Method for controlling ai tasks, terminal and base station

Abstract

A control method of an AI task, a terminal and a base station, the base station comprising: a first function module, configured to perform at least one of the following functions by interacting with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation and data management. The application provides a network architecture with intelligent internalization and a corresponding AI interaction process. On the basis of fully exploiting the computing power of the terminal, the application realizes a network with intelligent internalization by taking intelligence as part of the air interface transmission, thereby improving the intelligence of the network. In the embodiment of the application, the terminal collects data, pre-processes the data and reports the data to the base station, thereby reducing the amount of data to be uploaded. In addition, since the uploaded data is not original user data, the security of the terminal data can be improved.

Description

Technical Field

[0001] This invention relates to the field of mobile communication technology, specifically to a control method, terminal, and base station for artificial intelligence (AI) tasks. Background Technology

[0002] 5G and AI are gradually integrating into people's lives and bringing new economic growth trends. With the rapid development of AI technology, the network architecture in the future will increasingly integrate with AI technology. In the upcoming 6G era, AI will inevitably become an indispensable part of the network architecture.

[0003] The integration of existing network architecture with AI is still in its early stages of research. Many application scenarios have not yet been considered in the network architecture. For example, AI enhancement usually adopts an external solution, and no intelligent, intrinsically intelligent network architecture has been proposed. Furthermore, although terminals have certain computing power, their AI capabilities are not considered when designing the network. Additionally, existing AI models are trained on the network side, typically by collecting network or terminal data, using this data to build AI models, and then optimizing the design based on the desired goals. However, these solutions face several challenges: 1) Increasing amounts of data need to be reported, consuming more network bandwidth. 2) Terminals are becoming increasingly intelligent, but this architecture only allows AI computation on the network side, failing to fully utilize terminal computing power. 3) Terminals in communication networks typically receive more multi-dimensional data than the network side, but the security of terminal data is a sensitive factor. Summary of the Invention

[0004] At least one embodiment of the present invention provides a control method, terminal and network device for AI tasks, which can realize an intelligent endogenous network by utilizing the computing power of the terminal and improve the intelligence of the network.

[0005] According to one aspect of the present invention, at least one embodiment provides a base station comprising:

[0006] The first functional module is used to perform at least one of the following functions through interaction with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management.

[0007] According to another aspect of the present invention, at least one embodiment provides a terminal including a second functional module;

[0008] The second functional module is used to perform at least one of the following functions by interacting with the base station: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution.

[0009] According to another aspect of the present invention, at least one embodiment provides a control method for an AI task, applied to a base station, comprising:

[0010] Through the first functional module of the base station, at least one of the following functions is performed by interacting with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management.

[0011] According to another aspect of the present invention, at least one embodiment provides a method for controlling an AI task, applied to a terminal, comprising:

[0012] The terminal interacts with the base station through its second functional module to perform at least one of the following functions: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution.

[0013] According to another aspect of the present invention, at least one embodiment provides a base station, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method described above.

[0014] According to another aspect of the present invention, at least one embodiment provides a terminal, characterized in that it includes: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method described above.

[0015] According to another aspect of the present invention, at least one embodiment provides a computer-readable storage medium on which a program is stored, which, when executed by a processor, implements the steps of the method described above.

[0016] Compared with existing technologies, the AI ​​task control method, terminal, and base station provided in this invention offer an intelligent, intrinsically linked network architecture and a corresponding AI interaction process. By fully leveraging the terminal's computing power and integrating intelligence as part of air interface transmission, it achieves an intelligent, intrinsically linked network, thereby enhancing network intelligence. In this invention, the terminal collects data, preprocesses it, and then reports it to the base station, thus reducing the amount of data that needs to be uploaded. Furthermore, since the uploaded data is not raw user data, the security of terminal data is improved. Attached Figure Description

[0017] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings:

[0018] Figure 1 This is a schematic diagram of an application scenario according to an embodiment of the present invention;

[0019] Figure 2 This is an example diagram of the AI ​​endogenous module in the user plane protocol stack according to an embodiment of the present invention;

[0020] Figure 3 This is an example diagram of the AI ​​endogenous module in the signaling plane protocol stack according to an embodiment of the present invention;

[0021] Figure 4 This is a schematic diagram of the structure of the first functional module in an embodiment of the present invention;

[0022] Figure 5 This is a schematic diagram of the structure of the second functional module in an embodiment of the present invention;

[0023] Figure 6 This is an example diagram of an AI endogenous overall architecture according to an embodiment of the present invention;

[0024] Figure 7 This is a flowchart illustrating the AI ​​task control method of this invention when applied to the base station side;

[0025] Figure 8 This is a flowchart illustrating the AI ​​task control method of this invention when applied to the terminal side.

[0026] Figure 9 This is a schematic diagram of a base station structure provided in an embodiment of the present invention;

[0027] Figure 10 This is a schematic diagram of a terminal provided in an embodiment of the present invention. Detailed Implementation

[0028] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this invention will be thorough and complete, and will fully convey the scope of the invention to those skilled in the art.

[0029] The terms “first,” “second,” etc., used in the specification and claims of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented, for example, in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus. The terms “and / or” in the specification and claims indicate at least one of the connected objects.

[0030] The technologies described in this document are not limited to NR systems and Long Time Evolution (LTE) / LTE-Advanced (LTE-A) systems, but can also be used in various wireless communication systems such as Code Division Multiple Access (CDMA), 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 other systems. The terms "system" and "network" are often used interchangeably. CDMA systems can implement radio technologies such as CDMA2000 and Universal Terrestrial Radio Access (UTRA). UTRA includes Wideband Code Division Multiple Access (WCDMA) and other CDMA variants. TDMA systems can implement radio technologies such as the Global System for Mobile Communication (GSM). OFDMA systems can implement radio technologies such as Ultra Mobile Broadband (UMB), Evolution-UTRA (E-UTRA), IEEE 802.21 (Wi-Fi), IEEE 802.16 (WiMAX), IEEE 802.20, and Flash-OFDM. UTRA and E-UTRA are part of the Universal Mobile Telecommunications System (UMTS). LTE and more advanced LTE (such as LTE-A) are newer versions of UMTS that use E-UTRA. UTRA, E-UTRA, UMTS, LTE, LTE-A, and GSM are described in documents from an organization called the 3rd Generation Partnership Project (3GPP). CDMA2000 and UMB are described in documents from an organization called 3rd Generation Partnership Project 2 (3GPP2).The techniques described herein can be used with the systems and radio technologies mentioned above, as well as with other systems and radio technologies. However, the following description describes NR systems for illustrative purposes, and NR terminology is used in most of the following description, although these techniques can also be applied to applications beyond NR systems.

