first parameter for each semantic feature vector; the first data is from a terminal device; the first parameter represents a semantic value of the semantic feature vector; and is further configured to encode each semantic feature vector based on the first parameter and the first knowledge base to obtain second data;

[0059] The first communication unit is used to send the second data to the base station.

[0060] In a fourth aspect, an embodiment of the present invention further provides a data transmission device, which is applied to a base station and includes: a second communication unit and a second processing unit; wherein,

[0061] The second communication unit is configured to receive second data sent by the first node;

[0062] The second processing unit is used to decode the second data based on a local second knowledge base, and obtain multiple semantic feature vectors if the decoding is successful; drive related tasks based on the multiple semantic feature vectors, or convert the multiple semantic feature vectors based on the second knowledge base to obtain first data.

[0063] In a fifth aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the data transmission method described in the first or second aspect of the embodiment of the present invention.

[0064] In the sixth aspect, an embodiment of the present invention further provides a communication device, comprising a memory, a processor, and a computer program stored in the memory and runnable on the processor, wherein when the processor executes the program, the steps of the data transmission method described in the first or second aspect of the embodiment of the present invention are implemented.

[0065] In a seventh aspect, an embodiment of the present invention further provides a computer program product, comprising computer program instructions, which enable a computer to execute the steps of the data transmission method described in the first or second aspect of the embodiment of the present invention.

[0066] The data transmission method, apparatus, communication device, and storage medium provided by the embodiments of the present invention include: a first node performs semantic analysis on first data based on a local first knowledge base to obtain multiple semantic feature vectors and determine a first parameter for each semantic feature vector; the first data comes from a terminal device; the first parameter represents the semantic value of the semantic feature vector; the first node encodes each semantic feature vector based on the first parameter and the first knowledge base to obtain second data, and sends the second data to a base station; the base station decodes the second data based on a local second knowledge base, and obtains multiple semantic feature vectors if the decoding is successful; the base station drives related tasks based on the multiple semantic feature vectors, or converts the multiple semantic feature vectors based on the second knowledge base to obtain the first data. The technical solution of the embodiments of the present invention is adopted, and a first node is introduced as a model deployment node for semantic encoding of UE original data. The first node's strong computing and storage capabilities are utilized to establish a local first knowledge base at the first node for converting original data into semantic symbols, and functions such as source-channel semantic joint encoding are introduced to adapt to the transmission of semantic signals on the channel. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] Figure 1 A schematic diagram of a communication architecture applied to a data transmission method according to an embodiment of the present invention;

[0068] Figure 2 Schematic diagram of the data transmission method according to an embodiment of the present invention Figure 1 ;

[0069] Figure 3a and Figure 3b Schematic diagram of the MAC PDU structure in the data transmission method according to an embodiment of the present invention;

[0070] Figure 4 Schematic diagram of the data transmission method according to an embodiment of the present invention Figure 2 ;

[0071] Figure 5 A schematic diagram of an interactive process of a data transmission method according to an embodiment of the present invention;

[0072] Figure 6 Schematic diagram of the structure of the data transmission device according to an embodiment of the present invention Figure 1 ;

[0073] Figure 7 Schematic diagram of the structure of the data transmission device according to an embodiment of the present invention Figure 2 ;

[0074] Figure 8 Schematic diagram of the hardware structure of a communication device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0075] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0076] The technical solutions of the embodiments of the present invention can be applied to various communication systems, such as a Global System of Mobile Communication (GSM) system, a Long Term Evolution (LTE) system, or a 5G system. Optionally, the 5G system or 5G network may also be referred to as a New Radio (NR) system or NR network.

[0077] Exemplarily, the communication system applied in the embodiment of the present invention may include a network device and a terminal device (also referred to as a terminal, a communication terminal, etc.); the network device may be a device that communicates with the terminal device. The network device may provide communication coverage within a certain area and may communicate with terminals located in the area. Optionally, the network device may be a base station in each communication system, such as an evolved base station (eNB, Evolutional Node B) in an LTE system, or a base station (gNB) in a 5G system or NR system.

[0078] It should be understood that in the embodiments of the present application, devices having communication functions in the network / system may be referred to as communication devices. Communication devices may include network devices and terminals having communication functions. The network devices and terminal devices may be the specific devices described above and will not be described in detail here. Communication devices may also include other devices in the communication system, such as network controllers, mobility management entities, and other network entities, which are not limited in the embodiments of the present invention.

[0079] It should be understood that the terms "system" and "network" are often used interchangeably herein. The term "and / or" is simply a description of an association between related objects, indicating that three possible relationships exist. For example, "A and / or B" can represent: A exists alone, A and B exist simultaneously, or B exists alone. Furthermore, the character " / " generally indicates that the related objects are in an "or" relationship.

[0080] The terms "first", "second" etc. in the specification and claims of the present application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable in appropriate circumstances, so that the embodiments of the present application described herein can, for example, be implemented in an order other than those illustrated or described herein. In addition, the terms "comprise" and "have" and any of their variations are intended to cover non-exclusive inclusions, for example, the process, method, system, product or equipment comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or that are inherent to these processes, methods, products or equipment.

[0081] Figure 1 Schematic diagram of a communication architecture for a data transmission method according to an embodiment of the present invention; Figure 1 As shown, the communication architecture of this embodiment of the present invention introduces a first node (also called a central node) as a data collector. A source knowledge base is established on the first node, along with a neural network deployed to convert raw data into semantic symbols. Furthermore, functions such as source-channel semantic joint encoding and decoding are introduced to accommodate the transmission of semantic signals across the channel.

[0082] The following combination Figure 1 The communication architecture shown is used to illustrate the data transmission method according to an embodiment of the present invention.

[0083] An embodiment of the present invention provides a data transmission method. Figure 2 Schematic diagram of the data transmission method according to an embodiment of the present invention Figure 1 ;like Figure 2 As shown, the method includes:

[0084] Step 101: A first node performs semantic analysis on first data based on a local first knowledge base to obtain a plurality of semantic feature vectors and determine a first parameter of each semantic feature vector; the first data comes from a terminal device; and the first parameter represents a semantic value of the semantic feature vector.

[0085] Step 102: The first node encodes each semantic feature vector based on the first parameter and the first knowledge base to obtain second data, and sends the second data to a base station.

[0086] In this embodiment, the first node obtains first data from the terminal device, and the terminal device generates source data (first data) as a source. In this embodiment, the type of the first data may include one or more of the following: text, audio, image, video, point cloud data, volumetric video, etc.

[0087] In this embodiment, refer to Figure 1As shown, the first node can perform semantic analysis on the first data in combination with the local first knowledge base through the semantic analysis conversion module to obtain multiple semantic feature vectors. The first node pre-establishes a local knowledge base (recorded as the first knowledge base), which can also be called a semantic knowledge base. It is a structured knowledge network model with memory capabilities that can provide relevant semantic knowledge descriptions for data information. Exemplarily, the knowledge base (recorded as the first knowledge base) includes multiple elements and the association relationship between elements; wherein the elements can be single words, phrases, etc.; the association relationship can specifically refer to a semantic relationship, such as a synonym relationship, antonym relationship, superordinate / subordinate relationship, whole / part relationship, etc. In this embodiment, the knowledge base (recorded as the first knowledge base) can also include a neural network required for semantic analysis conversion, semantic importance modeling, and encoding. In this embodiment, the first node performs semantic analysis on the first data based on the first knowledge base to obtain multiple semantic feature vectors.

