A model data management method, a model data management device, and a storage medium

Abstract

The present disclosure relates to a model data management method, a model data management device and a storage medium. The model data management method is applied to a radio access network device, and the method comprises: determining a model task completion state of a terminal in response to the terminal switching the radio access network device; and determining a first radio access network device for transmitting model data according to the model task completion state. The present disclosure can solve the problem that the wireless network AI cannot perform model training or the training result cannot be effectively delivered in the terminal high-speed moving scenario.

Description

Technical Field

[0001] This disclosure relates to the field of wireless communication technology, and in particular to a model data management method, a model data management device, and a storage medium. Background Technology

[0002] Artificial intelligence, such as machine learning or deep learning, requires a large amount of data for model training and inference to obtain a high-precision network model, which then provides accurate decision recommendations to the terminal. The process is as follows: After obtaining the trained model, Operation, Administration and Maintenance (OAM) sends it to the radio access network (RAN) device. The RAN device performs model inference and sends the inference results to the terminal. The terminal then executes the decision-making task based on the received inference results.

[0003] However, due to the high mobility of terminals, if a terminal moves before receiving the model inference results, it will be unable to receive the inference results. In other words, the wireless access device cannot deliver the model inference results to the requesting terminal. This forces the terminal to re-request the model subscription from the wireless access network device, resulting in wasted resources and increased network load. Summary of the Invention

[0004] To overcome the problems existing in related technologies, this disclosure provides a model data management method, a model data management device, and a storage medium.

[0005] According to a first aspect of the present disclosure, a model data management method is provided, applied to a wireless access network device, the method comprising:

[0006] In response to the terminal switching radio access network devices, the model task completion status of the terminal is determined; based on the model task completion status, the first radio access network device for transmitting model data is determined.

[0007] In one embodiment, the wireless access network device that the terminal switches to is a distributed wireless access network device;

[0008] The step of determining the first wireless access network device for transmitting model data based on the model task completion status includes:

[0009] In response to the terminal's model task completion status being "model training task not completed", the distributed radio access network device to which the terminal is switched is determined to be the first radio access network device.

[0010] In one embodiment, the model data includes supplementary model training data;

[0011] The method further includes:

[0012] In response to the wireless access network device switching to the terminal, the distributed wireless access network device acquires the model training supplementary data; and sends the model training supplementary data to the Operation and Maintenance Management (OAM), the model training supplementary data being used by the OAM to continue training the terminal's model.

[0013] In one embodiment, the wireless access network device that the terminal switches to is a distributed wireless access network device;

[0014] The step of determining the first wireless access network device for transmitting model data based on the model task completion status includes:

[0015] In response to the terminal's model task completion status being "model inference task not completed", the controlling wireless access network device is determined to be the first wireless access network device.

[0016] In one embodiment, the model data includes model inference result data;

[0017] The method further includes:

[0018] In response to the completion of the model inference task by the control radio access network device, the model inference result data is determined; the model inference result data is sent to the distributed radio access network device to which the terminal is switched.

[0019] In one embodiment, the wireless access network device switched by the terminal is a control wireless access network device;

[0020] The step of determining the first wireless access network device for transmitting model data based on the model task completion status includes:

[0021] In response to the terminal's model task completion status being "model training task not completed", the control radio access network device for terminal switching is determined to be the first radio access network device.

[0022] In one embodiment, the model data includes supplementary model training data;

[0023] The method further includes:

[0024] In response to the wireless access network device controlling the wireless access network device, the device acquires the model training supplementary data; and sends the model training supplementary data to the OAM, the model training supplementary data being used by the OAM to continue training the terminal's model.

[0025] In one embodiment, the wireless access network device switched by the terminal is a control wireless access network device;

[0026] The step of determining the first wireless access network device for transmitting model data based on the model task completion status includes:

[0027] In response to the terminal's model task completion status being "model inference task not completed", the terminal source control wireless access network device is determined to be the first wireless access network device.

[0028] In one embodiment, the model data includes model inference result data;

[0029] The method further includes:

[0030] In response to the completion of the model inference task by the terminal source control radio access network device, the model inference result data is determined; the model inference result data is sent to the control radio access network device for terminal switching.

[0031] In one embodiment, the method further includes:

[0032] In response to the fact that the radio access network device is the first radio access network device, a model subscription request is sent to the OAM, the model subscription request being used to request the OAM to update the terminal's information.

[0033] According to a second aspect of the present disclosure, a model data management method is provided, applied to an OAM entity, the method comprising:

[0034] In response to a terminal switching a radio access network device, the terminal receives model data transmitted by a first radio access network device, the first radio access network device determining the model task completion status based on the terminal; and trains the terminal's model based on the model data.

[0035] In one embodiment, the model data includes supplementary model training data;

[0036] The training of the terminal request model based on the model data includes:

[0037] Obtain the local model training data of the OAM; train the terminal model based on the local model training data and model training supplementary data.

[0038] In one embodiment, the method further includes:

[0039] Receive a model subscription request sent by a first wireless access network device; update the terminal's information based on the model subscription request.

[0040] According to a third aspect of the present disclosure, a model data management apparatus is provided, applied to a wireless access network device, the apparatus comprising:

[0041] The determination module is used to determine the model task completion status of the terminal in response to the terminal switching radio access network devices; and to determine the first radio access network device for transmitting model data based on the model task completion status.

[0042] In one embodiment, the wireless access network device that the terminal switches to is a distributed wireless access network device;

[0043] The determining module is used for:

[0044] In response to the terminal's model task completion status being "model training task not completed", the distributed radio access network device to which the terminal is switched is determined to be the first radio access network device.

[0045] In one embodiment, the model data includes supplementary model training data;

[0046] The device further includes: an acquisition module;

[0047] The acquisition module is configured to acquire the model training supplementary data in response to the distributed radio access network device switching for the terminal by the radio access network device; and send the model training supplementary data to the operation and maintenance management (OAM), wherein the model training supplementary data is used by the OAM to continue training the model of the terminal.

[0048] In one embodiment, the wireless access network device that the terminal switches to is a distributed wireless access network device;

[0049] The determining module is used for:

[0050] In response to the terminal's model task completion status being "model inference task not completed", the controlling wireless access network device is determined to be the first wireless access network device.

[0051] In one embodiment, the model data includes model inference result data;

[0052] The determining module is further configured to:

[0053] In response to the completion of the model inference task by the control radio access network device, the model inference result data is determined; the model inference result data is sent to the distributed radio access network device to which the terminal is switched.

[0054] In one embodiment, the wireless access network device switched by the terminal is a control wireless access network device;

[0055] The determining module is used for:

[0056] In response to the terminal's model task completion status being "model training task not completed", the control radio access network device for terminal switching is determined to be the first radio access network device.

[0057] In one embodiment, the model data includes supplementary model training data;

[0058] The acquisition module is also used for:

[0059] In response to the wireless access network device controlling the wireless access network device, the device acquires the model training supplementary data; and sends the model training supplementary data to the OAM, the model training supplementary data being used by the OAM to continue training the terminal's model.

[0060] In one embodiment, the wireless access network device switched by the terminal is a control wireless access network device;

[0061] The determining module is used for:

[0062] In response to the terminal's model task completion status being "model inference task not completed", the terminal source control wireless access network device is determined to be the first wireless access network device.

[0063] In one embodiment, the model data includes model inference result data;

[0064] The determining module is further configured to:

[0065] In response to the completion of the model inference task by the terminal source control radio access network device, the model inference result data is determined; the model inference result data is sent to the control radio access network device for terminal switching.

[0066] In one embodiment, the device further includes: a transmitting module;

[0067] The sending module is configured to send a model subscription request to OAM in response to the fact that the wireless access network device is the first wireless access network device. The model subscription request is used to request OAM to update the information of the terminal.

[0068] According to a fourth aspect of the present disclosure, a model data management apparatus is provided for use with an OAM entity, the apparatus comprising:

[0069] A receiving module is used to receive model data transmitted by a first wireless access network device in response to a terminal switching wireless access network devices, wherein the first wireless access network device determines the model task completion status based on the terminal; a training module is used to train the terminal's model based on the model data.

[0070] In one embodiment, the model data includes supplementary model training data;

[0071] The training module is used for:

[0072] Obtain the local model training data of the OAM; train the terminal model based on the local model training data and model training supplementary data.

[0073] In one embodiment, the receiving module is further configured to:

[0074] Receive a model subscription request sent by a first wireless access network device; update the terminal's information based on the model subscription request.

[0075] According to a fifth aspect of the present disclosure, a model data management apparatus is provided, comprising:

[0076] A processor; a memory for storing processor-executable instructions; wherein the processor is configured to: execute the model data management method according to the first aspect or any embodiment of the first aspect, or execute the model data management method according to the second aspect or any embodiment of the second aspect.

[0077] According to a sixth aspect of the present disclosure, a non-transitory computer-readable storage medium is provided, wherein when instructions in the storage medium are executed by a processor of a mobile terminal, the mobile terminal is enabled to execute the model data management method described in the first aspect or any embodiment of the first aspect, or the mobile terminal is enabled to execute the model data management method described in the second aspect or any embodiment of the second aspect.

[0078] According to this disclosure, a mobility management method for wireless artificial intelligence is provided, the method comprising:

[0079] The terminal initiates an analysis subscription request. The gNB-CU, based on its AI processing capabilities and the analysis subscription request information, generates a model subscription request and sends it to the OAM. The OAM, based on the model subscription request, initiates a training supplementary data subscription request to the gNB-CU. Relevant network elements collect and process the data and upload it to the OAM. The OAM uses local training data and training supplementary data to train the model, obtaining a model that meets the model subscription request, and sends the trained model to the gNB-CU. The gNB-CU initiates a model inference data subscription request. Relevant network elements collect and process the data and upload it to the gNB-CU. The gNB-CU uses the model inference data to perform model inference and sends the inference results to the terminal. The terminal adjusts its strategy accordingly based on the inference results and uploads terminal performance feedback data to the gNB-CU. The gNB-CU collects and processes model performance data and terminal performance feedback data and reports it to the OAM. The OAM trains and optimizes the model and sends the updated model to the gNB-CU.