[0031] The following description provides examples and is not intended to limit the scope, applicability, or configuration set forth in the claims. Changes may be made to the function and arrangement of the elements discussed without departing from the spirit and scope of this disclosure. Various procedures or components may be appropriately omitted, substituted, or added to the examples. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Furthermore, features described with reference to certain examples may be combined in other examples.

[0032] Please see Figure 1 , Figure 1 This diagram illustrates a block diagram of a wireless communication system applicable to an embodiment of the present invention. The wireless communication system includes a terminal 11 and a network device 12. The terminal 11 can also be referred to as a user terminal or user equipment (UE). The terminal 11 can be a mobile phone, tablet computer, laptop computer, personal digital assistant (PDA), mobile internet device (MID), wearable device, or in-vehicle device, etc. It should be noted that the specific type of terminal 11 is not limited in this embodiment of the present invention. Network device 12 can be a base station and / or a core network element. The base station can be a 5G or later version base station (e.g., gNB, 5G NR NB, etc.), or a base station in other communication systems (e.g., eNB, WLAN access point, or other access point, etc.). The base station can be referred to as a node B, evolved node B, access point, base transceiver station (BTS), radio base station, radio transceiver, basic service set (BSS), extended service set (ESS), B node, evolved B node (eNB), home B node, home evolved B node, WLAN access point, WiFi node, or any other suitable term in the field, as long as the same technical effect is achieved. The base station is not limited to specific technical terms. It should be noted that in the embodiments of the present invention, only the base station in the NR system is used as an example, but the specific type of base station is not limited.

[0033] The base station can communicate with terminal 11 under the control of a base station controller, which in various examples may be part of the core network or some base stations. Some base stations may communicate control information or user data with the core network via backhaul. In some examples, some of these base stations may communicate with each other directly or indirectly via backhaul links, which may be wired or wireless communication links. The wireless communication system may support operation on multiple carriers (waveform signals of different frequencies). A multi-carrier transmitter can transmit modulated signals on multiple carriers simultaneously. For example, each communication link may be a multi-carrier signal modulated according to various radio technologies. Each modulated signal may be transmitted on a different carrier and may carry control information (e.g., reference signals, control channels, etc.), overhead information, data, etc.

[0034] The base station can wirelessly communicate with terminal 11 via one or more access point antennas. Each base station can provide communication coverage for its respective coverage area. The coverage area of ​​an access point can be divided into sectors that constitute only a part of that coverage area. The wireless communication system may include different types of base stations (e.g., macro base stations, micro base stations, or pico base stations). Base stations may also utilize different radio technologies, such as cellular or WLAN radio access technologies. Base stations may be associated with the same or different access networks or operator deployments. The coverage areas of different base stations (including coverage areas of the same or different types of base stations, coverage areas using the same or different radio technologies, or coverage areas belonging to the same or different access networks) may overlap.

[0035] Communication links in a wireless communication system may include an uplink for carrying uplink (UL) transmissions (e.g., from terminal 11 to network device 12) or a downlink for carrying downlink (DL) transmissions (e.g., from network device 12 to terminal 11). UL transmissions may also be referred to as reverse link transmissions, and DL transmissions may also be referred to as forward link transmissions. Downlink transmissions may be carried out using licensed frequency bands, unlicensed frequency bands, or both. Similarly, uplink transmissions may be carried out using licensed frequency bands, unlicensed frequency bands, or both.

[0036] As described in the background section, the existing network architecture fails to fully utilize terminal computing power, requires terminals to upload large amounts of data which consumes excessive bandwidth and is detrimental to the security of terminal data. To address at least one of the above problems, this invention provides an AI task control method that can realize an intelligent, intrinsically generated network based on the utilization of terminal computing power, thereby enhancing the intelligence of the network.

[0037] like Figure 2 and Figure 3As shown, in this embodiment of the invention, an AI-endogenous module is added to both the base station (including base station 1 and base station 2) and the terminal. For ease of description, the AI-endogenous module in the base station is referred to as the first functional module, and the AI-endogenous module in the terminal is referred to as the second functional module. Figure 2 An example of an AI-native module in the user plane protocol stack is given. Figure 3 An example of an AI-native module in the signaling plane protocol stack (control plane protocol stack) is given.

[0038] The first functional module in the base station is used to perform at least one of the following functions by interacting with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management.

[0039] The second functional module in the terminal is used to perform at least one of the following functions by interacting with the base station: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution.

[0040] Through the interaction between the AI-intrinsic modules of the base station and the terminal, in this embodiment of the invention, the terminal collects data, preprocesses the data, and then reports it to the base station, thereby reducing the amount of data that needs to be uploaded. Furthermore, since the uploaded data is not raw user data, the security of the terminal data can be improved. This embodiment of the invention can also utilize the terminal's computing power to execute AI tasks and report the execution results, thereby fully utilizing the terminal's existing computing power to achieve an intelligent, intrinsically linked network and enhance network intelligence.

[0041] like Figure 4 As shown, the first functional module on the base station side specifically includes:

[0042] AI task control unit 41 is used to determine one or more of the following configurations according to the task requirements outside or inside the base station and configure them to the terminal and / or local: the first AI model and algorithm to be used, data collection requirements, data preprocessing method, AI task configuration, AI model evaluation index; and to receive the data collection results and / or AI model evaluation results sent by the terminal.

[0043] AI task execution unit 42 is used to generate corresponding execution results based on the data collection results according to the first AI model and algorithm;

[0044] The decision generation unit 43 is used to generate network decisions based on the execution results generated by the AI ​​task execution unit;

[0045] The data management unit 44 is used to maintain the task queue for AI tasks, the AI ​​model and algorithm library, and the collected data sets. The data sets are used to manage the data collected by the terminal in a set-based manner.