[0088] In some optional embodiments, the first node extracts semantic features of the first data based on the first knowledge base and generates semantic annotations, and the semantic features carrying the semantic annotations are divided into multiple semantic streams, each semantic stream includes a semantic feature vector; wherein, the elements in each semantic feature vector belong to the same semantic object.

[0089] In some optional embodiments, determining the first parameter of each semantic feature vector includes: the first node determines the first type corresponding to the first data based on the first knowledge base, evaluates the semantic value of each semantic feature vector based on the first type, and determines the first parameter of each semantic feature vector; the first parameter represents the semantic value of the semantic feature vector corresponding to the first type; wherein the first type is a human perception-oriented type or a task accuracy-oriented type.

[0090] In this embodiment, refer to Figure 1 As shown, the first node can determine a first parameter of each semantic feature vector using a semantic importance modeling module in conjunction with a local first knowledge base. Optionally, the first parameter represents the semantic value of the semantic feature vector. The first node evaluates the first parameter of each semantic feature vector, which is related to the communication purpose in different scenarios.

[0091] As an implementation method, for modeling importance for human perception, that is, for scenarios oriented to human perception or human communication, since humans can find different visually salient areas in a scene under a given background, this model combines entropy modeling and saliency detection to generate a value index of a semantic feature vector.

[0092] As another implementation, to model the importance of task-oriented accuracy, for the machine-type communication scenario, since the gradient of the task target can directly flow into the neural network to highlight the important areas in the feature map, this mode generates a value index of the semantic feature vector by jointly using entropy modeling and gradient-weighted class activation mapping.

[0093] Based on this, in this embodiment, the first data is divided into the above two types. The first node determines that the first data corresponds to the first type based on the local first knowledge base, and then evaluates the semantic value of each semantic feature vector according to the first type to determine the first parameter of each semantic feature vector. Exemplarily, the first knowledge base may include a neural network model, for example, it may include the entropy modeling and significance detection model in the above first embodiment, and the semantic feature vectors belonging to the first type are processed based on the entropy modeling and significance detection model to obtain the corresponding first parameters. For another example, it may include the entropy modeling and gradient weighted class activation mapping model in the above second embodiment, and the semantic feature vectors belonging to the second type are processed based on the entropy modeling and gradient weighted class activation mapping to obtain the corresponding first parameters.

[0094] In this embodiment, the first parameter represents the semantic value of the semantic feature vector, and the semantic value can reflect the importance of the semantic feature vector in semantic communication, that is, the numerical value of the first parameter can reflect the importance of the semantic feature vector in semantic communication. For example, the larger the first parameter is, the higher the semantic value of the corresponding semantic feature vector is, that is, the more important the semantic feature vector is in semantic communication; correspondingly, the smaller the first parameter is, the lower the semantic value of the corresponding semantic feature vector is, that is, the less important the semantic feature vector is in semantic communication.

[0095] In some optional embodiments, the method further includes: the first node sending the first parameter to the base station to assist the base station in decoding the received data.

[0096] In this embodiment, refer to Figure 1 As shown, the first node can encode the semantic feature vector through the source-channel joint coding module in combination with the local first knowledge base. Specifically, the first node encodes each semantic feature vector based on the first parameter and the first knowledge base, specifically semantically encodes the semantic feature vector to generate a signal suitable for transmission on the channel. The compression (source coding) and error correction (channel coding) of the encoder are implemented in an integrated manner of source-channel joint coding. In some optional embodiments, the first node encodes the semantic feature vector based on the first parameter and the first neural network model in the local first knowledge base. The first neural network is mainly used to encode the semantic feature vector to obtain the second data; it can also be responsible for transmission resource allocation.

[0097] In some optional embodiments of the present invention, the first node encodes each semantic feature vector based on the first parameter and the first knowledge base, including: the first node determines whether it can or cannot provide the first service corresponding to the first data based on the first knowledge base, determines a mapping method based on the judgment result, and encodes each semantic feature vector based on the mapping method and the first parameter.

[0098] In some optional embodiments, the mapping method includes a first mapping method, a second mapping method or a third mapping method; wherein the amount of the second data encoded using the first mapping method, the second mapping method and the third mapping method respectively increases successively; and / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

[0099] In this embodiment, the mapping method provided in the first knowledge base of the first node can be divided into three levels, such as low-level, intermediate and high-level, or three levels of mapping, such as low-level mapping, intermediate mapping and high-level mapping; wherein, the low-level or low-level mapping corresponds to the third mapping method, the intermediate or intermediate mapping corresponds to the second mapping method, and the high-level or high-level mapping corresponds to the first mapping method. Different levels or different mapping methods correspond to different levels of semantic abstraction, or to different amounts of data. Exemplarily, the high-level or high-level mapping (first mapping method) corresponds to the highest level of semantic abstraction, and the amount of data encoded using this method is the smallest; the low-level or low-level mapping (third mapping method) corresponds to the lowest level of semantic abstraction, or even retains the original data, and the amount of data encoded using this method is the largest.

[0100] In this embodiment, a service may be pre-configured in the first knowledge base of the first node, and the service may be implemented through a network model. It can be understood that a network model may be pre-configured in the first knowledge base of the first node, and different network models may provide different services. Exemplarily, it may include model training services, reasoning services, modeling services, and the like. Accordingly, the first knowledge base of the first node may pre-configure a network model for model training, a network model for reasoning, a network model for modeling, and the like. Before encoding the semantic feature vector, the first node first queries the local first knowledge base to determine whether the first knowledge base has a first service (or network model) that can provide a matching service, determines a mapping method based on whether the first knowledge base can provide the first service (or whether the first knowledge base has a network model corresponding to the first service), and encodes each semantic feature vector based on the mapping method and the first parameter.

[0101] In some optional embodiments, determining the mapping method based on the judgment result includes: determining the first mapping method when the judgment result is that the first service corresponding to the first data can be provided; and determining the third mapping method when the judgment result is that the first service corresponding to the first data cannot be provided.

[0102] In this embodiment, if the first knowledge base can provide the first service (or the first knowledge base has a network model corresponding to the first service), the first mapping method (i.e., high-level mapping) is determined, and the first mapping method (i.e., high-level mapping) is used for encoding; if the first knowledge base cannot provide the first service (or the first knowledge base does not have a network model corresponding to the first service), the third mapping method (i.e., low-level mapping) is determined, that is, the semantic feature vector is directly low-level encoded and transmitted.

[0103] In some optional embodiments, encoding each semantic feature vector based on the mapping method and the first parameter includes: the first node encoding each semantic feature vector based on the mapping method, the first parameter and a first neural network model.

[0104] In this embodiment, after the first node determines the mapping mode, it encodes the semantic feature vector based on the mapping mode and in combination with the first parameter and the first neural network model in the local first knowledge base.

[0105] In some optional embodiments of the present invention, the method further includes: the first node determining or updating parameters of the first neural network model based on wireless channel environment information and / or wireless resource status information.

[0106] In this embodiment, the first node may collect wireless channel environment information and / or wireless resource status information, regularly optimize the local first neural network model, and train the model parameters of the first neural network model.

[0107] In some optional embodiments of the present invention, after sending the second data to the base station, the method further includes: the first node receiving first information sent by the base station, where the first information is used to request or instruct re-encoding of the second data.