[0080] In scenarios where terminals are highly mobile, the work of training and inferring wireless network models can be divided into the following two scenarios:

[0081] 1) When the terminal switches to a new gNB-DU under the same gNB-CU, the terminal re-initiates the analysis subscription request, the gNB-CU updates the terminal's analysis subscription request information, and determines the current task completion status.

[0082] If the current training task is not completed, the gNB-CU resends the model subscription request to the OAM. The OAM updates and analyzes the subscription request based on the information reported by the gNB-CU. The OAM then re-initiates a supplementary training data subscription request, and relevant network elements collect, process, and upload the data to the OAM. The OAM continues model training using local training data and supplementary training data to obtain a model that satisfies the subscription request, and sends it to the gNB-CU. After the terminal handover is completed, the newly connected gNB-DU is responsible for the relevant data collection and forwarding tasks.

[0083] If the current inference task is not completed, the gNB-CU sends an analytics subscription update request to the OAM, and the OAM updates the analytics subscription request. The gNB-CU continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the gNB-DU currently connected to the terminal according to the access location in the updated analytics subscription request message. This gNB-DU then sends the inference result to the terminal, and the terminal adjusts its strategy accordingly based on the inference result. After the terminal handover is completed, the newly connected gNB-DU is responsible for the relevant data collection and data forwarding tasks.

[0084] 2) When the terminal switches to the new gNB-CU, the terminal resends the analysis subscription request, and the newly connected gNB-CU sends a model subscription request to the OAM. The OAM updates the terminal's analysis subscription request and sends the updated analysis subscription request information to the terminal's source gNB-CU. After updating the analysis subscription request message, the source gNB-CU determines the current task completion status.

[0085] If the current training task is not completed, the source gNB-CU will no longer send supplementary training data to the OAM. The OAM initiates a supplementary training data subscription request to the newly connected gNB-CU of the terminal. The relevant network elements collect and process the data and upload it to the OAM. The OAM continues to train the model using the local training data and the supplementary training data, obtains a model that satisfies the model subscription request, and sends it to the newly connected gNB-CU of the terminal. After the terminal handover is completed, the newly connected gNB-DU and the newly connected gNB-CU of the terminal are responsible for tasks such as data collection, forwarding, model inference, and data feedback.

[0086] If the current inference task is not completed, the source gNB-CU continues to complete the inference task. After obtaining the inference result, it sends the inference result to the newly connected gNB-CU based on the access location in the update analysis request information. The source gNB-CU is no longer responsible for the analysis request-related tasks of this terminal. The newly connected gNB-CU sends the inference result to the newly connected gNB-DU, which then sends the inference result to the terminal. The terminal adjusts its strategy accordingly based on the inference result. After the terminal handover is completed, the newly connected gNB-DU and the newly connected gNB-CU are responsible for related data collection, forwarding, and model inference tasks.

[0087] Alternatively, in some publicly disclosed embodiments, the OAM is responsible for the entire process of data collection, model training, and model inference, while intermediate network elements are only responsible for forwarding data and model inference results. Specifically, the alternative process involves the terminal initiating an analytics subscription request. The gNB-DU and gNB-CU forward this request to the OAM. The OAM collects local data, trains the model, requests inference data, performs model inference, and then sends the inference results to the terminal. When a terminal handover occurs, each network element only needs to report an analytics subscription update request. The OAM then requests model inference data or sends inference results based on the location information in the updated analytics subscription request. Upon receiving the inference results, the terminal makes corresponding adjustments based on these results.

[0088] The technical solutions provided by the embodiments of this disclosure can include the following beneficial effects: by determining the model task completion status of the terminal and the wireless access network device transmitting data, the wireless network architecture supporting AI can have higher stability and efficiency in mobile terminal scenarios, further providing mobile terminals with higher-quality AI analysis services. Furthermore, it provides a method to ensure the continuity of wireless network AI model training in high-speed terminal movement scenarios, solving the problem that wireless network AI cannot perform model training or that training results cannot be effectively delivered in high-speed terminal movement scenarios, and solving the problem of inference result loss caused by terminal switching. This ensures the efficiency and stability of wireless network AI services, improves the terminal service experience, and avoids interruption of AI analysis services required by users during switching, ensuring the continuity and efficiency of AI analysis services for mobile users, while also improving the operating efficiency of the wireless network.

[0089] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0090] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0091] Figure 1 This is a schematic diagram of the system structure of a model data management method according to an exemplary embodiment.

[0092] Figure 2 This is a flowchart illustrating model training and model inference in a model data management method according to an exemplary embodiment.

[0093] Figure 3 This is a schematic diagram of the protocol and interface of a mobility management method for a model data management method, according to an exemplary embodiment.

[0094] Figure 4 This is a flowchart illustrating a model data management method according to an exemplary embodiment.

[0095] Figure 5 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0096] Figure 6 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0097] Figure 7 This is a schematic diagram illustrating the protocol and interface principles of a model data management method in which a terminal switches between the same gNB-CU when the training task is not completed, according to an exemplary embodiment.

[0098] Figure 8 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0099] Figure 9 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0100] Figure 10 This is a schematic diagram illustrating the protocol and interface principles of a model data management method in which a terminal switches between the same gNB-CU when the inference task is not completed, according to an exemplary embodiment.

[0101] Figure 11 This is a flowchart illustrating the delivery of AI tasks when terminals switch under the same gNB-CU in a model data management method according to an exemplary embodiment.

[0102] Figure 12 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0103] Figure 13 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0104] Figure 14 This is a schematic diagram illustrating the protocol and interface principle of a terminal switching across gNB-CU when the training task is not completed in a model data management method according to an exemplary embodiment.

[0105] Figure 15 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0106] Figure 16 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0107] Figure 17 This is a schematic diagram illustrating the protocol and interface principles for terminal switching across gNB-CU when the inference task is not completed in a model data management method according to an exemplary embodiment.

[0108] Figure 18 This is a flowchart illustrating the delivery of AI tasks during mid-terminal handover across gNB-CUs in a model data management method according to an exemplary embodiment.

[0109] Figure 19 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0110] Figure 20 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0111] Figure 21 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0112] Figure 22 This is a flowchart illustrating yet another model data management method according to an exemplary embodiment.

[0113] Figure 23 This is a block diagram of a model data management device according to an exemplary embodiment.

[0114] Figure 24 This is a block diagram of another model data management device according to an exemplary embodiment.

[0115] Figure 25 This is a block diagram illustrating an apparatus for model data management according to an exemplary embodiment.

[0116] Figure 26 This is a block diagram illustrating yet another apparatus for model data management according to an exemplary embodiment. Detailed Implementation

[0117] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this disclosure. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this disclosure as detailed in the appended claims.

[0118] Artificial intelligence, such as machine learning and deep learning, requires massive amounts of data for model training and inference to obtain high-precision network models, providing accurate decision recommendations for terminals. Terminals or next-generation wireless networks can achieve significant performance improvements by relying on AI-driven decision recommendations. To realize a big data-enabled AI wireless network and obtain models that can improve wireless network performance, it is necessary to determine the wireless network AI framework, the functions of AI modules, and the output-output relationships of various network elements.

[0119] At the 3rd Generation Partnership Project (3GPP) Wireless Access Network (RAN) #88 meeting, a research project on RAN-side intelligence optimization was adopted: a study on enhanced data collection for New Radio (NR) and EUTRA-NR Dual Connectivity (ENDC). The RAN3 #110e meeting began discussions on its design guidelines, basic concepts, applicable cases, and standard implications, with a basic functional framework agreed upon as the initial architecture. Figure 1 This is a schematic diagram of the system architecture of a model data management method according to an exemplary embodiment. For example... Figure 1 As shown, based on the discussion, a potential wireless network architecture supporting artificial intelligence includes the following functional units:

[0120] (1) Data collection and preparation: This includes data acquisition and data preprocessing functions. Data acquisition can be performed on multiple network elements. The data provided includes measurement data, feedback performance data, and model performance data.

[0121] (2) Model Training: Iterating machine learning models through computation and processing to obtain better models for inference. Inputs include training data and model performance feedback.

[0122] (3) Model inference: Using a trained artificial intelligence (machine learning / deep learning) model to generate prediction results or decision results.

[0123] (4) Action: Develop and execute strategies using the model inference results, and feed back the relevant performance results after execution to the Data collection.

[0124] Figure 1 The system architecture diagram shown provides a foundation for realizing wireless artificial intelligence. In scenarios where terminals have high-speed mobility, in order to ensure the continuity of model training and inference, and the continuity of AI analysis services obtained by the terminals, mobility management of wireless artificial intelligence is considered. At the same time, the interaction between various network elements with AI functions is further standardized and optimized, so that wireless network artificial intelligence has more robust and efficient performance.