[0046] In this embodiment of the invention, the AI ​​task control unit 41 can determine the data acquisition requirements and data preprocessing methods based on external needs or optimization objectives. For example, when external configuration resource allocation optimization is enabled, the first functional module determines the data acquisition requirements, data preprocessing methods, and data maintenance methods corresponding to the resource allocation optimization. At this time, the first functional module can also configure the physical layer to collect the current user's channel quality, the SDAP and PDCP layers to collect the user's data type, and the MAC layer to collect the user's throughput, and configure the data acquisition requirements and data preprocessing methods for the terminal.

[0047] Optionally, the AI ​​task control unit is further configured to form an AI model library for performing the task requirements according to the task requirements, and to determine AI model evaluation indicators, evaluate the AI ​​models in the AI ​​model library, and determine the first AI model to be used, wherein the evaluation indicators include model convergence time and / or prediction accuracy.

[0048] Optionally, the AI ​​task control unit is further configured to send the data collection request to the corresponding module in the base station, and to receive the data collection results sent by the module in the base station; the decision generation unit is further configured to send the network decision to the corresponding module in the base station.

[0049] Optionally, the AI ​​task control unit is further configured to determine other network elements that need to participate in the task requirement calculation, schedule and control the other network elements to participate in the training task of the first AI model, and synchronize the first AI model and transfer model parameters between the base station and the other network elements.

[0050] In this embodiment of the invention, the base station further includes an RRC layer functional module, which is set in... Figure 2 In the RRC layer of the base station shown, at least one of the following operations is performed:

[0051] Send the first task request to the terminal;

[0052] Receive the second task request from the terminal and send it to the first functional module;

[0053] Receive one or more of the following configurations determined by the first functional module and send them to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators;

[0054] Receive at least one of the data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes, and send them to the first functional module;

[0055] The system receives network decisions regarding the second task requirements from the first functional module and sends them to the terminal.

[0056] Optionally, the base station further includes a MAC layer function module, which is set in Figure 2 In the MAC layer of the base station shown, at least one of the following operations is performed:

[0057] Send a third task request to the terminal;

[0058] Receive the fourth task request from the terminal and send it to the first functional module;

[0059] Receive one or more of the following configurations determined by the first functional module and send them to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators;

[0060] Receive at least one of the data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes, and send them to the first functional module;

[0061] The system receives network decisions regarding the second task requirements from the first functional module and sends them to the terminal.

[0062] In addition, the base station also includes other protocol layer modules, which may specifically be functional modules of the PDCP layer, functional modules of the SDAP layer, etc., and these functional modules are used to perform at least one of the following operations:

[0063] Issue the fifth task requirement to the first functional module;

[0064] Receive network decisions based on the requirements of the fifth task from the parameter feedback of the first functional module.

[0065] As can be seen from the above, the interaction between the first functional module on the base station side and other modules includes:

[0066] A) Within the network element: Based on requirements and optimization goals, the first functional module sends network decisions to the corresponding modules; according to data collection needs, it sends data collection instructions to other modules and receives the corresponding data collection results. Here, "network element" refers to the base station, and "within the network element" refers to the area within the base station to which the first functional module belongs.

[0067] b) Between network elements and between network elements and terminals: For AI computing requiring the participation of multiple network elements, the first functional module is responsible for scheduling the training tasks of the participants, controlling the training tasks of the participants, synchronizing the AI ​​models, and transferring AI models and weights between network elements. If data collection requires the cooperation of terminals, it also configures the terminals for data collection, including raw data or preprocessed data collection messages. Here, "between network elements" refers to the interaction between various base stations.

[0068] On the network side, the protocol layers that interact with the first functional module include one or more of the following:

[0069] 1) RRC: RRC issues corresponding task requirements (such as optimization goals or requirements) to the first functional module, such as switching parameter optimization. The first functional module generates corresponding data collection requirements and AI model parameters and task configurations, which are sent to the terminal via RRC signaling. The first functional module then feeds back the decision to RRC.

[0070] 2) MAC: MAC sends task requirements (such as optimization goals or requirements) to the first functional module. The first functional module notifies the MAC of the data collection requirements and / or AI model parameters. The MAC receives data and / or AI model parameters from other network element nodes or terminals and notifies the first functional module. The MAC layer is responsible for the interaction between the AI ​​model and / or data collection and the terminal.

[0071] 3) Other protocol layers of the base station: Other protocol layers of the base station send corresponding task requirements (such as optimization goals or requirements) to the first functional module. The first functional module generates corresponding data collection requirements, and combines the data collection results to obtain optimization results and feed them back to the corresponding protocol layer.

[0072] By introducing an endogenous intelligent architecture, this embodiment of the invention incorporates intelligence as part of the base station. Through integration with the wireless architecture, it achieves the organic fusion of AI model algorithms, interaction processes, and wireless communication, thereby enabling the base station to self-optimize, self-evolve, and self-generate.

[0073] like Figure 5 As shown, the second functional module on the terminal side specifically includes:

[0074] The receiving unit 51 is configured to receive one or more of the following configurations sent by the base station: a first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0075] The task execution unit 52 is used to execute AI tasks according to the first AI model and algorithm configured by the base station, generate inference results, evaluate according to the AI ​​model evaluation indicators configured by the base station, obtain AI model evaluation results and / or AI model parameters, and send them to the base station.

[0076] The model synchronization unit 53 is used to synchronize the AI ​​model between the terminal and the base station;

[0077] The data reporting unit 54 is used to collect data according to data collection requirements and data preprocessing methods, generate collected data and / or tags, and report them to the base station.

[0078] In this embodiment of the invention, the terminal further includes an RRC layer functional module, which is set in... Figure 2 In the RRC layer of the terminal shown, at least one of the following operations is performed:

[0079] RRC layer functional modules are used to perform at least one of the following operations:

[0080] The first task request from the base station is sent to the second functional module.

[0081] The second task requirement generated by the second functional module itself is sent to the base station through the RRC layer;

[0082] The system receives one or more of the following configurations sent by the base station and sends them to the configuration receiving unit: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0083] Send at least one of the collected data, tags, AI model evaluation results, and AI model parameters to the base station.