[0108] In this embodiment, after the first node encodes the semantic feature vector according to the determined mapping method to obtain the second data and sends the second data to the base station, the local second knowledge base of the base station may provide the corresponding service, that is, the second knowledge base may be pre-configured with a network model of the corresponding service; or the second knowledge base of the base station may not be able to provide the corresponding service, that is, the second knowledge base is not pre-configured with a network model of the corresponding service. The base station may successfully decode the received second data or fail to decode it; if the base station side (including the base station and other base stations) determines that the second data cannot be successfully decoded, that is, the base station side cannot understand the corresponding semantics, then the base station sends a first message to the first node for requesting or instructing the re-encoding of the second data.

[0109] In some optional embodiments, when the mapping method is the first mapping method, the method further includes: the first node re-encodes each semantic feature vector according to the second mapping method, obtains the re-encoded second data, and sends the re-encoded second data to the base station.

[0110] In this embodiment, if in step 102, the first node uses the first mapping method to encode the semantic feature vector, and receives the first information sent by the base station, it further lowers the mapping level, uses the second mapping method to re-encode each semantic feature vector, obtains the re-encoded second data, and sends the re-encoded second data to the base station.

[0111] In other optional embodiments, when the mapping method is the second mapping method, the method further includes: the first node re-encodes each semantic feature vector according to the third mapping method, obtains the re-encoded second data, and sends the re-encoded second data to the base station.

[0112] In this embodiment, if in step 102, the first node uses the second mapping method to encode the semantic feature vector and receives the first information sent by the base station, the mapping level is further lowered and the third mapping method is used to re-encode each semantic feature vector to obtain the re-encoded second data, and the re-encoded second data is sent to the base station.

[0113] In some optional embodiments, for different mapping modes of the same service, the first node may use different network models to perform encoding processing.

[0114] In some optional embodiments, the first node receives the first information sent by the base station, including: the first node receives a first medium access control (MAC) protocol data unit (PDU) sent by the base station, the first MAC PDU includes a first MAC CE, and the first MAC CE includes the first information.

[0115] In this embodiment, due to the presence of three mapping methods, each level of mapping corresponds to a different amount of transmitted data. The first mapping method corresponds to a smaller amount of data than the second mapping method, which in turn corresponds to a smaller amount of data than the third mapping method. Furthermore, different services and scenarios have different transmission requirements. Therefore, this embodiment of the present invention proposes a frame structure based on hierarchical semantic transmission, which can dynamically adjust the frame length to accommodate different transmission requirements.

[0116] In this embodiment, for downlink, a first MAC CE (also referred to as a downlink MAC CE) is used to carry the first information. The first MAC CE is located in a first MAC PDU (also referred to as a downlink MAC PDU).

[0117] In some optional embodiments, the first MAC CE is located at the beginning of the first MAC PDU.

[0118] In this embodiment, the first MAC CE is located at the beginning of the first MAC PDU. For details, see Figure 3a As shown, when the re-encoding flag is 1, it indicates that the first information requests or instructs the re-encoding of the second data. This placement is conducive to low-latency operation of the terminal device.

[0119] In some optional embodiments, the first node has a counting module for counting the received first information (or Figure 3a ) is counted. Exemplarily, when the first node sends the second data encoded using the first mapping method for the first time, the count of the counting module is 0. Due to the small amount of data, it is carried in the uplink MAC CE (the second MAC CE); when the first information is received once, that is, the recoding flag bit in the received first MAC CE is 1, the second data encoded using the second mapping method is sent and carried in the uplink MAC CE (the second MAC CE); when the first information is received again, that is, the recoding flag bit in the received first MAC CE is 1 twice, the second data encoded using the third mapping method is used.

[0120] In some optional embodiments, the second data is carried by a second MAC PDU.

[0121] In some optional embodiments, the second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

[0122] In this embodiment, refer to Figure 3b As shown, for uplink, a second MAC CE (also called uplink MAC CE) is used to carry the second data. The second MAC CE is located in a second MAC PDU (also called uplink MAC PDU). The second MAC CE includes counting information, such as Figure 3b The re-encoding count shown in represents the number of times the first information is received, or the number of times the second data is retransmitted, using binary notation 00, 01, or 10. For example, 00 represents 0 re-encoding retransmissions, 01 represents 1 re-encoding retransmission, and 10 represents 2 re-encoding retransmissions.

[0123] In some optional embodiments, the second MAC CE is located at the end of the second MAC PDU. Figure 3b shown.

[0124] In some optional embodiments, when the second data is encoded according to the first mapping method or the second mapping method, the second data is carried in the second MAC CE; when the second data is encoded according to the third mapping method, the second data is carried in other fields of the second MAC PDU except the second MAC CE.

[0125] In this embodiment, the data volume corresponding to the first mapping mode (high-level semantic coding) and the second mapping mode (intermediate semantic coding) is relatively small. In these two mapping modes, the corresponding data can be carried by the second MAC CE. The data volume corresponding to the third mapping mode (low-level semantic coding) is relatively large. In addition to the second MAC CE, other fields in the second MAC PDU are used to carry the data. For details, see Figure 3b shown.

[0126] Based on the above embodiments, an embodiment of the present invention further provides a data transmission method. Figure 4 Schematic diagram of the data transmission method according to an embodiment of the present invention Figure 2 ;like Figure 4 As shown, the method includes:

[0127] Step 201: The base station receives second data sent by the first node, decodes the second data based on a local second knowledge base, and obtains multiple semantic feature vectors if the decoding is successful.

[0128] Step 202: The base station drives related tasks based on the multiple semantic feature vectors, or converts the multiple semantic feature vectors based on the second knowledge base to obtain first data.

[0129] In this embodiment, the base station receives second data sent by the first node, where the second data is data obtained by encoding the semantic feature vector.

[0130] In this embodiment, refer to Figure 1 As shown, the base station can decode the second data through the source-channel decoding module in combination with the local second knowledge base to restore the semantic feature vector. The base station pre-establishes a local knowledge base (recorded as the second knowledge base), which can also be called a semantic knowledge base. It is a structured knowledge network model with memory capabilities that can provide relevant semantic knowledge descriptions for data information. Exemplarily, the knowledge base (recorded as the second knowledge base) includes multiple elements and the association relationship between elements; wherein the elements can be single words, phrases, etc.; the association relationship can specifically refer to a semantic relationship, such as a synonymous relationship, antonym relationship, superordinate / subordinate relationship, whole / part relationship, etc. In this embodiment, the knowledge base (recorded as the second knowledge base) can also include a neural network required for semantic synthesis conversion and decoding.

[0131] In this embodiment, the semantic feature vectors obtained by decoding by the base station can be directly used to drive related tasks, such as machine-type communication tasks. In other optional embodiments, the base station can also convert multiple semantic feature vectors in combination with the local second knowledge base, fuse the semantic features, and reconstruct the first data as the source data to drive human communication tasks. In this embodiment, refer to Figure 1 As shown, the base station can convert the semantic feature vector through the semantic synthesis conversion module to obtain the first data.

[0132] In some optional embodiments, converting the multiple semantic feature vectors based on the second knowledge base to obtain first data includes: the base station correcting errors of the multiple semantic feature vectors based on the second knowledge base, and converting the multiple semantic feature vectors after error correction to obtain first data.