[0125] In related technologies, if a terminal switches to a different radio access network device before obtaining inference results, the terminal will lose the current inference results and re-initiate an analytics subscription request. The OAM and other related network elements will then perform new model training and inference. For example, the terminal initiates an analytics subscription request to the 5G next generation Node B Distributed Unit (gNB-DU). The gNB-DU forwards this request to the 5G next generation Node B Control Unit (gNB-CU), which then reports the request to the OAM. The OAM selects a suitable model to train based on the request and requests supplementary training data. After obtaining the training data, the OAM begins model training. Once the OAM obtains the training model, it sends it to the gNB-CU. The gNB-CU requests inference data and begins model inference. The gNB-CU then sends the obtained inference results to the gNB-DU, which in turn sends the inference results back to the terminal. If the terminal switches during the training or inference phase, the inference result will not be delivered along with the service data during the switch in traditional mobility management due to the latency of model training or inference. As a result, the terminal will not receive the final model inference result. At this time, the terminal will re-initiate the analysis subscription request, and each network element will re-perform the entire process of model training and inference.

[0126] Therefore, the following technical problems exist in the relevant technologies:

[0127] (1) When the terminal switches between training and inference phases, the terminal will lose the previous model inference results, and the resource overhead generated during the first model training and model inference will be wasted.

[0128] (2) When the terminal switches between training and inference phases, the terminal will re-initiate the analysis subscription request. Each network element continues to complete the initial model training and inference process, and also trains and infers the model for the newly initiated analysis subscription request of the terminal. Both training and inference require real-time data transmission. Under the condition of limited wireless communication resources, this scheme will increase the network load.

[0129] (3) When a terminal switches during the training or inference phase, it cannot obtain the model inference results requested from the source base station and re-initiates the analysis subscription request. Each network element then performs model training and inference again. Throughout the process, the total latency for the terminal to obtain the inference results includes the latency from the initial initiation of the analysis subscription request to the terminal switching, and the latency of re-training and inference. These two parts of latency are relatively large, which will cause the inference results to be not fed back in a timely manner, affecting the terminal's service experience.

[0130] (4) When the terminal switches frequently, OAM may need to train the model multiple times for the same analysis subscription request for the same terminal, which will lead to insufficient computing power of OAM and reduce the efficiency of the system.

[0131] Based on this, this disclosure provides a model data management method that enables AI-enabled wireless network architectures to have higher stability and efficiency in mobile terminal scenarios, further providing higher-quality AI analysis services for mobile terminals. The embodiments of this disclosure provide a method to ensure the continuity of wireless network AI model training in high-speed mobile terminal scenarios, solving the problem that wireless network AI cannot perform model training or the training results cannot be effectively delivered in high-speed mobile terminal scenarios, and resolving the problem of inference result loss caused by terminal switching. This ensures the efficiency and stability of wireless network AI services, improves the terminal service experience, and also helps to improve the operating efficiency of the wireless network.

[0132] It is further understood that the wireless communication system of this disclosure is a network providing wireless communication functionality. The wireless communication system can employ different communication technologies, such as code division multiple access (CDMA), wideband code division multiple access (WCDMA), time division multiple access (TDMA), frequency division multiple access (FDMA), orthogonal frequency-division multiple access (OFDMA), single-carrier frequency division multiple access (SC-FDMA), and carrier sense multiple access with collision avoidance. Based on factors such as capacity, speed, and latency, networks can be categorized as 2G networks, 3G networks, 4G networks, or future evolution networks, such as 5G networks. 5G networks can also be referred to as New Radio (NR). For ease of description, this disclosure may sometimes simply refer to the wireless communication network as a network.

[0133] Furthermore, the network device involved in this disclosure can also be referred to as a wireless access network device. This wireless access network device can be: a base station, an evolved Node B (eB) base station, a home base station, an access point (AP) in a Wireless Fidelity (WIFI) system, a wireless relay node, a wireless backhaul node, a transmission point (TP), or a transmission and reception point (TRP), etc. It can also be a gNB in ​​an NR system, or a component or part of a base station. When it is a vehicle-to-everything (V2X) communication system, the network device can also be an in-vehicle device. It should be understood that the specific technologies and device forms used in the embodiments of this disclosure are not limited.

[0134] Furthermore, the terminal involved in this disclosure can also be referred to as a terminal device, user equipment (UE), mobile station (MS), mobile terminal (MT), etc., and is a device that provides voice and / or data connectivity to a user. For example, a terminal can be a handheld device with wireless connectivity, an in-vehicle device, etc. Currently, some examples of terminals include: smartphones (Mobile Phones), pocket personal computers (PPCs), handheld computers, personal digital assistants (PDAs), laptops, tablets, wearable devices, or in-vehicle devices, etc. In addition, when it is a vehicle-to-everything (V2X) communication system, the terminal device can also be an in-vehicle device. It should be understood that the embodiments of this disclosure do not limit the specific technology or specific device form adopted by the terminal.

[0135] In this embodiment of the disclosure, based on Figure 1 The system architecture in this disclosure implements the model data management method provided in this disclosure. For example, Figure 1 As shown, the system includes a terminal, a gNB-DU, a gNB-CU, and an OAM. The terminal connects to the gNB-DU via a wireless channel, multiple gNB-DUs connect to the gNB-CU via F1 interfaces, and the gNB-CUs are connected to each other via Xn interfaces. The OAM is primarily responsible for the model training function unit in the wireless network architecture supporting AI. The gNB-CU is responsible for the model inference function unit, performing model inference. The gNB-DU is primarily responsible for the data collection function unit, collecting real-time inference data and terminal performance feedback data. The terminal is responsible for the action execution function unit, making corresponding strategy adjustments based on the model inference results.

[0136] Figure 2 This is a flowchart illustrating model training and model inference in a model data management method according to an exemplary embodiment. For example... Figure 2 As shown, the general model training and inference process includes the following steps:

[0137] Step S11: The terminal initiates an analysis subscription request.

[0138] In this embodiment of the disclosure, the terminal initiating an analysis subscription request includes the following steps: the terminal sends an analysis subscription request to the currently accessed gNB-DU, the analysis subscription request includes the access location, UE identifier, and analysis request type, and the gNB-DU currently accessed by the terminal sends the analysis subscription request to the gNB-CU.

[0139] In one embodiment, the terminal accesses gNB-DU1, and gNB-DU1 and gNB-DU2 access gNB-CU1. The UE identifier is a 5G globally unique temporary UE identity (GUTI), and the analysis request type is represented by an analysis ID, such as analysis ID 1: location prediction analysis service, analysis ID 2: load prediction analysis service. The access location mainly includes the gNB-CU and gNB-DU information currently accessed by the terminal.

[0140] In step S12, gNB-CU sends a model subscription request to OAM. The model subscription request includes its own AI processing capability information and terminal analysis subscription request information.

[0141] In this embodiment of the disclosure, the AI ​​processing capability information includes the base station server's computing speed and current surplus computing power.

[0142] Step S13: OAM performs initial model selection based on the model subscription request.

[0143] Step S14: OAM collects and processes local training data and training supplementary data.

[0144] In this embodiment of the disclosure, the OAM collects local training data and training supplementary data, including the following steps: OAM initiates a training supplementary data subscription request to gNB-CU; gNB-CU initiates a training supplementary data subscription request to gNB-DU; gNB-DU collects training data and sends training supplementary data to gNB-CU; gNB-CU collects and processes the local training data and the received training data and uploads them to OAM; and OAM collects and processes the local training data and training supplementary data as model training data.

[0145] In step S15, OAM uses the model training data to train the model, obtains a model that satisfies the model subscription request information, and sends the trained model to gNB-CU.

[0146] In step S16, the gNB-CU initiates a model inference data subscription request, and the relevant network elements collect and process the data and upload it to the gNB-CU.

[0147] In this embodiment of the disclosure, the gNB-CU initiates a model inference data subscription request, and the relevant network elements collect data and upload it to the gNB-CU, including the following steps: the gNB-CU initiates a model inference data subscription request to the gNB-DU currently connected to the terminal (optionally, other gNB-DUs connected to the gNB-CU), and the gNB-DU currently connected to the terminal (optionally, other gNB-DUs connected to the gNB-CU) collects model inference data and uploads it to the gNB-CU.

[0148] Step S17: gNB-CU uses model inference data to perform model inference and sends the inference results to the terminal. The terminal adjusts its strategy accordingly based on the inference results and then collects and feeds back performance data.

[0149] In this embodiment of the disclosure, the gNB-CU uses model inference data to perform model inference and sends the inference results to the terminal. The terminal makes corresponding policy adjustments based on the inference results, including the following steps: the gNB-CU uses model inference data to perform model inference and sends the inference results to the gNB-DU accessed by the terminal; the gNB-DU sends the received inference results to the terminal; and the terminal makes corresponding policy adjustments based on the inference results.

[0150] In step S18, gNB-CU collects model performance data and terminal performance feedback data and reports them to OAM. OAM trains and optimizes the model and sends the updated model to gNB-CU.

[0151] In this embodiment of the disclosure, the gNB-CU collects model performance data and terminal performance feedback data and reports them to OAM. OAM trains and optimizes the model and sends the updated model to gNB-CU, including the following steps: gNB-CU compares the inference results with the real data to obtain model performance data; the terminal sends performance feedback data to gNB-DU; gNB-DU sends it to gNB-CU; gNB-CU processes the model performance data and terminal performance feedback data and sends it to OAM; OAM trains and optimizes the model based on the model performance data and performance feedback data and sends the updated model parameters to gNB-CU.

[0152] Among them, model performance data refers to model accuracy, and performance feedback data is the quantification of performance improvement brought by AI analysis services. For example, after the terminal subscribes to a certain analysis and implements corresponding strategy adjustments based on the analysis results, power saving can be achieved, for example, power saving can reach 5%.