[0084] Optionally, the terminal also includes a MAC layer function module, which is set in Figure 2 In the MAC layer of the terminal shown, at least one of the following operations is performed:

[0085] The third task request from the base station is received and sent to the second functional module;

[0086] The fourth task requirement generated by the second functional module itself is sent to the base station through the MAC layer;

[0087] The system receives one or more of the following configurations sent by the base station and sends them to the configuration receiving unit: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0088] Send at least one of the collected data, tags, AI model evaluation results, and AI model parameters to the base station.

[0089] In addition, the terminal may also include other access protocol layer modules, which may specifically be functional modules of the PDCP layer, functional modules of the RLC layer, etc., and these functional modules are used to perform at least one of the following operations:

[0090] Issue the sixth task requirement to the second functional module;

[0091] Collect data according to the data collection requirements generated by the second functional module and feed it back to the second functional module.

[0092] As can be seen from the above, on the terminal side, the protocol layer that interacts with the second functional module includes one or more of the following:

[0093] 1) RRC: a. The terminal's RRC sends the corresponding task requirements (such as optimization goals or requirements) to the second functional module, or the second functional module generates the corresponding task requirements and feeds them back to the first functional module on the base station side through the RRC layer. b. Data collection and reporting to the terminal are performed according to the configuration. The second functional module receives the data collection configuration carried by the network-side RRC signaling, processes the collected data, generates collected data and tags according to the data collection requirements, and sends them to the access network through the terminal's RRC. c. The initial configuration of the AI ​​model from the first functional module on the base station side is received through the RRC layer.

[0094] 2) MAC: The terminal MAC receives data collection requests and / or AI model parameter notifications from the network-side MAC and other nodes and transmits them to the second functional module. The terminal MAC layer is responsible for interacting with the network-side AI model, collecting and gathering terminal data, and reporting it through MAC data packets.

[0095] 3) Other access protocol layers: Collect data according to the corresponding data collection requirements generated by the terminal's second functional module and feed it back to the endogenous AI module.

[0096] Please refer to Figure 6 This invention provides an example of an endogenous AI architecture. In this embodiment, end-to-end AI task control includes: the network acting as the AI ​​task manager, responsible for end-to-end AI task management, data acquisition and preprocessing configuration, model synchronization, and other functions. Specifically,

[0097] End-to-end AI task management specifically includes: determining the tasks the network needs to execute based on external or internal optimization tasks, such as externally triggered load balancing or internal network resource allocation optimization; selecting the corresponding models and algorithms; and configuring them locally and on the terminal. The network side sends the initial configuration of the AI ​​task to the terminal side. After receiving the configuration information, the terminal establishes the AI ​​task and decomposes the local task to one or more AI training and inference execution entities for computation. The initial configuration and reconfiguration of the AI ​​task, the selection of AI models and algorithms, and the configuration of data acquisition and preprocessing can be sent to the terminal via RRC signaling or MAC CE.

[0098] Data collection and reporting specifically includes: according to the collection requirements of the AI ​​task, the terminal preprocesses and cleans the data according to the specified data processing method. The data is then formed into data packets according to a predefined or configured format and sent to the network. The results of data collection can be transmitted via protocol layer data packets or control unit data packets such as MACCE, RLC control packets, PDCP control packets, and RRC.

[0099] AI function execution specifically includes: the AI's intrinsic algorithms and computational components. Based on the model and algorithm configuration for the AI ​​task and the training dataset collected from data collection, the model is trained to predict or provide corresponding decisions. AI function execution also includes end-to-end synchronous maintenance of model parameters during AI training. For example, in a federated learning scenario, the network needs to calculate global parameters based on the parameters reported by each terminal and then distribute them to each terminal. Model parameter synchronization can be achieved through RRC signaling or MAC CE.

[0100] Decision-making specifically includes determining the final decision based on the results obtained from the AI ​​function execution, such as network configuration, resource allocation results, or physical layer channel estimation. Terminal-related decisions can be communicated to the peer via RRC signaling or MAC CE.

[0101] Data management specifically includes: the unified database used in this embodiment of the invention, which is divided into three parts: an AI task queue, an algorithm set, and a data set. This database provides data storage services for these data. The AI ​​task queue refers to the cached tasks received from external or internal sources if the network cannot execute them immediately, and they are processed according to task priority. The algorithm set is the total collection of available algorithms in the system. The task scheduling and decomposition module determines the model and algorithm used for each task, and the AI ​​function execution unit retrieves the corresponding model and algorithm from the set. Simultaneously, the algorithm set can also store pre-trained models and algorithms for direct inference or distribution to the terminal for inference. The data set stores data already processed by the terminal or network according to the data collection and processing requirements of the AI ​​tasks. This data includes public information such as user throughput and user status, cell status, and specific information required for each task, such as a user's movement trajectory.

[0102] exist Figure 6 Under the architecture shown, one possible process for switching optimization is as follows:

[0103] 1) Upon receiving a handover optimization task request, the access network stores it in the task queue, schedules the task queue, establishes an end-to-end AI task (handover optimization), and sends RRC signaling to the terminal, including configuring the AI ​​model and notifying the UE to collect location information and channel quality information. On the network side, the first functional model executes the AI ​​model.

[0104] 2) After receiving the RRC signaling, the terminal sends the task configuration to the first functional module. The first functional module collects the configuration according to the AI ​​model configuration, executes the AI ​​function, and collects data. It then notifies the access network through RRC signaling and / or MAC data packets.

[0105] 3) The access network receives RRC signaling / MAC data packets, parses the data collection results, saves the collected data in the data management system, and obtains corresponding decisions through AI model training and inference. Simultaneously, the trained AI module is saved to the model library.

[0106] 4) The first functional module verifies whether the current decision is feasible. If feasible, the decision is sent to the RRC signaling configuration terminal.

[0107] 5) Upon receiving the decision, the terminal performs an AI task evaluation. If the evaluation performance is not met, the first functional module triggers a handover optimization request, which is sent to the access network via RRC signaling.

