Model monitoring method and device

Abstract

The invention provides a model monitoring method and device, and the method comprises the steps: transmitting first information to a second node, the first information comprising the monitoring information of a first model at a first node side; first monitoring configuration information from the second node is received, the first monitoring configuration information is used for configuring terminal equipment to execute the monitoring operation of the first model, and the first monitoring configuration information is configured by the second node. And the performance of the model / function is ensured.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and in particular to a model monitoring method and apparatus. Background Technology

[0002] Currently, with the increasing diversity of supported services, networks need to support various requirements such as ultra-high speed, ultra-low latency, ultra-high reliability, and massive connectivity. This makes network planning, configuration, and resource scheduling increasingly complex. Relying on human experience or simple algorithms for network planning, self-optimization of network configuration, and resource scheduling suffers from drawbacks such as high time consumption, high cost, and poor adaptability of self-optimization and scheduling algorithms, making it unable to meet these new challenges. Introducing artificial intelligence (AI) and machine learning into mobile networks can significantly improve the efficiency of network planning, configuration, and resource scheduling, achieving network intelligence. AI and machine learning refer to acquiring large amounts of data, using machine learning algorithms to train models and / or make decision inferences on the acquired data, and outputting AI functions or models and / or decision results (such as predicting the amount of business data within a certain future time period).

[0003] However, when a terminal device is moving, the source node may detect that the performance or prediction accuracy of an artificial intelligence function / model is not very good. But when the terminal device moves from the source node to the target node, the target node may still configure and apply the function or model because it lacks awareness of the performance of the function or model. This may result in poor performance of the function or model on the target node side. Therefore, how to achieve the continuity of function or model monitoring and ensure the performance of the function or model during the movement of the terminal device is a technical problem that people in this field are solving. Summary of the Invention

[0004] This application proposes a model monitoring method and apparatus that can ensure the continuity of model / function monitoring and guarantee the performance of the model / function.

[0005] In a first aspect, embodiments of this application provide a model monitoring method. This method can be applied to a first node, which may be an access network device, a component within the access network device (e.g., a processor, chip, circuit, or chip system), or a logic module or software capable of implementing all or part of the functions of the access network device. The method includes: sending first information to a second node, the first information including monitoring information of a first model on the first node side; and receiving first monitoring configuration information from the second node, the first monitoring configuration information being used to configure a terminal device to perform monitoring operations of the first model, the first monitoring configuration information being configured by the second node.

[0006] The first monitoring configuration information is determined based on this first information.

[0007] In the above method, during the process of the terminal device moving from the first node to the second node, by sending the monitoring information of the first model on the first node side to the second node, the second node can determine the historical performance or historical monitoring status of the first model based on the monitoring information of the first model on the first node side. Therefore, the second node can select a function / model (i.e., the first model) for the terminal device based on the historical performance or historical monitoring status of the first model, and configure first monitoring configuration information. This first monitoring configuration information is used by the terminal device to monitor the first model, thereby ensuring the continuity of model monitoring. For example, when the second node determines that the historical performance of the first model is good based on the monitoring information of the first model on the first node side, the first monitoring configuration information determined by the second node can be a relatively sparse model monitoring configuration; when the second node determines that the historical performance of the first model is poor based on the monitoring information of the first model on the first node side, the first monitoring configuration information determined by the second node can be a relatively dense model monitoring configuration. Furthermore, when the terminal device moves from the first node to the second node, the second node can determine the historical performance or historical monitoring status of the first model based on the monitoring information of the first model on the first node side. This avoids the problem that the second node may still configure and apply the first model due to a lack of performance awareness of the first model during the process of the terminal device moving from the first node to the second node, which would lead to poor performance of the first model on the second node side, thus ensuring the performance of the model.

[0008] In one possible implementation, the monitoring information of the first model on the first node side includes at least one of the following: identification information of the cell where the terminal device has stayed and / or identification information of the cell where the terminal device is currently located; performance information of the first model; or second monitoring configuration information, which is configuration information determined by the first node for monitoring the first model.

[0009] Optionally, the monitoring information of the first model on the first node side includes the monitoring information of the first model under one or more cells on the first node side, with one cell corresponding to one monitoring information. For example, the identification information of one of the one or more cells corresponds to one monitoring information.

[0010] In the above method, the monitoring information of the first model under one or more cells on the first node side can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0011] In another possible implementation, the second monitoring configuration information includes at least one of the following: configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node.

[0012] The configuration information configured by the first node for the terminal device includes: configuration information configured by the first node for the terminal device and used by the terminal device to perform the monitoring operation of the first model; the configuration information recommended by the first node to the second node includes: configuration information recommended by the first node to the second node for the second node to configure the terminal device to perform the monitoring operation of the first model.

[0013] In the above method, the second monitoring configuration information includes the monitoring configuration information configured by the first node for the terminal device and / or the monitoring configuration information configured by the second node for the terminal device recommended by the first node to the second node. The second monitoring configuration information can be used as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of monitoring of the first model and the performance of the first model.

[0014] In another possible implementation, the performance information of the first model is reported by the terminal device or determined by the first node.

[0015] In another possible implementation, the performance information of the first model is reported by the terminal device, and the performance information of the first model includes at least one of the following: the predicted value determined based on the first model and the measured value determined by the terminal device; the difference between the predicted value determined based on the first model and the measured value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0016] In the above method, the performance information of the first model can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0017] In another possible implementation, the performance information of the first model is determined by the first node, and the performance information of the first model includes at least one of the following: the measurement result determined by the first node in measuring the performance of the first model; the predicted value determined based on the first model and the measurement value determined by the terminal device; the difference between the predicted value determined based on the first model and the measurement value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0018] In the above method, the performance information of the first model can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0019] In another possible implementation, the first information may further include at least one of the following: the identification information of the first model, or the network-side condition identification information corresponding to the first model.