[0153] Figure 3 This is a schematic diagram illustrating the protocol and interface principles of a mobility management method within a model data management method, according to an exemplary embodiment. For example... Figure 3 As shown, this mainly involves the terminal, the gNB-DU accessed by the terminal, the gNB-CU accessed by the terminal, and the OAM provided in the embodiments of the present invention. Specifically, as follows:

[0154] 1a. The terminal sends an analytics subscription request signaling to the gNB-DU, indicating that it is initiating an analytics subscription request to the gNB-DU. 1b. The gNB-DU sends an analytics subscription request signaling to the gNB-CU, indicating that it is initiating an analytics subscription request to the gNB-CU. 2. The gNB-CU generates model subscription request information based on its AI processing capabilities and the analytics subscription request information. 3. The gNB-CU sends a model subscription request signaling to the OAM, indicating that it is initiating a model subscription request to the OAM. 4. The OAM performs initial model selection based on the model subscription request information, selecting a model to be trained that meets the analytics subscription request. 5a. The OAM sends a training supplementary data subscription request signaling to the gNB-CU, indicating that it is initiating a training supplementary data subscription request to the gNB-CU. 5b. The gNB-CU sends a training supplementary data subscription request signaling to the gNB-DU, indicating that it is initiating a training supplementary data subscription request to the gNB-DU. 6a. The gNB-DU collects training data. 6b. gNB-DU sends training data to gNB-CU. 6c. gNB-CU collects and processes local training data and training data uploaded by gNB-DU. 6d. gNB-CU sends the processed training data to OAM. 7. OAM collects and processes local training data and uploaded supplementary training data as model training data. 8. OAM uses the model training data to train the model and obtain a model that meets the model subscription request information. 9. OAM sends the model to gNB-CU. 10. gNB-CU sends a model inference data subscription request signaling to gNB-DU, indicating that a model inference data subscription request is being initiated to gNB-DU. 11. gNB-DU collects model inference data. 12. gNB-DU sends model inference data to gNB-CU. 13. gNB-CU uses the model inference data to perform model inference and obtain the model inference result. 14a. gNB-CU sends the model inference result to gNB-DU. 14b. gNB-DU sends the model inference result to the terminal. 15. The terminal adjusts its strategy based on the inference results and collects performance feedback data. 16a. The terminal sends the performance feedback data to gNB-DU. 16b. gNB-DU sends the performance feedback data to gNB-CU. 17. gNB-CU compares the inference results with the actual data to obtain model performance data. 18. gNB-CU processes the model performance data and the terminal performance feedback data. 19. gNB-CU sends the model performance data and the terminal performance feedback data to OAM. 20. OAM uses the model performance data and performance feedback data to train and optimize the model. 21. OAM sends the updated model parameters to gNB-CU.

[0155] Figure 4This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 4 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0156] In step S21, in response to the terminal switching the wireless access network device, the model task completion status of the terminal is determined.

[0157] In this embodiment of the disclosure, as described above, during the training and inference process of the model requested by the terminal, if the terminal moves and switches to a different wireless access device, the terminal re-initiates its analytics subscription request to the new distributed wireless access network device. This distributed wireless access network device reports the terminal's analytics subscription request to the control wireless access network device to which the terminal is connected. The control wireless access network device updates the analytics subscription request based on the terminal's request and determines the current model task completion status of the terminal. Based on the terminal's model task completion status, the first wireless access network device for transmitting model data is determined.

[0158] Here, model data can be model training data, model training supplementary data, model inference data, or data related to the terminal model.

[0159] In step S22, the first wireless access network device for transmitting model data is determined based on the model task completion status.

[0160] In this embodiment of the disclosure, the model task completion status includes model training task not completed and model inference task not completed. After the terminal switches to a different wireless access network device, the wireless access network device determines the first wireless access network device for transmitting model data.

[0161] The model data management method provided in this disclosure can determine the wireless access network device currently transmitting training model data or inference model data based on the model task completion status. This solves the problem that wireless network AI cannot perform model training or cannot effectively deliver training results in high-speed mobile terminal scenarios, and also solves the problem of inference result loss caused by terminal switching, thus ensuring the efficiency and stability of wireless network AI services.

[0162] In some embodiments of this disclosure, the terminal switching of radio access network devices may involve switching distributed radio access network devices without switching control radio access network devices, or it may involve switching both distributed and control radio access network devices. It should be noted that the communication range of the control radio access network device can cover the communication ranges of multiple distributed radio access network devices.

[0163] If the terminal switches the distributed radio access network device without switching the control radio access network device, the first radio access network device for transmitting model data can be determined based on the model task completion status. The following implementation method will be described in conjunction with the accompanying drawings.

[0164] One implementation method, Figure 5 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 5 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0165] In step S31, in response to the terminal's model task completion status being that the model training task is not completed, the distributed radio access network device to which the terminal switches is determined to be the first radio access network device.

[0166] In this embodiment, when the terminal switches between distributed radio access network (DRN) devices without switching between control RDN devices, if the terminal's model task completion status is "model training task incomplete," the terminal re-initiates an analytics subscription request to the switched DRN device. The DRN device reports the terminal's analytics subscription request to the control RDN device to which the terminal is connected. The control RDN device updates the analytics subscription request based on the terminal's request and determines the current model task completion status of the terminal. It then resends the model subscription request to the OAM. This model subscription request includes the control RRN device's own AI processing capability information and the terminal's analytics subscription request.

[0167] OAM updates the terminal's analysis subscription request based on the information reported by gNB-CU. OAM re-initiates the model training supplementary data subscription request, determining that the distributed radio access network device to which the terminal is switching provides model training supplementary data to OAM. That is, it determines that the distributed radio access network device to which the terminal is switching is the first radio access network device.

[0168] Figure 6 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 6 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0169] In step S41, in response to the wireless access network device switching to a terminal, the distributed wireless access network device acquires supplementary data for model training.

[0170] In step S42, supplementary model training data is sent to Operation and Maintenance Management (OAM).

[0171] The supplementary data for model training is used by OAM to continue training the terminal's model.

[0172] In this embodiment, the OAM initiates a model training supplementary data subscription request to the control radio access network device. The control radio access network device also initiates a model training supplementary data subscription request to a newly connected distributed radio access network device. The newly connected distributed radio access network device collects the terminal's training data and sends it to the control radio access network device. The control radio access network device collects and processes its local training data, merges the local training data and the terminal training data, determines the model training supplementary data, and uploads the model training supplementary data to the OAM. The OAM collects and processes its local training data, using the OAM local training data and the model training supplementary data as model training data. The OAM uses the model training data to continue model training, obtaining a model that satisfies the model subscription request, and sends it to the control radio access network device.

[0173] Figure 7 This is a schematic diagram illustrating the protocol and interface principles for switching between terminals under the same gNB-CU when the training task is not completed in a model data management method according to an exemplary embodiment. Figure 7 As shown, this mainly involves the terminal, the terminal's source gNB-DU (gNB-DU1), the newly accessed gNB-DU (gNB-DU3), the accessed gNB-CU, and OAM provided in the embodiments of this disclosure. See below:

[0174] 1a. The terminal sends an analysis subscription request signaling to gNB-DU3, indicating that it is initiating an analysis subscription request to gNB-DU3. 1b. gNB-DU3 sends an analysis subscription request signaling to gNB-CU, indicating that it is initiating an analysis subscription request to gNB-CU. 2. gNB-CU updates the analysis subscription request information and determines that the current training task is not yet complete. 3. gNB-CU generates model subscription request information based on its AI processing capabilities and the analysis subscription request information. 4. gNB-CU sends a model subscription request signaling to OAM, indicating that it is initiating a model subscription request to OAM. 5. OAM updates the analysis subscription request information based on the model subscription request information. 6a. OAM sends a training supplementary data subscription request signaling to gNB-CU, indicating that it is initiating a training supplementary data subscription request to gNB-CU. 6b. gNB-CU sends a training supplementary data subscription request signaling to gNB-DU3, indicating that it is initiating a training supplementary data subscription request to gNB-DU3. 7a. gNB-DU3 collects training data. 7b. gNB-DU3 sends supplementary training data to gNB-CU. 7c. gNB-CU collects and processes local training data and supplementary training data uploaded by gNB-DU3. 7d. gNB-CU sends the processed supplementary training data to OAM. 8. OAM collects and processes local training data and supplementary training data as model training data. 9. OAM uses the model training data to train the model and obtain a model that meets the model subscription request information. 10. OAM sends the model to gNB-CU. 11. gNB-DU3 is responsible for tasks such as collecting and forwarding terminal analysis request-related data.

[0175] Another implementation method, Figure 8 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 8 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0176] In step S51, in response to the terminal's model task completion status being that the model inference task is not completed, the controlling wireless access network device is determined to be the first wireless access network device.

[0177] In this embodiment, when the terminal switches between distributed radio access network devices without switching between control radio access network devices, if the terminal's model task completion status is "model inference task incomplete," the control radio access network device sends an analysis subscription update request to the OAM to update the terminal's analysis subscription request. The OAM updates the analysis subscription request based on the reported information. The control radio access network device continues to complete the inference task and obtains model inference result data, which is then sent to the terminal. That is, the control radio access network device is the first radio access network device, determining that the terminal makes corresponding decision adjustments based on the inference result data.

[0178] Figure 9 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 9 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0179] In step S61, in response to the completion of the model inference task performed by the control radio access network device, the model inference result data is determined.

[0180] In step S62, the model inference result data is sent to the distributed wireless access network device for terminal switching.

[0181] In this embodiment, after the control radio access network device continues to complete the inference task and obtains the model inference result data, the control radio access network device sends the model inference result data to the newly connected distributed radio access network device based on the terminal access location in the update analysis subscription request. The newly connected distributed radio access network device forwards the model inference result data to the terminal, and the terminal makes corresponding policy adjustments based on the model inference result data. After the terminal handover is completed, the newly connected distributed radio access network device is responsible for tasks such as data collection and forwarding.