[0108] 6) Repeat steps 1 to 5 above until the decision meets the evaluation criteria or the maximum number of adjustments is reached.

[0109] Please refer to Figure 7The present invention provides a method for controlling an AI task, which, when applied to a base station, includes:

[0110] Step 71: Through the first functional module of the base station, at least one of the following functions is executed by interacting with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management.

[0111] Through the above steps, this embodiment of the invention realizes an intelligent, intrinsically generated network by utilizing terminal computing power, thereby enhancing the intelligence of the network.

[0112] Optionally, in the above method, the base station specifically performs the following steps:

[0113] Based on the task requirements external to or internal to the base station, determine one or more of the following configurations and configure them to the terminal and / or local: the first AI model and algorithm to be used, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; and receive the data collection results and / or AI model evaluation results sent by the terminal.

[0114] Based on the data collection results, the corresponding execution results are generated according to the first AI model and algorithm.

[0115] Network decisions are generated based on the execution results produced by the AI ​​task execution unit;

[0116] Maintain the task queue for AI tasks, the AI ​​model and algorithm library, and the collected data sets.

[0117] Optionally, the above methods also include:

[0118] The first functional module forms an AI model library to perform the task requirements according to the task requirements, and determines AI model evaluation indicators, evaluates the AI ​​models in the AI ​​model library, and determines the first AI model to be used. The evaluation indicators include model convergence time and / or prediction accuracy.

[0119] Preferably, the above method further includes:

[0120] The first functional module sends the data collection request to the corresponding module in the base station, and receives the data collection results sent by the module in the base station; and sends the network decision to the corresponding module in the base station.

[0121] Optionally, the above methods also include:

[0122] The first functional module determines other network elements that need to participate in the calculation of the task requirements, schedules and controls the other network elements to participate in the training task of the first AI model, and performs synchronization of the first AI model and transmission of model parameters between the base station and the other network elements.

[0123] Optionally, the above methods also include:

[0124] The first functional module also interacts with the RRC layer functional module of the base station to perform at least one of the following operations:

[0125] The first task request is sent to the terminal through the RRC layer functional module;

[0126] Receive the second task request from the terminal forwarded by the RRC layer functional module;

[0127] Through the RRC layer functional module, one or more of the following configurations are sent to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators;

[0128] The RRC layer functional module sends network decisions for the second task requirements to the terminal.

[0129] Receive at least one of the following: data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes and forwarded by the RRC layer functional module.

[0130] Optionally, the above methods also include:

[0131] The first functional module also interacts with the MAC layer functional module of the base station to perform at least one of the following operations:

[0132] The third task request is sent to the terminal through the MAC layer functional module.

[0133] Receive the fourth task request from the terminal forwarded by the MAC layer functional module;

[0134] The MAC layer functional module sends one or more of the following configurations to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0135] The MAC layer functional module sends network decisions for the second task requirements to the terminal.

[0136] Receive at least one of the following: data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes and forwarded by the MAC layer functional module.

[0137] Optionally, the above methods also include:

[0138] Other protocol layer modules are used to perform at least one of the following operations:

[0139] Issue the fifth task requirement to the first functional module;

[0140] Receive network decisions based on the requirements of the third task from the parameter feedback of the first functional module.

[0141] Please refer to Figure 8 The present invention provides an AI task control method, which, when applied to a terminal, includes:

[0142] Step 81: Interact with the base station through the second functional module of the terminal to perform at least one of the following functions: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution.

[0143] Through the above steps, the embodiments of the present invention can utilize terminal computing power to realize an intelligent, intrinsically generated network, thereby enhancing the intelligence of the network.

[0144] Optionally, the execution of at least one of the following functions: end-to-end artificial intelligence (AI) task control, data acquisition and reporting, and AI task execution, including:

[0145] Receive one or more of the following configurations sent by the base station: first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators;

[0146] According to the first AI model and algorithm configured by the base station, the AI ​​task is executed to generate inference results; the AI ​​model is evaluated according to the AI ​​model evaluation indicators configured by the base station to obtain the AI ​​model evaluation results and / or AI model parameters and send them to the base station.

[0147] Synchronization of the AI ​​model is performed between the terminal and the base station;

[0148] Data is collected according to data collection requirements and data preprocessing methods, and the collected data and / or tags are generated and reported to the base station.

[0149] Optionally, the method further includes:

[0150] The second functional module also interacts with the terminal's RRC layer functional module to perform at least one of the following operations:

[0151] Receive the first task request from the base station forwarded by the RRC layer functional module;

[0152] The RRC layer functional module sends the second task requirement generated by the second functional module itself to the base station.

[0153] The system receives one or more of the following configurations forwarded by the RRC layer functional module from the base station: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0154] The RRC layer functional module sends at least one of the collected data, tags, AI model evaluation results, and AI model parameters to the base station.

[0155] Optionally, the method further includes:

[0156] The second functional module also interacts with the terminal's MAC layer functional module to perform at least one of the following operations:

[0157] Receive the third task request from the base station forwarded by the MAC layer functional module;

[0158] Through the MAC layer function, the fourth task requirement generated by the second functional module itself is sent to the base station;

[0159] The configuration receiving unit receives one or more of the following configurations forwarded by the MAC layer functional module from the base station and sends them to the configuration receiving unit: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0160] Through the MAC layer function, at least one of the collected data, tags, AI model evaluation results, and AI model parameters is sent to the base station.

[0161] Optionally, the method further includes:

[0162] Other access protocol layer modules are used to perform at least one of the following operations:

[0163] Issue the sixth task requirement to the second functional module;

[0164] Collect data according to the data collection requirements generated by the second functional module and feed it back to the second functional module.

[0165] The various methods of the embodiments of the present invention have been described above. Apparatus for implementing the above methods will now be provided.

[0166] Please refer to Figure 9 This invention provides a base station, comprising: a processor 901, a transceiver 902, a memory 903, and a bus interface, wherein:

[0167] In this embodiment of the invention, the network-side device further includes: a program stored on a memory 903 and executable on a processor 901, wherein the program, when executed by the processor 901, performs the following steps:

[0168] Through the first functional module of the base station, at least one of the following functions is performed by interacting with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management.