[0133] In this embodiment, refer to Figure 1As shown, the semantic comprehensive conversion mainly includes two functions: semantic distortion correction and semantic feature fusion, both of which are nonlinear transformations. Among them, semantic distortion correction uses the correlation within and between semantic feature vectors to further correct the residual error left by source-channel decoding. Semantic feature fusion reconstructs the source data using the feature mapping after semantic distortion correction. Semantic feature fusion not only retains the effective identification information of multiple features involved in the fusion, but also eliminates the redundancy of information to a large extent, realizes information compression, and is conducive to real-time processing of information. The base station corrects the errors of multiple semantic feature vectors based on the local second knowledge base, and performs feature mapping on the multiple semantic feature vectors after error correction, thereby converting them into the first data.

[0134] In some optional embodiments of the present invention, the method further includes: the base station receiving a first parameter sent by the first node, where the first parameter represents a semantic value of the semantic feature vector.

[0135] In some optional embodiments, decoding the second data based on a local second knowledge base includes: the base station decoding the second data based on the second knowledge base and the first parameter.

[0136] In this embodiment, before encoding the semantic feature vector, the first node determines, based on the application scenario (or communication purpose) of the source data, that the source data belongs to a first type, where the first type is either a human perception-oriented type or a task accuracy-oriented type. The first node then evaluates the semantic value of the semantic feature vector based on the first type, thereby obtaining a first parameter for each semantic feature vector. Each semantic feature vector is then encoded based on the first parameter to obtain second data. During decoding of the second data, the base station decodes the second data based on the local second knowledge base and the first parameter.

[0137] In some optional embodiments, decoding the second data based on the local second knowledge base includes: decoding the second data by the base station based on a second neural network model in the second knowledge base. The second neural network is mainly used to convert the second data into multiple semantic feature vectors.

[0138] In some optional embodiments of the present invention, the method further includes: the base station determining or updating parameters of the second neural network model based on wireless channel environment information and / or wireless resource status information.

[0139] In this embodiment, the base station may collect wireless channel environment information and / or wireless resource status information, regularly optimize the local second neural network model, and train the model parameters of the second neural network model.

[0140] In some optional embodiments of the present invention, the method further includes: in case the decoding is unsuccessful, the base station sends the second data to other base stations, and receives processing results corresponding to the second data sent by the other base stations.

[0141] In this embodiment, a service may be pre-configured in the second knowledge base of the base station, and the service may be implemented through a network model. It can be understood that the second knowledge base of the base station may be pre-configured with a network model, and different network models may provide different services. Exemplarily, it may include model training services, reasoning services, modeling services, etc. Accordingly, the second knowledge base of the base station may be pre-configured with a network model for model training, a network model for reasoning, a network model for modeling, etc. Whether the second knowledge base of the base station has a first service (or network model) that can provide a match with the second data determines whether the decoding is successful.

[0142] As an implementation method, the base station determines whether it can or cannot provide the first service corresponding to the second data based on the second knowledge base; if the judgment result is that the first service corresponding to the second data can be provided, the second data is decoded according to the network model corresponding to the first service to obtain multiple semantic feature vectors; if the judgment result is that the first service corresponding to the second data cannot be provided, the base station determines that the second data cannot be decoded, that is, the decoding is unsuccessful; then, the base station sends the second data to other base stations to inquire whether the other base stations can provide the first service corresponding to the second data; if the other base stations can provide the first service corresponding to the second data, the other base stations decode the second data and receive the processing results sent by the other base stations.

[0143] In some optional embodiments, the method further includes: when the processing result is that the other base station cannot successfully decode the second data, the base station sends first information to the first node, and the first information is used to request or instruct re-encoding of the second data.

[0144] In this embodiment, when the base station side (including the base station and other base stations) cannot successfully decode the second data, that is, when the base station side (including the base station and other base stations) cannot provide the first service corresponding to the second data, the base station sends a first information to the first node to request or instruct the re-encoding of the second data.

[0145] In some optional embodiments of the present invention, the method further includes: the base station receiving the re-encoded second data sent by the first node.

[0146] In this embodiment, the re-encoded second data has a lower encoding level than the previously encoded second data, which means that the corresponding data volume is increased. As one embodiment, in step 201, if the second data is obtained by encoding the semantic feature vector using the first mapping method, then in this embodiment, the second data is obtained by encoding the semantic feature vector using the second mapping method. As another embodiment, if the second data is obtained by encoding the semantic feature vector using the second mapping method, then in this embodiment, the second data is obtained by encoding the semantic feature vector using the third mapping method.

[0147] In this embodiment, different mapping methods for the same service can be decoded and processed corresponding to different network models. Whether the second knowledge base of the base station has a first service (or network model) that can provide a match with the second data, and whether the first service (or network model) matches the mapping method determines whether the decoding can be successful. In other words, the second knowledge base can be divided into three levels: low-level, intermediate and advanced, and the second data is mapped to low-level, intermediate and advanced semantics respectively. That is, the base station decodes the second data based on the network model in the second knowledge base. If the corresponding network model does not exist in the second knowledge base, the decoding of the second data is unsuccessful; correspondingly, if the corresponding network model exists in the second knowledge base, the decoding of the second data is successful.

[0148] In some optional embodiments, knowledge can be shared between the first knowledge base of the first node and the second knowledge base of the base station, thereby reducing the training cost of the semantic model (network model).

[0149] In some optional embodiments, the base station sends the first information to the first node, including: the base station receives a first MAC PDU sent by the first node, the first MAC PDU includes a first MAC CE, and the first MAC CE includes the first information.

[0150] In this embodiment, for downlink, a first MAC CE (also referred to as a downlink MAC CE) is used to carry the first information. The first MAC CE is located in a first MAC PDU (also referred to as a downlink MAC PDU).

[0151] In some optional embodiments, the first MAC CE is located at the beginning of the first MAC PDU.

[0152] In this embodiment, the first MAC CE is located at the beginning of the first MAC PDU. For details, see Figure 3a As shown, when the re-encoding flag is 1, it indicates that the first information requests or instructs the re-encoding of the second data. This placement is conducive to low-latency operation of the terminal device.

[0153] In some optional embodiments, the second data is carried by a second MAC PDU.

[0154] In some optional embodiments, the second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

[0155] In some optional embodiments, the first node has a counting module for counting the received first information (or Figure 3a ) is counted. Exemplarily, when the first node sends the second data encoded using the first mapping method for the first time, the count of the counting module is 0. Due to the small amount of data, it is carried in the uplink MAC CE (the second MAC CE); when the first information is received once, that is, the recoding flag bit in the received first MAC CE is 1, the second data encoded using the second mapping method is sent and carried in the uplink MAC CE (the second MAC CE); when the first information is received again, that is, the recoding flag bit in the received first MAC CE is 1 twice, the second data encoded using the third mapping method is used.

[0156] In this embodiment, refer to Figure 3b As shown, for uplink, a second MAC CE (also called uplink MAC CE) is used to carry the second data. The second MAC CE is located in a second MAC PDU (also called uplink MAC PDU). The second MAC CE includes counting information, such as Figure 3b The re-encoding count shown in represents the number of times the first information is received, or the number of times the second data is retransmitted, using binary notation 00, 01, or 10. For example, 00 represents 0 re-encoding retransmissions, 01 represents 1 re-encoding retransmission, and 10 represents 2 re-encoding retransmissions.

[0157] In some optional embodiments, the second MAC CE is located at the end of the second MAC PDU. Figure 3b shown.