[0182] Figure 10 This is a schematic diagram illustrating the protocol and interface principles of a model data management method, where the terminal switches between the same gNB-CU when the inference task is not completed, according to an exemplary embodiment. Figure 10 As shown, this mainly involves the terminal, the terminal's source gNB-DU (gNB-DU1), the newly accessed gNB-DU (gNB-DU3), the accessed gNB-CU, and OAM provided in the embodiments of this disclosure. See below:

[0183] 1a. The terminal sends an analytics subscription request signaling to gNB-DU3, indicating that it is initiating an analytics subscription request to gNB-DU3. 1b. gNB-DU3 sends an analytics subscription request signaling to gNB-CU, indicating that it is initiating an analytics subscription request to gNB-CU. 2. gNB-CU updates the terminal's analytics subscription request information and determines that the current inference task is not yet complete. 3. gNB-CU sends an analytics subscription update request signaling to OAM, indicating that it is initiating an analytics subscription update request to OAM. 4. OAM updates the terminal's analytics subscription request information. 5. gNB-CU continues to complete the model inference task and obtains the inference results. 6a. gNB-CU sends the model inference results to gNB-DU3. 6b. gNB-DU3 sends the model inference results to the terminal. 7. gNB-DU3 is responsible for tasks such as collecting and forwarding data related to the terminal's analytics request.

[0184] In some embodiments of this disclosure, Figure 11 This is a flowchart illustrating the AI ​​task delivery process when terminals switch between the same gNB-CU in a model data management method according to an exemplary embodiment. Figure 11 As shown, the terminal re-initiates the analysis subscription request. The gNB-CU updates the analysis subscription request based on the reported information and determines the current task completion status. If the training task is not completed, the gNB-CU resends the model subscription request to the OAM (this subscription request includes its own AI processing capability information and the terminal's analysis subscription request). The OAM updates the analysis subscription request based on the reported information. The OAM re-collects and processes the training data and supplementary training data as model training data. The OAM uses the training data to continue model training, obtaining a model that meets the model subscription request, and sends it to the gNB-CU. After the terminal switchover is completed, the newly connected gNB-DU is responsible for related data collection and data forwarding tasks. If the inference task is not completed, the gNB-CU sends an analysis subscription update request to the OAM. The OAM updates the analysis subscription request based on the reported information. The gNB-CU continues to complete the inference task and obtains the inference result. The gNB-CU sends the inference result to the terminal, and the terminal makes corresponding decisions and adjustments based on the inference result. After the terminal switchover is completed, the newly connected gNB-DU is responsible for related data collection and data forwarding tasks. In this embodiment of the disclosure, the model training supplementary data can also be referred to as training supplementary data, and the model inference result data can also be referred to as inference result.

[0185] In particular, in some embodiments of this disclosure, the terminal re-initiating the analysis subscription request may include the following steps: the terminal initiates an analysis subscription request to the newly accessed gNB-DU and the newly accessed gNB-DU reports the analysis subscription request to the gNB-CU.

[0186] Specifically, in some embodiments of this disclosure, the process of OAM re-collecting and processing training data and training supplementary data as model training data may include the following steps: OAM initiates a training supplementary data subscription request to gNB-CU; gNB-CU initiates a training supplementary data subscription request to the newly connected gNB-DU of the terminal; the newly connected gNB-DU of the terminal collects training data and sends training supplementary data to gNB-CU; gNB-CU collects and processes local training data and received training data and uploads them to OAM; and OAM collects and processes local training data and training supplementary data as model training data.

[0187] Specifically, in some embodiments of this disclosure, the gNB-CU sends the inference results to the terminal, and the terminal makes corresponding decision adjustments based on the inference results, which may include the following steps: the gNB-CU sends the inference results to the gNB-DU that the terminal has newly accessed based on the terminal access location in the update analysis subscription request, and the gNB-DU that the terminal has newly accessed forwards the inference results to the terminal, and the terminal makes corresponding policy adjustments based on the inference results.

[0188] In some embodiments of this disclosure, if the terminal switches between distributed radio access network devices and switching control radio access network devices, the first radio access network device for transmitting model data can be determined based on the model task completion status. The following embodiments will be described in conjunction with the accompanying drawings.

[0189] In this embodiment, when a terminal switches between a distributed radio access network (DRB) device and a control RDB device, the terminal re-initiates an analytics subscription request to the newly connected DRB device. The newly connected DRB device then sends the analytics subscription request to the newly connected control RDB device. The newly connected control RDB device sends a model subscription request to the OAM device, which includes its own AI processing capability information and the analytics subscription request. The OAM device updates the analytics subscription request based on the model subscription request and initiates an analytics subscription update request to the source control RDB device. The source control RDB device updates the analytics subscription request and determines the current task completion status.

[0190] One implementation method, Figure 12 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 12 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0191] In step S71, in response to the terminal's model task completion status being that the model training task is not completed, the control radio access network device for terminal switching is determined to be the first radio access network device.

[0192] In this embodiment of the disclosure, when the terminal switches between the distributed radio access network device and the control radio access network device, if the terminal's current model task completion status is that the model training task is not completed, the source control radio access network device will no longer send supplementary model training data to the OAM. In other words, the source control radio access network device will no longer send supplementary model training data to the OAM and will no longer be responsible for the terminal's analysis subscription requests. That is, the control radio access network device switched by the terminal is the first radio access network device.

[0193] Figure 13 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 13 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0194] In step S81, in response to the wireless access network device, the system controls the wireless access network device to obtain supplementary data for model training.

[0195] In step S82, supplementary model training data is sent to OAM.

[0196] The supplementary data for model training is used by OAM to continue training the terminal's model.

[0197] In this embodiment, the OAM initiates a model training supplementary data subscription request to the newly connected control radio access network device, and the newly connected control radio access network device initiates a model training supplementary data subscription request to the newly connected distributed radio access network device. The newly connected distributed radio access network device collects training data and sends the model training supplementary data to the newly connected control radio access network device. The newly connected control radio access network device collects and processes the local training data and the received training data and uploads them to the OAM.

[0198] OAM continues model training using local training data and supplementary training data to obtain a model that satisfies the model subscription request, and sends it to the newly connected control radio access network device. After the terminal handover is completed, the newly connected distributed radio access network device and the newly connected control radio access network device are responsible for tasks such as data collection, forwarding, model inference, and data feedback.

[0199] Figure 14 This is a schematic diagram illustrating the protocol and interface principles for terminal switching across gNB-CUs when the training task is not completed in a model data management method according to an exemplary embodiment. Figure 14As shown, this mainly involves the terminal provided in the embodiments of the present invention, the terminal's source gNB-DU (gNB-DU1), the newly accessed gNB-DU (gNB-DU3), the terminal's source gNB-CU (gNB-CU1), the newly accessed gNB-CU (gNB-CU2), and OAM. Specifically, as follows:

[0200] 1a. The terminal sends an analysis subscription request signaling to gNB-DU3, indicating that it is initiating an analysis subscription request to gNB-DU3. 1b. gNB-DU3 sends an analysis subscription request signaling to gNB-CU2, indicating that it is initiating an analysis subscription request to gNB-CU2. 2. gNB-CU2 generates model subscription request information based on its AI processing capabilities and the analysis subscription request information. 3. gNB-CU2 sends a model subscription request signaling to OAM, indicating that it is initiating a model subscription request to OAM. 4. OAM updates the analysis subscription request information based on the model subscription request information. 5. OAM sends an analysis subscription update request signaling to gNB-CU1, indicating that it is initiating an analysis subscription update request to gNB-CU1. 6. gNB-CU1 updates the analysis subscription request information and determines that the current training task is not completed. 7. It stops uploading supplementary training data and is no longer responsible for tasks related to the analysis subscription request of this terminal. 8a. OAM sends a training supplementary data subscription request signaling to gNB-CU2, indicating that it is initiating a training supplementary data subscription request to gNB-CU2. 8b. gNB-CU2 sends a training supplementary data subscription request signaling to gNB-DU3, indicating that it is initiating a training supplementary data subscription request to gNB-DU3. 9a. gNB-DU3 collects training data. 9b. gNB-DU3 sends the training data to gNB-CU2. 9c. gNB-CU2 collects and processes local training data and training data uploaded by gNB-DU3. 9d. gNB-CU2 sends the processed training data to OAM. 10. OAM collects and processes local training data and training supplementary data as model training data. 11. OAM uses the model training data to continue model training and obtain a model that meets the model subscription request information. 12. OAM sends the model to gNB-CU2. 13. After the handover is complete, gNB-DU3 is responsible for tasks such as collecting and forwarding data related to the terminal analysis request. 14. After the switch is completed, gNB-CU2 will be responsible for tasks such as terminal analysis request-related model inference, data collection, processing and feedback.

[0201] Another implementation method, Figure 15 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 15 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0202] In step S91, in response to the terminal's model task completion status being that the model inference task is not completed, the terminal source control radio access network device is determined to be the first radio access network device.

[0203] In this embodiment of the disclosure, when the terminal switches between distributed radio access network devices and switching control radio access network devices, if the current model task completion status of the terminal is that the model inference task is not completed, it is determined that the source control radio access network device continues to complete the inference task. After obtaining the model inference result data, the model inference result data is sent to the control radio access network device newly accessed by the terminal according to the access location in the update analysis request information. That is, the terminal source control radio access network device is determined to be the first radio access network device.

[0204] Figure 16 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 16 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0205] In step S101, in response to the completion of the model inference task by the terminal source controlling the wireless access network device, the model inference result data is determined.

[0206] In step S102, the model inference result data is sent to the control wireless access network device for terminal switching.