[0169] Optionally, when the processor executes the program, it further performs the following steps:

[0170] Based on the task requirements external to or internal to the base station, determine one or more of the following configurations and configure them to the terminal and / or local: the first AI model and algorithm to be used, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; and receive the data collection results and / or AI model evaluation results sent by the terminal.

[0171] Based on the data collection results, the corresponding execution results are generated according to the first AI model and algorithm.

[0172] Network decisions are generated based on the execution results produced by the AI ​​task execution unit;

[0173] Maintain the task queue for AI tasks, the AI ​​model and algorithm library, and the collected data sets.

[0174] Optionally, when the processor executes the program, it further performs the following steps:

[0175] Using the first functional module, an AI model library for executing the task requirements is formed according to the task requirements. AI model evaluation metrics are determined, and the AI ​​models in the AI ​​model library are evaluated to determine the first AI model to be used. The evaluation metrics include model convergence time and / or prediction accuracy.

[0176] Optionally, when the processor executes the program, it further performs the following steps:

[0177] The first functional module sends the data collection request to the corresponding module in the base station and receives the data collection results sent by the module in the base station; and sends the network decision to the corresponding module in the base station.

[0178] Optionally, when the processor executes the program, it further performs the following steps:

[0179] The first functional module determines other network elements that need to participate in the calculation of the task requirements, schedules and controls the other network elements to participate in the training task of the first AI model, and performs synchronization of the first AI model and transmission of model parameters between the base station and the other network elements.

[0180] Optionally, when the processor executes the program, it further performs the following steps:

[0181] Through the first functional module, it interacts with the RRC layer functional module of the base station to perform at least one of the following operations:

[0182] The first task request is sent to the terminal through the RRC layer functional module;

[0183] Receive the second task request from the terminal forwarded by the RRC layer functional module;

[0184] Through the RRC layer functional module, one or more of the following configurations are sent to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators;

[0185] The RRC layer functional module sends network decisions for the second task requirements to the terminal.

[0186] Receive at least one of the following: data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes and forwarded by the RRC layer functional module.

[0187] Optionally, when the processor executes the program, it further performs the following steps:

[0188] The first functional module interacts with the MAC layer functional module of the base station to perform at least one of the following operations:

[0189] The third task request is sent to the terminal through the MAC layer functional module.

[0190] Receive the fourth task request from the terminal forwarded by the MAC layer functional module;

[0191] The MAC layer functional module sends one or more of the following configurations to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0192] The MAC layer functional module sends network decisions for the second task requirements to the terminal.

[0193] Receive at least one of the following: data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes and forwarded by the MAC layer functional module.

[0194] Optionally, when the processor executes the program, it further performs the following steps:

[0195] Perform at least one of the following operations through other protocol layer modules:

[0196] Issue the fifth task requirement to the first functional module;

[0197] Receive network decisions based on the requirements of the third task from the parameter feedback of the first functional module.

[0198] Understandably, in this embodiment of the invention, the computer program executed by the processor 901 can achieve the above-mentioned functions. Figure 7 The various processes of the method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0199] exist Figure 9 In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits together, represented by one or more processors (processor 901) and memory (memory 903). The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 902 can be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium.

[0200] The processor 901 is responsible for managing the bus architecture and general processing, while the memory 903 can store the data used by the processor 901 when performing operations.

[0201] It should be noted that the terminal in this embodiment is the same as the one described above. Figure 7 The device corresponding to the method shown above, and the implementation methods in each embodiment, are all applicable to the embodiments of this terminal, and can achieve the same technical effect. In this device, the transceiver 902 and the memory 903, as well as the transceiver 902 and the processor 901, can be connected for communication via a bus interface. The function of the processor 901 can also be implemented by the transceiver 902, and the function of the transceiver 902 can also be implemented by the processor 901. It should be noted that the device provided by the embodiments of the present invention can implement all the method steps implemented in the above method embodiments and can achieve the same technical effect. Therefore, the parts that are the same as those in the method embodiments and the beneficial effects will not be described in detail here.

[0202] In some embodiments of the present invention, a computer-readable storage medium is also provided, on which a program is stored, which, when executed by a processor, performs the following steps:

[0203] Through the first functional module of the base station, at least one of the following functions is performed by interacting with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management.

[0204] When executed by the processor, this program can implement all the control methods of the AI ​​tasks applied to the base station side described above, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0205] Please refer to Figure 10 A schematic diagram of a terminal provided in an embodiment of the present invention includes: a processor 1001, a transceiver 1002, a memory 1003, a user interface 1004, and a bus interface.

[0206] In this embodiment of the invention, the terminal further includes a program stored on memory 1003 and executable on processor 1001.

[0207] When the processor 1001 executes the program, it performs the following steps:

[0208] The terminal interacts with the base station through its second functional module to perform at least one of the following functions: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution.

[0209] Optionally, when the processor executes the program, it further performs the following steps:

[0210] Receive one or more of the following configurations sent by the base station: first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators;

[0211] According to the first AI model and algorithm configured by the base station, the AI ​​task is executed to generate inference results; the AI ​​model is evaluated according to the AI ​​model evaluation indicators configured by the base station to obtain the AI ​​model evaluation results and / or AI model parameters and send them to the base station.

[0212] Synchronization of the AI ​​model is performed between the terminal and the base station;

[0213] Data is collected according to data collection requirements and data preprocessing methods, and the collected data and / or tags are generated and reported to the base station.

[0214] Optionally, when the processor executes the program, it further performs the following steps:

[0215] Through the second functional module, it interacts with the terminal's RRC layer functional module to perform at least one of the following operations:

[0216] Receive the first task request from the base station forwarded by the RRC layer functional module;

[0217] The RRC layer functional module sends the second task requirement generated by the second functional module itself to the base station.

[0218] The system receives one or more of the following configurations forwarded by the RRC layer functional module from the base station: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0219] The RRC layer functional module sends at least one of the collected data, tags, AI model evaluation results, and AI model parameters to the base station.