[0158] In some optional embodiments, when the second data is encoded according to the first mapping method or the second mapping method, the second data is carried in the second MAC CE; when the second data is encoded according to the third mapping method, the second data is carried in other fields of the second MAC PDU except the second MAC CE; wherein the data amount of the second data encoded using the first mapping method, the second mapping method and the third mapping method respectively increases successively; and / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

[0159] In this embodiment, since there are three mapping methods, the amount of data transmitted corresponding to each level of mapping method is different. The amount of data corresponding to the first mapping method is smaller than the amount of data corresponding to the second mapping method, and the amount of data corresponding to the second mapping method is smaller than the amount of data corresponding to the third mapping method.

[0160] In this embodiment, the data volume corresponding to the first mapping mode (high-level semantic coding) and the second mapping mode (intermediate semantic coding) is relatively small. In these two mapping modes, the corresponding data can be carried by the second MAC CE. The data volume corresponding to the third mapping mode (low-level semantic coding) is relatively large. In addition to the second MAC CE, other fields in the second MAC PDU are used to carry the data. For details, see Figure 3b shown.

[0161] Adopting the technical solution of the embodiment of the present invention, on the first aspect, a first node is introduced as a model deployment node for semantic coding of UE original data, and the strong computing and storage capabilities of the first node are utilized to establish a local first knowledge base at the first node for the conversion of original data to semantic symbols, and introduce functions such as source-channel semantic joint coding to adapt to the transmission of semantic signals in the channel. On the second aspect, through hierarchical semantic transmission, three levels of mapping can be performed in terms of semantics: high-level (first mapping method), intermediate (second mapping method), and low-level (third mapping method). The selection of the coding level can be adjusted through multiple interactions between the first node and the base station, so that the sender and receiver can correctly understand the semantic intention and improve the transmission efficiency. In addition, this method can also realize the update of the network model in the semantic library.

[0162] Figure 5 FIG. 1 is a schematic diagram of an interactive flow of a data transmission method according to an embodiment of the present invention; FIG. Figure 5 As shown, the method includes:

[0163] Step 301: A first node receives first data and a service request from a UE, where the service request is used to request a first service.

[0164] Here, the service request may include a service identifier or a model identifier, for example, an identifier of the first service.

[0165] Here, after the first node receives the first data and the service request from the UE, the retransmission counter is reset to zero.

[0166] Step 302: The first node performs semantic analysis on the first data based on the first knowledge base, obtains a plurality of semantic feature vectors, and determines the semantic value of each semantic feature vector;

[0167] Step 303: The first node queries the local first knowledge base to find a matching AI service or model.

[0168] Here, the first node may query the first knowledge base based on the service identifier or model identifier to determine whether there is an AI service or model matching the service identifier or model identifier in the first knowledge base. If there is no AI service or model matching the service identifier or model identifier in the first knowledge base, then execute step 304: encode the semantic feature vector using a third mapping method (low-level semantics) to obtain second data, and send the second data to the base station; if there is an AI service or model matching the service identifier or model identifier in the first knowledge base, execute step 305.

[0169] Step 305: The first node uses the first mapping method to perform source-channel coding on the semantic feature vector in combination with the semantic value to obtain second data, and transmits the second data to the base station.

[0170] Here, the first node uses the first mapping method to directly abstract the AI ​​service requirements accurately, then performs source-channel encoding and transmits it to the base station.

[0171] Step 306: The base station determines whether the AI ​​service corresponding to the second data can be provided based on the local second knowledge base; if the judgment result is that the AI ​​service corresponding to the second data can be provided, execute step 307; if the judgment result is that the AI ​​service corresponding to the second data cannot be provided, execute step 308.

[0172] Here, the second data may include a service identifier or a model identifier, or in addition to sending the second identifier to the base station, the first node also sends a service request to the base station, and the service request may include a service identifier or a model identifier. The base station can then query the local second knowledge base based on the service identifier or model identifier to determine whether there is an AI service or model matching the service identifier or model identifier in the second knowledge base. If there is no AI service or model matching the service identifier or model identifier in the second knowledge base, it means that the base station cannot provide the AI ​​service corresponding to the second data; if there is an AI service or model matching the service identifier or model identifier in the second knowledge base, it means that the base station can provide the AI ​​service corresponding to the second data.

[0173] Step 307: The base station decodes the second data based on the AI ​​model of the second knowledge base, obtains multiple semantic feature vectors, and sends feedback results to the first node.

[0174] Here, the recoding flag is displayed as 0, indicating that there is an AI model in the local second knowledge base of the base station that has the same function as the first knowledge base in the first node and can identify the service requirements, and no recoding and retransmission are required.

[0175] Steps 308 to 309: the base station sends a broadcast message, which may include the second data; and receives feedback results sent by other base stations.

[0176] Here, there are two situations:

[0177] a. If there is at least one other base station in the network that can provide the service, the other base station with the closest physical location will use the second database local to the other base station to decode the second data, obtain multiple semantic feature vectors, and send feedback information to the base station. The feedback information includes the decoding processing results, that is, the feedback information includes multiple semantic feature vectors obtained by decoding; in this case, the base station sends feedback information to the first node. Here, the recoding flag is displayed as 0, indicating that the second knowledge base local to the base station has an AI model that has the same function as the first knowledge base in the first node and can identify the service requirements, and no recoding and retransmission are required.

[0178] b. If other base stations in the network cannot provide the service, the other base stations send feedback information to the base station, and the feedback information indicates that the service cannot be provided. In this case, step 310 is executed.

[0179] Step 310: The base station sends a re-encoding and retransmission instruction to the first node.

[0180] Here, the re-encoding and retransmission instruction is equivalent to the first information in the above embodiment. The re-encoding and re-transmission instruction is carried by the downlink MAC CE. In the downlink MAC CE, a re-encoding flag bit 1 is added to indicate that the base station failed to identify the service and requires the first node to re-encode and retransmit.

[0181] In some optional embodiments, after receiving the re-encoding and retransmission instructions, the first node uploads the corresponding AI model parameters to assist the base station in updating the model knowledge base.

[0182] Here, the recoding flag is displayed as 1, indicating that there is no AI model in the local second knowledge base of the base station that has the same function as the first knowledge base in the first node and can identify the service requirement, and it needs to be re-encoded and retransmitted.

[0183] Step 311: The first node receives the re-encoding and retransmission instruction for the first time, performs source-channel coding using a second mapping method to obtain second data, and transmits the second data to the base station.

[0184] Here, the retransmission counter counts up by 1.

[0185] Step 312: The base station determines again whether the AI ​​service corresponding to the second data can be provided based on the local second knowledge base; if the judgment result is that the AI ​​service corresponding to the second data can be provided, execute step 313; if the judgment result is that the AI ​​service corresponding to the second data cannot be provided, execute steps 314-315.

[0186] Step 313: The base station decodes the second data based on the AI ​​model of the second knowledge base, obtains multiple semantic feature vectors, and sends feedback results to the first node.

[0187] Steps 314 to 315: the base station sends a broadcast message, which may include the second data; and receives feedback results sent by other base stations.

[0188] Similar to the above steps 308 and 309, two situations are included. In case b, that is, if other base stations in the network cannot provide the service, the other base stations send feedback information to the base station, indicating that the service cannot be provided. In this case, step 316 is executed.

[0189] Step 316: The base station sends a re-encoding and retransmission instruction to the first node again.