[0207] In this embodiment of the disclosure, the source control radio access network device continues to complete the inference task and obtains the inference result. The source control radio access network device sends the inference result to the control radio access network device to which the terminal has newly accessed, based on the access location in the update analysis subscription request. Afterward, the source control radio access network device is no longer responsible for tasks related to the terminal's analysis request.

[0208] The newly connected control radio access network (SAR) device sends the inference results to the terminal, and the terminal adjusts its strategy accordingly based on the inference results. The newly connected control SAR device then sends the inference results to the newly connected distributed SAR network (DNR) device. The newly connected DNR device sends the inference results to the terminal, and the terminal adjusts its strategy accordingly based on the inference results. After the terminal handover is complete, the newly connected DNR device and the newly connected control SAR device are responsible for tasks such as data collection, forwarding, model inference, and performance feedback.

[0209] Figure 17 This is a schematic diagram illustrating the protocol and interface principles for terminal handover across gNB-CUs when the inference task is not completed in a model data management method according to an exemplary embodiment. Figure 17As shown, this mainly involves the terminal, the source gNB-DU (gNB-DU1) of the terminal, the newly accessed gNB-DU (gNB-DU3) of the terminal, the source gNB-CU (gNB-CU1) of the terminal, the newly accessed gNB-CU (gNB-CU2) of the terminal, and OAM provided in the embodiments of this disclosure. See below:

[0210] 1a. The terminal sends an analysis subscription request signaling to gNB-DU3, indicating that it is initiating an analysis subscription request to gNB-DU3. 1b. gNB-DU3 sends an analysis subscription request signaling to gNB-CU2, indicating that it is initiating an analysis subscription request to gNB-CU2. 2. gNB-CU2 generates model subscription request information based on its AI processing capabilities and the analysis subscription request information. 3. gNB-CU2 sends a model subscription request signaling to OAM, indicating that it is initiating a model subscription request to OAM. 4. OAM updates the analysis subscription request information based on the model subscription request information. 5. OAM sends an analysis subscription update request signaling to gNB-CU1, indicating that it is initiating an analysis subscription update request to gNB-CU1. 6. gNB-CU1 updates the analysis subscription request information and determines that the current inference task is not yet complete. 7. gNB-CU1 continues to complete the inference task and obtains new inference results. 8a. gNB-CU1 sends the model inference results to gNB-CU2. 8b. gNB-CU1 is no longer responsible for tasks related to this terminal analysis request. 8c. gNB-CU2 sends the model inference results to gNB-DU3. 8d. gNB-DU3 sends the model inference results to the terminal. 9. After the handover, gNB-DU3 is responsible for tasks such as collecting and forwarding data related to the terminal analysis request. 10. After the handover, gNB-CU2 is responsible for tasks such as model inference, data collection, and processing related to the terminal analysis request.

[0211] In some embodiments of this disclosure, Figure 18 This is a flowchart illustrating the AI ​​task delivery during mid-terminal handover across gNB-CUs in a model data management method according to an exemplary embodiment. Figure 18As shown, the terminal re-initiates the analysis subscription request. The newly connected gNB-CU sends a model subscription request to the OAM. The OAM updates the analysis subscription request and sends an analysis subscription update request to the source gNB-CU. The source gNB-CU updates the analysis subscription request information and determines the current task completion status. If the training task is not completed, the source gNB-CU no longer sends supplementary training data to the OAM and is no longer responsible for the terminal's analysis subscription request. The OAM re-collects local training data and supplementary training data as model training data. The OAM uses the model training data to continue model training, obtains a model that meets the model subscription request, and sends it to the newly connected gNB-CU. After the terminal switchover is completed, the newly connected gNB-DU and the newly connected gNB-CU are responsible for related data collection, forwarding, model inference, performance feedback, and other tasks. If the inference task is not completed, the source gNB-CU continues to complete the inference task and obtains the inference result. The source gNB-CU sends the inference result to the newly connected gNB-CU on the terminal. The source gNB-CU is no longer responsible for the terminal's analysis of subscription request-related tasks. The newly connected gNB-CU on the terminal sends the inference result to the terminal. The terminal makes corresponding policy adjustments based on the inference result. After the terminal switch is completed, the newly connected gNB-DU and the newly connected gNB-CU on the terminal are responsible for related data collection, forwarding, model inference, performance feedback and other tasks.

[0212] In particular, in some embodiments of this disclosure, the terminal re-initiating the analysis subscription request may include the following steps: the terminal initiates an analysis subscription request to the newly accessed gNB-DU and the newly accessed gNB-DU reports the analysis subscription request to the gNB-CU.

[0213] Specifically, in some embodiments of this disclosure, the source gNB-CU no longer sending training supplementary data to OAM and no longer being responsible for the terminal's analysis subscription requests may include the following steps: OAM initiates a training supplementary data subscription request to the newly connected gNB-CU of the terminal; the newly connected gNB-CU of the terminal initiates a training supplementary data subscription request to the newly connected gNB-DU of the terminal; the newly connected gNB-DU of the terminal collects training data and sends training supplementary data to the newly connected gNB-CU of the terminal; the newly connected gNB-CU of the terminal collects and processes local training data and received training data and uploads it to OAM; and OAM collects and processes local training data and training supplementary data as model training data.

[0214] Specifically, in some embodiments of this disclosure, the newly connected gNB-CU sends the inference result to the terminal, and the terminal makes corresponding policy adjustments based on the inference result, which may include the following steps: the newly connected gNB-CU sends the inference result to the newly connected gNB-DU and the newly connected gNB-DU sends the inference result to the terminal, and the terminal makes corresponding policy adjustments based on the inference result.

[0215] Figure 19 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 19 As shown, the model data management method is used in wireless access network devices and includes the following steps.

[0216] In step S111, in response to the radio access network device being the first radio access network device, a model subscription request is sent to OAM.

[0217] The model subscription request is used to request OAM to update the terminal's information.

[0218] In this embodiment, if the terminal switches to a new radio access network (RAN) device, it resends its analytics subscription request to the new RAN device. The first RAN device transmitting model data resends the model subscription request to the OAM. If the first RAN device is a distributed RAN device, it resends the model subscription request to the OAM via a control RAN device. If the first RAN device is a control RAN device, it resends the model subscription request to the OAM.

[0219] Based on the same / similar concept, embodiments of this disclosure also provide a model data management method.

[0220] Figure 20 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 20 As shown, the model data management method used in OAM includes the following steps.

[0221] In step S121, in response to the terminal switching wireless access network devices, the model data transmitted by the first wireless access network device is received.

[0222] Among them, the first wireless access network device determines the task completion status based on the terminal's model.

[0223] In step S122, the terminal's model is trained based on the model data.

[0224] In this embodiment of the disclosure, if OAM receives model data transmitted by the first radio access network device, it determines that the terminal has switched radio access network devices, and the current model task completion status of the terminal is that the model training task is not completed. Based on the received model data, OAM continues to train the model to obtain the terminal's trained model.

[0225] Figure 21 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 21 As shown, the model data management method used in OAM includes the following steps.

[0226] In step S131, the local model training data of OAM is obtained.

[0227] In step S132, the terminal's model is trained based on the local model training data and the model training supplementary data.

[0228] In this embodiment of the disclosure, OAM collects and processes local OAM training data, and uses the local OAM training data and supplementary model training data as model training data. OAM uses the model training data to continue model training, obtains a model that satisfies the model subscription request, and sends it to the control radio access network device.

[0229] Figure 22 This is a flowchart illustrating a model data management method according to an exemplary embodiment. For example... Figure 22 As shown, the model data management method used in OAM includes the following steps.

[0230] In step S141, a model subscription request sent by the first wireless access network device is received.

[0231] In step S142, the terminal information is updated based on the model subscription request.

[0232] In this embodiment of the disclosure, OAM receives a model subscription request sent by the first radio access network device and updates the terminal's information, including the terminal's access location information after switching radio access network devices.

[0233] Based on the same concept, embodiments of this disclosure also provide a model data management device.

[0234] It is understood that the model data management device provided in this disclosure includes hardware structures and / or software modules corresponding to each function in order to achieve the above-mentioned functions. In conjunction with the units and algorithm steps of the various examples disclosed in this disclosure, this disclosure can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the technical solutions of this disclosure.

[0235] Figure 23 This is a block diagram illustrating a model data management device according to an exemplary embodiment. (Refer to...) Figure 23 The model data management device 100 is applied to wireless access network equipment and includes a determination module 101.

[0236] The determination module 101 is used to determine the model task completion status of the terminal in response to the terminal switching radio access network devices. Based on the model task completion status, the first radio access network device for transmitting model data is determined.

[0237] In this embodiment of the disclosure, the wireless access network device for terminal switching is a distributed wireless access network device.

[0238] The determination module 101 is used to determine the distributed radio access network device that the terminal switches to as the first radio access network device in response to the terminal's model task completion status being that the model training task is not completed.

[0239] In this embodiment of the disclosure, the model data includes supplementary model training data. The apparatus also includes an acquisition module 102.

[0240] The acquisition module 102 is used to acquire model training supplementary data in response to the distributed radio access network device switching for the terminal. The model training supplementary data is sent to the Operation and Maintenance Management (OAM) department, and is used by OAM to continue training the terminal's model.

[0241] In this embodiment of the disclosure, the wireless access network device switched by the terminal is a distributed wireless access network device. The determination module 101 is configured to determine and control the wireless access network device as the first wireless access network device in response to the terminal's model task completion status being that the model inference task is not completed.

[0242] In this embodiment of the disclosure, the model data includes model inference result data. The determining module 101 is further configured to determine the model inference result data in response to the completion of the model inference task performed by the control radio access network device. The model inference result data is then sent to the distributed radio access network device to which the terminal is switched.