[0220] Optionally, when the processor executes the program, it further performs the following steps:

[0221] Through the second functional module, it interacts with the MAC layer functional module of the terminal to perform at least one of the following operations:

[0222] Receive the third task request from the base station forwarded by the MAC layer functional module;

[0223] Through the MAC layer function, the fourth task requirement generated by the second functional module itself is sent to the base station;

[0224] The configuration receiving unit receives one or more of the following configurations forwarded by the MAC layer functional module from the base station and sends them to the configuration receiving unit: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators.

[0225] Through the MAC layer function, at least one of the collected data, tags, AI model evaluation results, and AI model parameters is sent to the base station.

[0226] Optionally, when the processor executes the program, it further performs the following steps:

[0227] Perform at least one of the following operations through other access protocol layer modules:

[0228] Issue the sixth task requirement to the second functional module;

[0229] Collect data according to the data collection requirements generated by the second functional module and feed it back to the second functional module.

[0230] Understandably, in this embodiment of the invention, the computer program executed by the processor 1001 can achieve the above-mentioned functions. Figure 8 The various processes of the method embodiments shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0231] exist Figure 10 In this context, the bus architecture can include any number of interconnected buses and bridges, specifically linking various circuits of one or more processors represented by processor 1001 and memory represented by memory 1003 together. The bus architecture can also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. The bus interface provides an interface. The transceiver 1002 can be multiple elements, including transmitters and receivers, providing a unit for communicating with various other devices over a transmission medium. For different user equipment, the user interface 1004 can also be an interface capable of connecting external or internal devices, including but not limited to keypads, displays, speakers, microphones, joysticks, etc.

[0232] The processor 1001 is responsible for managing the bus architecture and general processing, and the memory 1003 can store the data used by the processor 1001 when performing operations.

[0233] It should be noted that the device in this embodiment is the same as the one described above. Figure 8 The device corresponding to the method shown above is applicable to the embodiments of this device and can achieve the same technical effect. In this device, the transceiver 1002 and the memory 1003, as well as the transceiver 1002 and the processor 1001, can be connected via a bus interface. The functions of the processor 1001 can also be implemented by the transceiver 1002, and vice versa. It should be noted that the device provided in this embodiment can implement all the method steps implemented in the above method embodiments and achieve the same technical effect. Therefore, the parts and beneficial effects that are the same as those in the method embodiments will not be described in detail here.

[0234] In some embodiments of the present invention, a computer-readable storage medium is also provided, on which a program is stored, which, when executed by a processor, performs the following steps:

[0235] The terminal interacts with the base station through its second functional module to perform at least one of the following functions: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution.

[0236] When executed by the processor, this program can implement all the control methods of the AI ​​tasks applied to the terminal side described above, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0237] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software 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 implementations should not be considered beyond the scope of this invention.

[0238] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0239] In the embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0240] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of the embodiments of the present invention, depending on actual needs.

[0241] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0242] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, essentially, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, ROM, RAM, magnetic disks, or optical disks.

[0243] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A base station, characterized in that, include: The first functional module is used to perform at least one of the following functions through interaction with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management. The first functional module includes: The AI ​​task control unit is used to determine one or more of the following configurations and configure them to the terminal and / or local machine according to the task requirements outside or inside the base station: the first AI model and algorithm to be used, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; and to receive the data collection results and / or AI model evaluation results sent by the terminal; wherein, the data collection results are the collected data and / or tags generated by the terminal in accordance with the data collection requirements and data preprocessing method. The AI ​​task execution unit is used to generate corresponding execution results based on the data collection results, according to the first AI model and algorithm. The decision generation unit is used to generate network decisions based on the execution results generated by the AI ​​task execution unit. The data management unit is used to maintain the task queue for AI tasks, the AI ​​model and algorithm library, and the collected data sets; The base station also includes: Configure the RRC layer functional module in the RRC layer of the base station to perform at least one of the following operations: Send the first task request to the terminal; Receive the second task request from the terminal and send it to the first functional module; Receive one or more of the following configurations determined by the first functional module and send them to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; The system receives at least one of the following: data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes, and sends them to the first functional module.

2. The base station as described in claim 1, characterized in that, The AI ​​task control unit is further configured to form an AI model library for executing the task requirements according to the task requirements, and to determine AI model evaluation indicators, evaluate the AI ​​models in the AI ​​model library, and determine the first AI model to be used. The evaluation indicators include model convergence time and / or prediction accuracy.

3. The base station as described in claim 1, characterized in that, The AI ​​task control unit is also used to send the data collection request to the corresponding module in the base station, and to receive the data collection result sent by the module in the base station; The decision generation unit is further configured to send the network decision to the corresponding module in the base station.

4. The base station as described in claim 1, characterized in that, The AI ​​task control unit is also used to determine other network elements that need to participate in the task requirement calculation, schedule and control the other network elements to participate in the training task of the first AI model, and synchronize the first AI model and transfer model parameters between the base station and the other network elements.

5. The base station as described in claim 1, characterized in that, Also includes: MAC layer functional modules are used to perform at least one of the following operations: Send a third task request to the terminal; Receive the fourth task request from the terminal and send it to the first functional module; Receive one or more of the following configurations determined by the first functional module and send them to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; Receive at least one of the data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes, and send them to the first functional module; The system receives network decisions regarding the second task requirements from the first functional module and sends them to the terminal.

6. The base station as described in claim 1, characterized in that, Also includes: Other protocol layer modules are used to perform at least one of the following operations: Issue the fifth task requirement to the first functional module; Receive network decisions based on the requirements of the fifth task from the parameter feedback of the first functional module.

7. A terminal, characterized in that, Includes a second functional module; The second functional module is used to perform at least one of the following functions by interacting with the base station: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution; The second functional module includes: Configure the receiving unit to receive one or more of the following configurations sent by the base station: first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; The task execution unit is used to execute AI tasks and generate inference results according to the first AI model and algorithm configured by the base station; to evaluate according to the AI ​​model evaluation indicators configured by the base station, to obtain AI model evaluation results and / or AI model parameters and send them to the base station. A model synchronization unit is used to synchronize AI models between the terminal and the base station; The data reporting unit is used to collect data according to data collection requirements and data preprocessing methods, generate collected data and / or tags, and report them to the base station. The terminal also includes: Configure the RRC layer function module in the RRC layer of the terminal to perform at least one of the following operations: The first task request from the base station is sent to the second functional module. The second task requirement generated by the second functional module itself is sent to the base station through the RRC layer; The system receives one or more of the following configurations sent by the base station and sends them to the configuration receiving unit: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators. Send at least one of the collected data, tags, AI model evaluation results, and AI model parameters to the base station.