[0190] Here, the recoding flag is displayed as 1, indicating that there is no AI model in the local second knowledge base of the base station that has the same function as the first knowledge base in the first node and can identify the service requirement, and it needs to be re-encoded and retransmitted.

[0191] Step 317: The first node receives the re-encoding and retransmission instruction again, performs source-channel coding using the third mapping method to obtain second data, and transmits the second data to the base station.

[0192] Here, the retransmission counter is incremented by 1.

[0193] Step 318: The base station receives and processes the second data, and feeds back the processing result to the first node.

[0194] Take semantic transfer of image classification tasks as an example:

[0195] First, the UE sends a picture of a zebra to the first node, and obtains a multi-level semantic feature vector of the picture based on a first knowledge base locally stored in the first node.

[0196] The high-level semantic feature vector is represented as "zebra".

[0197] The mid-level semantic feature vector describes the attributes of a zebra, which may include:

[0198] Color: black and white;

[0199] Silhouette: Horse;

[0200] Stripes: Yes;

[0201] The low-level semantic feature vector contains the original information of the image and is a pixel-level feature vector.

[0202] If the first node has a model corresponding to the image classification task, the model is used to perform source-channel encoding on the high-level semantic feature vector. After transmitting it to the base station through the channel, the semantic feature vector is obtained through source-channel decoding, and the mapping of the feature vector to the source data is realized using the second knowledge base local to the base station.

[0203] If the second local knowledge base of the base station can identify or parse the semantic description of the high-level semantic feature vector, that is, can identify the semantic feature description related to "zebra", then the task is completed.

[0204] Otherwise, the base station sends a semantic recognition request to other base stations. If the "zebra" semantics can be recognized in the knowledge base of other base stations, the task is completed; if the "zebra" semantics cannot be recognized in the knowledge base of other base stations, the base station requests the central node to transmit the intermediate semantic feature vector.

[0205] If the second knowledge base of the base station can parse the semantic attribute description of "zebra" as an intermediate semantic feature vector, the task is successful.

[0206] Otherwise, the base station queries other base stations. If other base stations still cannot recognize the semantics of the intermediate semantic feature vector, the base station requests the first node to send the low-level semantic feature vector, that is, the original image information, for the base station to understand the image.

[0207] Based on the above embodiment, an embodiment of the present invention further provides a data transmission device, which is applied to a first node. Figure 6 Schematic diagram of the structure of the data transmission device according to an embodiment of the present invention Figure 1 ;like Figure 6 As shown, the device includes: a first processing unit 11 and a first communication unit 12; wherein,

[0208] The first processing unit 11 is configured to perform semantic analysis on the first data based on a local first knowledge base to obtain a plurality of semantic feature vectors and determine a first parameter for each semantic feature vector; the first data is from a terminal device; the first parameter represents a semantic value of the semantic feature vector; and is further configured to encode each semantic feature vector based on the first parameter and the first knowledge base to obtain second data;

[0209] The first communication unit 12 is configured to send the second data to a base station.

[0210] In some optional embodiments of the present invention, the first processing unit 11 is used to determine the first type corresponding to the first data based on the first knowledge base, evaluate the semantic value of each semantic feature vector based on the first type, and determine the first parameter of each semantic feature vector; the first parameter represents the semantic value of the semantic feature vector corresponding to the first type; wherein the first type is a human perception-oriented type or a task accuracy-oriented type.

[0211] In some optional embodiments of the present invention, the first communication unit 12 is further configured to send the first parameter to the base station.

[0212] In some optional embodiments of the present invention, the first processing unit 11 is used to determine whether the first service corresponding to the first data can be provided based on the first knowledge base, determine a mapping method based on the judgment result, and encode each semantic feature vector based on the mapping method and the first parameter.

[0213] In some optional embodiments of the present invention, the mapping mode includes a first mapping mode, a second mapping mode, or a third mapping mode; wherein the data amounts of the second data encoded by respectively adopting the first mapping mode, the second mapping mode, and the third mapping mode increase in sequence;

[0214] And / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

[0215] In some optional embodiments of the present invention, the first processing unit 11 is used to determine a first mapping method when the judgment result is that the first service corresponding to the first data can be provided; and determine a third mapping method when the judgment result is that the first service corresponding to the first data cannot be provided.

[0216] In some optional embodiments of the present invention, the first processing unit 11 is configured to encode each semantic feature vector based on the mapping method, the first parameter, and the first neural network model.

[0217] In some optional embodiments of the present invention, the first processing unit 11 is further used to determine or update the parameters of the first neural network model based on wireless channel environment information and / or wireless resource status information.

[0218] In some optional embodiments of the present invention, the first communication unit 12 is further configured to receive first information sent by the base station after sending the second data to the base station, where the first information is used to request or instruct re-encoding of the second data.

[0219] In some optional embodiments of the present invention, the first processing unit 11 is further configured to, when the mapping mode is the first mapping mode, re-encode each semantic feature vector according to the second mapping mode to obtain re-encoded second data;

[0220] The first communication unit 12 is further configured to send the re-encoded second data to the base station.

[0221] In some optional embodiments of the present invention, the first processing unit 11 is further configured to, when the mapping mode is the second mapping mode, re-encode each semantic feature vector according to the third mapping mode to obtain re-encoded second data;

[0222] The first communication unit 12 is further configured to send the re-encoded second data to the base station.

[0223] In some optional embodiments of the present invention, the first communication unit 12 is configured to receive a first MAC PDU sent by the base station, where the first MAC PDU includes a first MAC CE, and the first MAC CE includes the first information.

[0224] In some optional embodiments of the present invention, the first MAC CE is located at the beginning of the first MAC PDU.

[0225] In some optional embodiments of the present invention, the second data is carried by a second MAC PDU.

[0226] In some optional embodiments of the present invention, the second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

[0227] In some optional embodiments of the present invention, the second MAC CE is located at the end of the second MAC PDU.

[0228] In some optional embodiments of the present invention, when the second data is encoded according to the first mapping mode or the second mapping mode, the second data is carried in the second MAC CE;

[0229] When the second data is obtained by encoding according to the third mapping manner, the second data is carried in other fields of the second MAC PDU except the second MAC CE.

[0230] In an embodiment of the present invention, the first processing unit 11 in the device can be implemented by a central processing unit (CPU), a digital signal processor (DSP), a microcontroller unit (MCU) or a programmable gate array (FPGA) in actual applications; the first communication unit 12 in the device can be implemented by a communication module (including: basic communication kit, operating system, communication module, standardized interface and protocol, etc.) and a transceiver antenna in actual applications.

[0231] An embodiment of the present invention further provides a data transmission device, which is applied to a base station. Figure 7 Schematic diagram of the structure of the data transmission device according to an embodiment of the present invention Figure 2 ;like Figure 7 As shown, the device includes: a second communication unit 21 and a second processing unit 22; wherein,

[0232] The second communication unit 21 is configured to receive second data sent by the first node;

[0233] The second processing unit 22 is used to decode the second data based on a local second knowledge base, and obtain multiple semantic feature vectors if the decoding is successful; drive related tasks based on the multiple semantic feature vectors, or convert the multiple semantic feature vectors based on the second knowledge base to obtain first data.

[0234] In some optional embodiments of the present invention, the second processing unit 22 is used to correct errors of the multiple semantic feature vectors based on the second knowledge base, and convert the multiple semantic feature vectors after error correction to obtain first data.