[0243] In this embodiment of the disclosure, the radio access network device switched by the terminal is a control radio access network device. The determination module 101 is configured to determine, in response to the terminal's model task completion status being that the model training task is not completed, that the control radio access network device switched by the terminal is the first radio access network device.

[0244] In this embodiment of the disclosure, the model data includes model training supplementary data. The acquisition module 102 is further configured to acquire the model training supplementary data in response to the radio access network device controlling the radio access network device. The model training supplementary data is sent to the OAM (Operational Access Management) system, and the model training supplementary data is used by the OAM to continue training the terminal's model.

[0245] In this embodiment of the disclosure, the wireless access network device switched by the terminal is a control wireless access network device. The determination module 101 is configured to determine, in response to the terminal's model task completion status being that the model inference task is not completed, that the terminal's source control wireless access network device is the first wireless access network device.

[0246] In this embodiment of the disclosure, the model data includes model inference result data. The determining module 101 is further configured to determine the model inference result data in response to the completion of the model inference task performed by the terminal source control radio access network device. The model inference result data is then sent to the control radio access network device for terminal handover.

[0247] In this embodiment of the disclosure, the device further includes a transmitting module 103.

[0248] The sending module 103 is used to send a model subscription request to OAM in response to the radio access network device being the first radio access network device. The model subscription request is used to request OAM to update the terminal's information.

[0249] Figure 24 This is a block diagram illustrating a model data management device according to an exemplary embodiment. (Refer to...) Figure 24 The model data management device 200, applied to OAM, includes a receiving module 201 and a training module 202.

[0250] In this embodiment of the disclosure, the receiving module 201 is configured to receive model data transmitted by the first wireless access network device in response to the terminal switching wireless access network devices, wherein the first wireless access network device determines the model task completion status based on the terminal. The training module 202 is configured to train the terminal's model based on the model data.

[0251] In this embodiment of the disclosure, the model data includes supplementary model training data. The training module 202 is used to acquire local model training data from OAM. Based on the local model training data and the supplementary model training data, the terminal's model is trained.

[0252] In this embodiment of the disclosure, the receiving module 201 is further configured to receive a model subscription request sent by the first wireless access network device. The terminal information is updated based on the model subscription request.

[0253] Regarding the apparatus in the above embodiments, the specific manner in which each module performs its operation has been described in detail in the embodiments related to the method, and will not be elaborated upon here.

[0254] Figure 25 This is a block diagram illustrating an apparatus 300 for model data management according to an exemplary embodiment. For example, apparatus 300 may be a mobile phone, computer, digital broadcasting terminal, messaging device, game console, tablet device, medical device, fitness equipment, personal digital assistant, etc.

[0255] Reference Figure 25 The device 300 may include one or more of the following components: processing component 302, memory 304, power component 306, multimedia component 308, audio component 310, input / output (I / O) interface 312, sensor component 314, and communication component 316.

[0256] Processing component 302 typically controls the overall operation of device 300, such as operations associated with display, telephone calls, data communication, camera operation, and recording. Processing component 302 may include one or more processors 320 to execute instructions to perform all or part of the steps of the methods described above. Furthermore, processing component 302 may include one or more modules to facilitate interaction between processing component 302 and other components. For example, processing component 302 may include a multimedia module to facilitate interaction between multimedia component 308 and processing component 302.

[0257] Memory 304 is configured to store various types of data to support the operation of device 300. Examples of such data include instructions for any application or method operating on device 300, contact data, phonebook data, messages, pictures, videos, etc. Memory 304 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk.

[0258] The power supply component 306 provides power to the various components of the device 300. The power supply component 306 may include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power to the device 300.

[0259] Multimedia component 308 includes a screen that provides an output interface between the device 300 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touchscreen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors may sense not only the boundaries of the touch or swipe action but also the duration and pressure associated with the touch or swipe operation. In some embodiments, multimedia component 308 includes a front-facing camera and / or a rear-facing camera. When the device 300 is in an operating mode, such as a shooting mode or a video mode, the front-facing camera and / or the rear-facing camera may receive external multimedia data. Each front-facing camera and rear-facing camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

[0260] Audio component 310 is configured to output and / or input audio signals. For example, audio component 310 includes a microphone (MIC) configured to receive external audio signals when device 300 is in an operating mode, such as call mode, recording mode, and voice recognition mode. The received audio signals may be further stored in memory 304 or transmitted via communication component 316. In some embodiments, audio component 310 also includes a speaker for outputting audio signals.

[0261] I / O interface 312 provides an interface between processing component 302 and peripheral interface modules, such as keyboards, click wheels, buttons, etc. These buttons may include, but are not limited to, home buttons, volume buttons, power buttons, and lock buttons.

[0262] Sensor assembly 314 includes one or more sensors for providing status assessments of various aspects of device 300. For example, sensor assembly 314 may detect the on / off state of device 300, the relative positioning of components such as the display and keypad of device 300, changes in the position of device 300 or a component of device 300, the presence or absence of user contact with device 300, the orientation or acceleration / deceleration of device 300, and temperature changes of device 300. Sensor assembly 314 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 314 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, sensor assembly 314 may also include an accelerometer, a gyroscope, a magnetometer, a pressure sensor, or a temperature sensor.

[0263] Communication component 316 is configured to facilitate wired or wireless communication between device 300 and other devices. Device 300 can access wireless networks based on communication standards, such as WiFi, 2G, or 3G, or combinations thereof. In one exemplary embodiment, communication component 316 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, communication component 316 also includes a near-field communication (NFC) module to facilitate short-range communication. For example, the NFC module may be implemented based on radio frequency identification (RFID) technology, Infrared Data Association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0264] In an exemplary embodiment, the apparatus 300 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the methods described above.

[0265] In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions is also provided, such as a memory 304 including instructions, which can be executed by a processor 320 of the device 300 to perform the above-described method. For example, the non-transitory computer-readable storage medium may be a ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, and optical data storage device, etc.

[0266] Figure 26 This is a block diagram illustrating an apparatus 400 for model data management according to an exemplary embodiment. For example, apparatus 400 may be provided as a server. (Refer to...) Figure 26 The apparatus 400 includes a processing component 422, which further includes one or more processors, and memory resources represented by memory 432 for storing instructions, such as application programs, that can be executed by the processing component 422. The application programs stored in memory 432 may include one or more modules, each corresponding to a set of instructions. Furthermore, the processing component 422 is configured to execute instructions to perform the methods described above.

[0267] Device 400 may also include a power supply component 426 configured to perform power management of device 400, a wired or wireless network interface 450 configured to connect device 400 to a network, and an input / output (I / O) interface 458. Device 400 may operate on an operating system stored in memory 432, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, or similar.

[0268] It can be further understood that in this disclosure, "multiple" refers to two or more, and other quantifiers are similar. "And / or" describes the relationship between related objects, indicating that three relationships can exist; for example, A and / or B can represent: A alone, A and B simultaneously, and B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. The singular forms "a," "the," and "the" are also intended to include the plural forms unless the context clearly indicates otherwise.

[0269] It is further understood that the terms "first," "second," etc., are used to describe various types of information, but this information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another, and do not indicate a specific order or degree of importance. In fact, the expressions "first," "second," etc., are completely interchangeable. For example, without departing from the scope of this disclosure, first information can also be referred to as second information, and similarly, second information can also be referred to as first information.

[0270] It is further understood that although operations are described in a specific order in the accompanying drawings in this disclosure, this should not be construed as requiring these operations to be performed in the specific order or serial order shown, or requiring all of the shown operations to obtain the desired result. In certain environments, multitasking and parallel processing may be advantageous.

[0271] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and examples are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the following claims.

[0272] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A model data management method, characterized in that, Applied to wireless access network devices, the method includes: In response to a terminal switching a radio access network device, the switched radio access network device and the model task completion status of the terminal are determined. The switched radio access network device includes a new distributed radio access network device under the same control radio access network device, or a new control radio access network device. The model task completion status includes model training task not completed or model inference task not completed. Based on the switched wireless access network device and the model task completion status, the first wireless access network device for transmitting model data is determined, wherein the model data includes model training supplementary data and model inference result data. In the case where the switched radio access network device is a new distributed radio access network device under the same control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new distributed radio access network device under the same control radio access network device, and the model data management is carried out in the following manner: the same control radio access network device resends the model subscription request to the Operation and Maintenance Management (OAM). The OAM updates and analyzes the subscription request based on the information reported by the same control radio access network device and initiates a training supplementary data subscription request. The OAM continues to train the model using local training data and training supplementary data to obtain a model that meets the model subscription request, and sends it to the same control radio access network device. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new distributed radio access network device under the same control radio access network device, and the terminal's model task completion status is that the model inference task is not completed, the first radio access network device is the same control radio access network device, and model data management is performed in the following manner: the same control radio access network device sends an analysis subscription update request to OAM, the OAM updates the analysis subscription request, the same control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new distributed radio access network device according to the access location in the update analysis subscription request message. The new distributed radio access network device sends the inference result to the terminal, and the terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new control radio access network device, and model data management is carried out in the following manner: the source control radio access network device no longer sends training supplementary data to the OAM, the OAM initiates a training supplementary data subscription request to the new control radio access network device, the OAM continues to train the model using local training data and training supplementary data, obtains a model that satisfies the model subscription request, and sends it to the new control radio access network device. After the terminal switch is completed, the new control radio access network device and the new distributed radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, model inference, and data feedback tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is "model inference task incomplete," the first radio access network device is the source control radio access network device, and model data management is performed as follows: the source control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new control radio access network device according to the access location in the update analysis request information. The source control radio access network device is no longer responsible for the terminal's analysis request task. The new control radio access network device sends the inference result to the new distributed radio access network device corresponding to the new control radio access network device. The new distributed radio access network device corresponding to the new control radio access network device sends the inference result to the terminal. The terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new control radio access network device and the new distributed radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, and model inference tasks.