8. The terminal as described in claim 7, characterized in that, Also includes: MAC layer functional modules are used to perform at least one of the following operations: The third task request from the base station is received and sent to the second functional module; The fourth task requirement generated by the second functional module itself is sent to the base station through the MAC layer; The system receives one or more of the following configurations sent by the base station and sends them to the configuration receiving unit: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators. Send at least one of the collected data, tags, AI model evaluation results, and AI model parameters to the base station.

9. The terminal as described in claim 7, characterized in that, Also includes: Other access protocol layer modules are used to perform at least one of the following operations: Issue the sixth task requirement to the second functional module; Collect data according to the data collection requirements generated by the second functional module and feed it back to the second functional module.

10. A control method for an AI task, applied to a base station, characterized in that, include: Through the first functional module of the base station, at least one of the following functions is performed by interacting with the terminal: end-to-end artificial intelligence (AI) task control, AI task execution, network decision generation, and data management. Specifically, the execution of at least one of the following functions: end-to-end AI task control, AI task execution, network decision generation, and data management includes: Based on the task requirements external or internal to the base station, determine one or more of the following configurations and configure them to the terminal and / or locally: the first AI model and algorithm to be used, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; and receive the data collection results and / or AI model evaluation results sent by the terminal; wherein, the data collection results are the collected data and / or tags generated by the terminal in accordance with the data collection requirements and data preprocessing method. Based on the data collection results, the corresponding execution results are generated according to the first AI model and algorithm. Network decisions are generated based on the execution results produced by the AI ​​task execution unit; Maintain the task queue for AI tasks, the AI ​​model and algorithm library, and the collected data sets; The method further includes: The first functional module also interacts with the RRC layer functional module of the base station to perform at least one of the following operations, wherein the RRC layer functional module sets the RRC layer of the base station: The first task request is sent to the terminal through the RRC layer functional module; Receive the second task request from the terminal forwarded by the RRC layer functional module; Through the RRC layer functional module, one or more of the following configurations are sent to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; The network decision for the second task requirement is sent to the terminal through the RRC layer functional module. Receive at least one of the following: data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes and forwarded by the RRC layer functional module.

11. The method as described in claim 10, characterized in that, Also includes: The first functional module forms an AI model library to perform the task requirements according to the task requirements, and determines AI model evaluation indicators, evaluates the AI ​​models in the AI ​​model library, and determines the first AI model to be used. The evaluation indicators include model convergence time and / or prediction accuracy.

12. The method as described in claim 10, characterized in that, Also includes: The first functional module sends the data collection request to the corresponding module in the base station, and receives the data collection results sent by the module in the base station; And, the network decision is sent to the corresponding module in the base station.

13. The method as described in claim 10, characterized in that, Also includes: The first functional module determines other network elements that need to participate in the calculation of the task requirements, schedules and controls the other network elements to participate in the training task of the first AI model, and performs synchronization of the first AI model and transmission of model parameters between the base station and the other network elements.

14. The method as described in claim 10, characterized in that, Also includes: The first functional module also interacts with the MAC layer functional module of the base station to perform at least one of the following operations: The third task request is sent to the terminal through the MAC layer functional module. Receive the fourth task request from the terminal forwarded by the MAC layer functional module; The MAC layer functional module sends one or more of the following configurations to the terminal: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators. The MAC layer functional module sends network decisions for the second task requirements to the terminal. Receive at least one of the following: data, tags, AI model evaluation results, and AI model parameters sent by the terminal or other network element nodes and forwarded by the MAC layer functional module.

15. A method for controlling an AI task, applied to a terminal, characterized in that, include: The terminal interacts with the base station through its second functional module to perform at least one of the following functions: end-to-end artificial intelligence (AI) task control, data collection and reporting, and AI task execution. The execution of at least one of the following functions: end-to-end artificial intelligence (AI) task control, data acquisition and reporting, and AI task execution includes: Receive one or more of the following configurations sent by the base station: first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators; According to the first AI model and algorithm configured by the base station, the AI ​​task is executed to generate inference results; the AI ​​model is evaluated according to the AI ​​model evaluation indicators configured by the base station to obtain the AI ​​model evaluation results and / or AI model parameters and send them to the base station. Synchronization of the AI ​​model is performed between the terminal and the base station; Data is collected according to data collection requirements and data preprocessing methods, and the collected data and / or tags are generated and reported to the base station. The method further includes: The second functional module also interacts with the terminal's RRC layer functional module to perform at least one of the following operations, wherein the RRC layer functional module sets the RRC layer of the terminal: Receive the first task request from the base station forwarded by the RRC layer functional module; The RRC layer functional module sends the second task requirement generated by the second functional module itself to the base station. The system receives one or more of the following configurations forwarded by the RRC layer functional module from the base station: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators. The RRC layer functional module sends at least one of the collected data, tags, AI model evaluation results, and AI model parameters to the base station.

16. The method as described in claim 15, characterized in that, The second functional module includes a configuration receiving unit; the method further includes: The second functional module also interacts with the terminal's MAC layer functional module to perform at least one of the following operations: Receive the third task request from the base station forwarded by the MAC layer functional module; Through the MAC layer function, the fourth task requirement generated by the second functional module itself is sent to the base station; The configuration receiving unit receives one or more of the following configurations forwarded by the MAC layer functional module from the base station and sends them to the configuration receiving unit: the first AI model and algorithm, data collection requirements, data preprocessing method, AI task configuration, and AI model evaluation indicators. Through the MAC layer function, at least one of the collected data, tags, AI model evaluation results, and AI model parameters is sent to the base station.

17. A base station, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, performs the steps of the method as described in any one of claims 10 to 14.

18. A terminal, characterized in that, include: A processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the method as described in any one of claims 15 to 16.

19. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the method as described in any one of claims 10 to 16.

Images