[0235] In some optional embodiments of the present invention, the second communication unit 21 is further configured to receive a first parameter sent by the first node, where the first parameter represents a semantic value of the semantic feature vector.

[0236] In some optional embodiments of the present invention, the second processing unit 22 is configured to decode the second data based on the second knowledge base and the first parameter.

[0237] In some optional embodiments of the present invention, the second processing unit 22 is used to decode the second data based on the second neural network model in the second knowledge base.

[0238] In some optional embodiments of the present invention, the second processing unit 22 is further used to determine or update the parameters of the second neural network model based on wireless channel environment information and / or wireless resource status information.

[0239] In some optional embodiments of the present invention, the second communication unit 21 is further configured to send the second data to other base stations and receive processing results corresponding to the second data sent by the other base stations when decoding by the second processing unit 22 is unsuccessful.

[0240] In some optional embodiments of the present invention, the second communication unit 21 is further used to send first information to the first node when the processing result is that the other base station cannot successfully decode the second data, and the first information is used to request or instruct re-encoding of the second data.

[0241] In some optional embodiments of the present invention, the second communication unit 21 is further configured to receive the re-encoded second data sent by the first node.

[0242] In some optional embodiments of the present invention, the second communication unit 21 is configured to receive a first MAC PDU sent by the first node, where the first MAC PDU includes a first MAC CE, and the first MAC CE includes the first information.

[0243] In some optional embodiments of the present invention, the first MAC CE is located at the beginning of the first MAC PDU.

[0244] In some optional embodiments of the present invention, the second data is carried by a second MAC PDU.

[0245] In some optional embodiments of the present invention, the second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

[0246] In some optional embodiments of the present invention, the second MAC CE is located at the end of the second MAC PDU.

[0247] In some optional embodiments of the present invention, when the second data is encoded according to the first mapping mode or the second mapping mode, the second data is carried in the second MAC CE;

[0248] When the second data is obtained by encoding according to the third mapping manner, the second data is carried in other fields except the second MAC CE in the second MAC PDU;

[0249] The amounts of the second data encoded by respectively adopting the first mapping mode, the second mapping mode and the third mapping mode increase in sequence;

[0250] And / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

[0251] In an embodiment of the present invention, the second processing unit 22 in the device can be implemented by a CPU, DSP, MCU or FPGA in actual applications; the second communication unit 21 in the device can be implemented by a communication module (including: basic communication kit, operating system, communication module, standardized interface and protocol, etc.) and a transceiver antenna in actual applications.

[0252] It should be noted that the data transmission device provided in the above embodiment is only illustrated by the division of the above-mentioned program modules when performing data transmission. In actual applications, the above-mentioned processing can be assigned to different program modules as needed, that is, the internal structure of the device can be divided into different program modules to complete all or part of the above-mentioned processing. In addition, the data transmission device provided in the above embodiment and the data transmission method embodiment are based on the same concept. The specific implementation process is detailed in the method embodiment and will not be repeated here.

[0253] An embodiment of the present invention further provides a communication device, which is a first node or a base station. Figure 8 FIG. 1 is a schematic diagram of the hardware structure of a communication device according to an embodiment of the present invention. Figure 8 As shown, the communication device includes a memory 32, a processor 31, and a computer program stored in the memory 32 and executable on the processor 31. When the processor 31 executes the program, the steps of the data transmission method applied to the first node or base station in an embodiment of the present invention are implemented.

[0254] Optionally, the communication device may further include at least one network interface 33. The various components in the communication device are coupled together via a bus system 34. It is understood that the bus system 34 is used to achieve connection and communication between these components. In addition to the data bus, the bus system 34 also includes a power bus, a control bus, and a status signal bus. However, for the sake of clarity, Figure 8 Various buses are labeled as bus system 34 .

[0255] It is understood that the memory 32 can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD-ROM); the magnetic surface memory can be a magnetic disk or a magnetic tape. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example and not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct rambus random access memory (DRRAM).The memory 32 described in the embodiments of the present invention is intended to include, but is not limited to, these and any other suitable types of memory.

[0256] The methods disclosed in the above embodiments of the present invention can be applied to or implemented by processor 31. Processor 31 may be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits in processor 31 or by software instructions. The above processor 31 may be a general-purpose processor, a DSP, or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, etc. Processor 31 can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present invention. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of the present invention can be directly implemented and executed by a hardware decoding processor, or by a combination of hardware and software modules in the decoding processor. The software module may be located in a storage medium located in memory 32. Processor 31 reads information from memory 32 and, in conjunction with its hardware, completes the steps of the above method.

[0257] In an exemplary embodiment, the communication device may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), FPGAs, general-purpose processors, controllers, MCUs, microprocessors, or other electronic components to perform the aforementioned method.

[0258] In an exemplary embodiment, the present invention further provides a computer-readable storage medium, such as a memory 32 including a computer program. The computer program can be executed by the processor 31 of the communication device to perform the steps of the aforementioned method. The computer-readable storage medium can be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface mount storage, optical disk, or CD-ROM; or various devices including any one or any combination of the aforementioned memories.

[0259] The computer-readable storage medium provided by the embodiment of the present invention stores a computer program thereon, which, when executed by a processor, implements the steps of the data transmission method applied to a first node or a base station according to the embodiment of the present invention.

[0260] An embodiment of the present application further provides a computer program product, including a computer program, which can be executed by a computer (such as the processor 31 of a communication device) to complete the steps of any of the aforementioned data transmission methods.

[0261] The methods disclosed in the several method embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments.

[0262] The features disclosed in the several product embodiments provided in this application can be arbitrarily combined without conflict to obtain new product embodiments.

[0263] The features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method embodiments or device embodiments.

[0264] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely schematic. For example, the division of the units is merely a logical function division. In actual implementation, there may be other division methods, such as: multiple units or components can be combined, or can be integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the components shown or discussed can be through some interfaces, and the indirect coupling or communication connection of the devices or units can be electrical, mechanical or other forms.

[0265] The units described above as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units; some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0266] In addition, all functional units in the embodiments of the present invention may be integrated into one processing unit, or each unit may be separately used as a unit, or two or more units may be integrated into one unit; the above-mentioned integrated units may be implemented in the form of hardware or in the form of hardware plus software functional units.

[0267] Those skilled in the art will understand that all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the aforementioned storage medium includes: mobile storage devices, ROM, RAM, disks or optical disks, etc. Various media that can store program codes.

[0268] Alternatively, if the above-mentioned integrated unit of the present invention is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the embodiment of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the methods described in each embodiment of the present invention. The aforementioned storage medium includes: various media that can store program codes, such as mobile storage devices, ROM, RAM, magnetic disks or optical disks.

[0269] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A data transmission method, characterized in that: The method is applied to a first node, and includes: The first node performs semantic analysis on the first data based on a local first knowledge base to obtain a plurality of semantic feature vectors and determine a first parameter of each semantic feature vector; the first data comes from a terminal device; and the first parameter represents a semantic value of the semantic feature vector; The first node encodes each semantic feature vector based on the first parameter and the first knowledge base to obtain second data, and sends the second data to the base station.

2. The method according to claim 1, characterized in that The determining of the first parameter of each semantic feature vector includes: Based on the first knowledge base, the first node determines the first type corresponding to the first data, evaluates the semantic value of each semantic feature vector based on the first type, and determines the first parameter of each semantic feature vector; the first parameter represents the semantic value of the semantic feature vector corresponding to the first type; wherein, the first type is a human perception-oriented type or a task accuracy-oriented type.