2. The model data management method according to claim 1, characterized in that, The method further includes: the control wireless access network device comparing the inference results with real data to obtain model performance data and terminal performance feedback data; The model performance data and the terminal performance feedback data are reported to OAM, and the updated model parameters sent by OAM are received. The updated model parameters are obtained by OAM through training and optimization of the model based on the model performance data and the terminal performance feedback data.

3. The model data management method according to claim 1, characterized in that, The method further includes: When the switched radio access network device is a new distributed radio access network device under the same control radio access network device, the terminal re-initiates the analysis subscription request, and the same control radio access network device updates the analysis subscription request information of the terminal. When the switched radio access network device is a new control radio access network device, the terminal resends the analysis subscription request, the new control radio access network device sends a model subscription request to the OAM, the OAM updates the terminal's analysis subscription request, and sends the updated analysis subscription request information to the terminal's source control radio access network device.

4. A model data management method, characterized in that, Applied to OAM entities, the method includes: In response to a terminal switching a radio access network device, the terminal receives model data transmitted by the first radio access network device. The model data includes model training supplementary data and model inference result data. The first radio access network device is determined based on the switched radio access network device and the terminal's model task completion status. The switched radio access network device includes a newly distributed radio access network device under the same control radio access network device, or a new control radio access network device. The model task completion status includes either the model training task not being completed or the model inference task not being completed. The model of the terminal is trained based on the model data; In the case where the switched radio access network device is a new distributed radio access network device under the same control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new distributed radio access network device under the same control radio access network device, and the model data management is carried out in the following manner: the same control radio access network device resends the model subscription request to the Operation and Maintenance Management (OAM). The OAM updates and analyzes the subscription request based on the information reported by the same control radio access network device and initiates a training supplementary data subscription request. The OAM continues to train the model using local training data and training supplementary data to obtain a model that meets the model subscription request, and sends it to the same control radio access network device. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new distributed radio access network device under the same control radio access network device, and the terminal's model task completion status is that the model inference task is not completed, the first radio access network device is the same control radio access network device, and model data management is performed in the following manner: the same control radio access network device sends an analysis subscription update request to OAM, the OAM updates the analysis subscription request, the same control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new distributed radio access network device according to the access location in the update analysis subscription request message. The new distributed radio access network device sends the inference result to the terminal, and the terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new control radio access network device, and model data management is carried out in the following manner: the source control radio access network device no longer sends training supplementary data to the OAM, the OAM initiates a training supplementary data subscription request to the new control radio access network device, the OAM continues to train the model using local training data and training supplementary data, obtains a model that satisfies the model subscription request, and sends it to the new control radio access network device. After the terminal switch is completed, the new control radio access network device and the new distributed radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, model inference, and data feedback tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is "model task completed" but "model inference task incomplete", the first radio access network device is the source control radio access network device, and model data management is performed in the following manner: the source control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new control radio access network device according to the access location in the update analysis request information. The source control radio access network device is no longer responsible for the terminal's analysis request task. The new control radio access network device sends the inference result to the new distribution radio access network device corresponding to the new control radio access network device. The new distribution radio access network device corresponding to the new control radio access network device sends the inference result to the terminal. The terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new control radio access network device and the new distribution radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, and model inference tasks.

5. The model data management method according to claim 4, characterized in that, The method further includes: The system receives model performance data and terminal performance feedback data reported by the control radio access network device. The model performance data is obtained by the control radio access network device by comparing the inference results with the actual data. Based on the model performance data and the terminal performance feedback data, the model of the terminal is trained and optimized to generate updated model parameters. The updated model parameters are sent to the control wireless access network device.

6. A model data management device, characterized in that, Applied to wireless access network equipment, the device includes: A determination module is used to respond to a terminal switching a radio access network device, determine the switched radio access network device and the model task completion status of the terminal; and determine the first radio access network device for transmitting model data based on the switched radio access network device and the model task completion status. The switched radio access network device includes a new distributed radio access network device under the same control radio access network device, or a new control radio access network device. The model task completion status includes model training task not completed or model inference task not completed. The model data includes model training supplementary data and model inference result data. In the case where the switched radio access network device is a new distributed radio access network device under the same control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new distributed radio access network device under the same control radio access network device, and the model data management is carried out in the following manner: the same control radio access network device resends the model subscription request to the Operation and Maintenance Management (OAM). The OAM updates and analyzes the subscription request based on the information reported by the same control radio access network device and initiates a training supplementary data subscription request. The OAM continues to train the model using local training data and training supplementary data to obtain a model that meets the model subscription request, and sends it to the same control radio access network device. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new distributed radio access network device under the same control radio access network device, and the terminal's model task completion status is that the model inference task is not completed, the first radio access network device is the same control radio access network device, and model data management is performed in the following manner: the same control radio access network device sends an analysis subscription update request to OAM, the OAM updates the analysis subscription request, the same control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new distributed radio access network device according to the access location in the update analysis subscription request message. The new distributed radio access network device sends the inference result to the terminal, and the terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new control radio access network device, and model data management is carried out in the following manner: the source control radio access network device no longer sends training supplementary data to the OAM, the OAM initiates a training supplementary data subscription request to the new control radio access network device, the OAM continues to train the model using local training data and training supplementary data, obtains a model that satisfies the model subscription request, and sends it to the new control radio access network device. After the terminal switch is completed, the new control radio access network device and the new distributed radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, model inference, and data feedback tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is "model task completed" but "model inference task incomplete", the first radio access network device is the source control radio access network device, and model data management is performed in the following manner: the source control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new control radio access network device according to the access location in the update analysis request information. The source control radio access network device is no longer responsible for the terminal's analysis request task. The new control radio access network device sends the inference result to the new distribution radio access network device corresponding to the new control radio access network device. The new distribution radio access network device corresponding to the new control radio access network device sends the inference result to the terminal. The terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new control radio access network device and the new distribution radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, and model inference tasks.

7. A model data management device, characterized in that, Applied to OAM entities, the device includes: A receiving module is configured to receive model data transmitted by a first wireless access network device in response to a terminal switching wireless access network devices. The model data includes model training supplementary data and model inference result data. The first wireless access network device is determined based on the switched wireless access network device and the model task completion status of the terminal. The switched wireless access network device includes a newly distributed wireless access network device under the same control wireless access network device, or a new control wireless access network device. The model task completion status includes either the model training task not being completed or the model inference task not being completed. A training module is used to train a model of the terminal based on the model data; In the case where the switched radio access network device is a new distributed radio access network device under the same control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new distributed radio access network device under the same control radio access network device, and the model data management is carried out in the following manner: the same control radio access network device resends the model subscription request to the Operation and Maintenance Management (OAM). The OAM updates and analyzes the subscription request based on the information reported by the same control radio access network device and initiates a training supplementary data subscription request. The OAM continues to train the model using local training data and training supplementary data to obtain a model that meets the model subscription request, and sends it to the same control radio access network device. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new distributed radio access network device under the same control radio access network device, and the terminal's model task completion status is that the model inference task is not completed, the first radio access network device is the same control radio access network device, and model data management is performed in the following manner: the same control radio access network device sends an analysis subscription update request to OAM, the OAM updates the analysis subscription request, the same control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new distributed radio access network device according to the access location in the update analysis subscription request message. The new distributed radio access network device sends the inference result to the terminal, and the terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new distributed radio access network device is responsible for data collection and data forwarding tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is that the model training task is not completed, the first radio access network device is the new control radio access network device, and model data management is carried out in the following manner: the source control radio access network device no longer sends training supplementary data to the OAM, the OAM initiates a training supplementary data subscription request to the new control radio access network device, the OAM continues to train the model using local training data and training supplementary data, obtains a model that satisfies the model subscription request, and sends it to the new control radio access network device. After the terminal switch is completed, the new control radio access network device and the new distributed radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, model inference, and data feedback tasks. When the switched radio access network device is a new control radio access network device and the terminal's model task completion status is "model task completed" but "model inference task incomplete", the first radio access network device is the source control radio access network device, and model data management is performed in the following manner: the source control radio access network device continues to complete the inference task, and after obtaining the inference result, it sends the inference result to the new control radio access network device according to the access location in the update analysis request information. The source control radio access network device is no longer responsible for the terminal's analysis request task. The new control radio access network device sends the inference result to the new distribution radio access network device corresponding to the new control radio access network device. The new distribution radio access network device corresponding to the new control radio access network device sends the inference result to the terminal. The terminal adjusts its strategy according to the inference result. After the terminal switch is completed, the new control radio access network device and the new distribution radio access network device corresponding to the new control radio access network device are responsible for data collection, forwarding, and model inference tasks.

8. A communication system, characterized in that, Including terminals, and at least one of the network devices: A wireless access network device that performs the model data management method according to any one of claims 1-3; An OAM entity that performs the model data management method according to any one of claims 4-5.

9. A model data management device, characterized in that, include: processor; Memory used to store processor-executable instructions; The processor is configured to: execute the model data management method according to any one of claims 1-3, or execute the model data management method according to any one of claims 4-5.

10. A non-transitory computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the wireless access network device, the wireless access network device is able to perform the model data management method according to any one of claims 1-3; or, when the instructions in the storage medium are executed by the processor of the OAM entity, the OAM entity is able to perform the model data management method according to any one of claims 4-5.

Images