3. The method according to claim 1, characterized in that The method further comprises: The first node sends the first parameter to the base station.

4. The method according to claim 1, wherein The first node encodes each semantic feature vector based on the first parameter and the first knowledge base, including: The first node determines whether it can provide the first service corresponding to the first data based on the first knowledge base, determines a mapping method based on the determination result, and encodes each semantic feature vector based on the mapping method and the first parameter.

5. The method according to claim 4, characterized in that The mapping mode includes a first mapping mode, a second mapping mode or a third mapping mode; The amounts of the second data encoded by respectively adopting the first mapping mode, the second mapping mode and the third mapping mode increase in sequence; And / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

6. The method according to claim 4 or 5, characterized in that The determining of the mapping mode based on the judgment result includes: If the result of the determination is that the first service corresponding to the first data can be provided, determining a first mapping mode; If the judgment result is that the first service corresponding to the first data cannot be provided, a third mapping method is determined.

7. The method according to claim 4, characterized in that The encoding of each semantic feature vector based on the mapping mode and the first parameter includes: The first node encodes each semantic feature vector based on the mapping method, the first parameter and the first neural network model.

8. The method according to claim 7, characterized in that The method further comprises: The first node determines or updates parameters of the first neural network model based on wireless channel environment information and / or wireless resource status information.

9. The method according to claim 5, characterized in that After sending the second data to the base station, the method further includes: The first node receives first information sent by the base station, where the first information is used to request or instruct re-encoding of the second data.

10. The method according to claim 9, characterized in that When the mapping mode is the first mapping mode, the method further includes: The first node re-encodes each semantic feature vector according to the second mapping manner to obtain re-encoded second data, and sends the re-encoded second data to the base station.

11. The method according to claim 9, characterized in that When the mapping mode is the second mapping mode, the method further includes: The first node re-encodes each semantic feature vector according to the third mapping manner to obtain re-encoded second data, and sends the re-encoded second data to the base station.

12. The method according to claim 9, characterized in that The first node receiving the first information sent by the base station includes: The first node receives a first medium access control protocol data unit (MAC PDU) sent by the base station, where the first MAC PDU includes a first medium access control element (MAC CE), and the first medium access control element (MAC CE) includes the first information.

13. The method according to claim 12, characterized in that The first medium access control element MAC CE is located at the beginning of the first medium access control protocol data unit MAC PDU.

14. The method according to claim 10 or 11, characterized in that The second data is carried by a second MAC PDU.

15. The method according to claim 14, characterized in that The second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

16. The method according to claim 15, characterized in that The second MAC CE is located at the end of the second MAC PDU.

17. The method according to claim 15, characterized in that When the second data is obtained by encoding according to the first mapping mode or the second mapping mode, the second data is carried in the second MAC CE; When the second data is obtained by encoding according to the third mapping method, the second data is carried in other fields of the second MAC PDU except the second MAC CE.

18. A data transmission method, characterized in that: The method is applied to a base station, and the method includes: The base station receives the second data sent by the first node, decodes the second data based on a local second knowledge base, and obtains a plurality of semantic feature vectors when the decoding is successful; The base station drives related tasks based on the multiple semantic feature vectors, or converts the multiple semantic feature vectors based on the second knowledge base to obtain first data.

19. The method according to claim 18, characterized in that The converting the plurality of semantic feature vectors based on the second knowledge base to obtain first data includes: The base station corrects errors of the multiple semantic feature vectors based on the second knowledge base, and converts the multiple semantic feature vectors after error correction to obtain first data.

20. The method according to claim 18, wherein The method further comprises: The base station receives a first parameter sent by the first node, where the first parameter represents a semantic value of a semantic feature vector.

21. The method according to claim 20, characterized in that The decoding the second data based on the local second knowledge base includes: The base station decodes the second data based on the second knowledge base and the first parameter.

22. The method according to claim 18, wherein The decoding the second data based on the local second knowledge base includes: The base station decodes the second data based on a second neural network model in the second knowledge base.

23. The method according to claim 22, characterized in that The method further comprises: The base station determines or updates parameters of the second neural network model based on wireless channel environment information and / or wireless resource status information.

24. The method according to claim 18, wherein The method further comprises: In the case that the decoding is unsuccessful, the base station sends the second data to other base stations, and receives a processing result corresponding to the second data sent by the other base stations.

25. The method according to claim 24, characterized in that The method further comprises: If the processing result is that the other base station cannot successfully decode the second data, the base station sends first information to the first node, where the first information is used to request or instruct re-encoding of the second data.

26. The method according to claim 25, characterized in that The method further comprises: The base station receives the re-encoded second data sent by the first node.

27. The method according to claim 25, characterized in that The base station sending first information to the first node includes: The base station receives a first MAC PDU sent by the first node, where the first MAC PDU includes a first MAC CE, and the first MAC CE includes the first information.

28. The method according to claim 27, characterized in that The first MAC CE is located at the beginning of the first MAC PDU.

29. The method according to claim 18 or 26, characterized in that The second data is carried by a second MAC PDU.

30. The method according to claim 29, wherein The second MAC PDU includes a second MAC CE, the second MAC CE includes a first identifier, and the first identifier indicates the number of retransmissions.

31. The method according to claim 30, wherein The second MAC CE is located at the end of the second MAC PDU.

32. The method according to claim 29, wherein When the second data is obtained by encoding according to the first mapping mode or the second mapping mode, the second data is carried in the second MAC CE; When the second data is obtained by encoding according to the third mapping manner, the second data is carried in other fields except the second MAC CE in the second MAC PDU; The amounts of the second data encoded by respectively adopting the first mapping mode, the second mapping mode and the third mapping mode increase in sequence; And / or, the second data encoded using the first mapping method represents semantics, the second data encoded using the second mapping method represents semantic attribute description information, and the second data encoded using the third mapping method includes the first data.

33. A data transmission device, characterized in that: The device is applied to a first node, and includes: a first processing unit and a first communication unit; wherein, The first processing unit is configured to perform semantic analysis on the first data based on a local first knowledge base to obtain a plurality of semantic feature vectors and determine a first parameter for each semantic feature vector; the first data is from a terminal device; the first parameter represents a semantic value of the semantic feature vector; and is further configured to encode each semantic feature vector based on the first parameter and the first knowledge base to obtain second data; The first communication unit is used to send the second data to the base station.

34. A data transmission device, characterized in that The device is applied to a base station, and includes: a second communication unit and a second processing unit; wherein, The second communication unit is configured to receive second data sent by the first node; The second processing unit is used to decode the second data based on a local second knowledge base, and obtain multiple semantic feature vectors if the decoding is successful; drive related tasks based on the multiple semantic feature vectors, or convert the multiple semantic feature vectors based on the second knowledge base to obtain first data.

35. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 17 are implemented; or When the program is executed by a processor, the steps of the method according to any one of claims 18 to 32 are implemented.

36. A communication device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the method described in any one of claims 1 to 17 are implemented; or, when the processor executes the program, the steps of the method described in any one of claims 18 to 32 are implemented.

37. A computer program product, characterized in that The method comprises computer program instructions, which enable a computer to execute the steps of the method according to any one of claims 1 to 17; or, the computer program instructions enable a computer to execute the steps of the method according to any one of claims 18 to 32.

Images