[0020] In another possible implementation, the method further includes: receiving model information supported by the second node or model information of interest from the second node; and determining to send the first information to the second node based on the model information supported by the second node or model information of interest.

[0021] In the above method, when the first model is not included in the model information supported by the second node or the model information of interest, it is determined not to send the first information to the second node; when the first model is included in the model information supported by the second node or the model information of interest, it is determined to send the first information to the second node. Furthermore, the second node can determine the first monitoring configuration information based on the monitoring information of the first model on the first node side in the first information to ensure the continuity of model monitoring.

[0022] In another possible implementation, the method further includes: determining whether the first model is available on the second node side.

[0023] In the above method, the availability of the first model can be judged, the normal operation of the first model can be guaranteed, and the effectiveness of the application of the first model can be ensured.

[0024] In another possible implementation, determining that the first model is available on the second node side includes: determining that the first node and the second node belong to a first cell group and that the first model is available in the first cell group; or determining that the first node and the second node are within a first geographical area and that the first model is available in the first geographical area.

[0025] In another possible implementation, determining that the first model is available on the second node side includes: determining that the network conditions corresponding to the second node side are consistent with or similar to the network conditions corresponding to the network side condition identifier information corresponding to the first model.

[0026] In another possible implementation, the method further includes: receiving first indication information from the terminal device, the first indication information being used to indicate that the first model is available on the second node side.

[0027] In the above method, the availability of the first model on the second node side further indicates that the first node can send the first information to the second node.

[0028] In another possible implementation, the first information is carried in a channel request (HO request) message; the first monitoring configuration information is carried in a channel acknowledgment (HO ACK) message.

[0029] Secondly, embodiments of this application provide a model monitoring method. This method can be applied to a second node, which can be an access network device, a component within the access network device (e.g., a processor, chip, circuit, or chip system), or a logic module or software capable of implementing all or part of the functions of the access network device. The method includes: receiving first information from a first node, the first information including monitoring information of a first model on the first node side; determining first monitoring configuration information based on the first information, the first monitoring configuration information being used to configure a terminal device to perform the monitoring operation of the first model; and sending the first monitoring configuration information to the first node.

[0030] In the above method, during the process of the terminal device moving from the first node to the second node, by sending the monitoring information of the first model on the first node side to the second node, the second node can determine the historical performance or historical monitoring status of the first model based on the monitoring information of the first model on the first node side. Therefore, the second node can configure the terminal device with the selected function / model (i.e., the first model) and first monitoring configuration information based on the historical performance or historical monitoring status of the first model, thus ensuring the continuity of model monitoring. For example, when the second node determines that the historical performance of the first model is good based on the monitoring information of the first model on the first node side, the first monitoring configuration information determined by the second node can be a relatively sparse model monitoring configuration; when the second node determines that the historical performance of the first model is poor based on the monitoring information of the first model on the first node side, the first monitoring configuration information determined by the second node can be a relatively dense model monitoring configuration. Furthermore, when the terminal device moves from the first node to the second node, the second node can determine the historical performance or historical monitoring status of the first model based on the monitoring information of the first model on the first node side. This avoids the problem that the second node may still configure and apply the first model due to a lack of performance awareness of the first model during the process of the terminal device moving from the first node to the second node, which would lead to poor performance of the first model on the second node side, thus ensuring the performance of the model.

[0031] In one possible implementation, the monitoring information of the first model on the first node side includes at least one of the following: identification information of the cell where the terminal device has stayed and / or identification information of the cell where the terminal device is currently located; performance information of the first model; or second monitoring configuration information, which is configuration information determined by the first node for monitoring the first model.

[0032] Optionally, the monitoring information of the first model on the first node side includes the monitoring information of the first model under one or more cells on the first node side, with one cell corresponding to one monitoring information. For example, the identification information of one of the one or more cells corresponds to one monitoring information.

[0033] In the above method, the monitoring information of the first model under one or more cells on the first node side can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0034] In another possible implementation, the second monitoring configuration information includes at least one of the following: configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node.

[0035] In the above method, the second monitoring configuration information includes the monitoring configuration information configured by the first node for the terminal device and / or the monitoring configuration information configured by the second node for the terminal device recommended by the first node to the second node. The second monitoring configuration information can be used as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of monitoring of the first model and the performance of the first model.

[0036] In another possible implementation, the performance information of the first model is reported by the terminal device or determined by the first node.

[0037] In another possible implementation, the performance information of the first model is reported by the terminal device, and the performance information of the first model includes at least one of the following: the predicted value determined based on the first model and the measured value determined by the terminal device; the difference between the predicted value determined based on the first model and the measured value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0038] In the above method, the performance information of the first model can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0039] In another possible implementation, the performance information of the first model is determined by the first node, and the performance information of the first model includes at least one of the following: the measurement result determined by the first node in measuring the performance of the first model; the predicted value determined based on the first model and the measurement value determined by the terminal device; the difference between the predicted value determined based on the first model and the measurement value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0040] In the above method, the performance information of the first model can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0041] In another possible implementation, the method further includes sending the model information supported by the second node or the model information of interest to the first node.

[0042] In another possible implementation, the first information may further include at least one of the following: the identification information of the first model, or the network-side condition identification information corresponding to the first model.

[0043] Thirdly, embodiments of this application provide a model monitoring method, which can be applied to a terminal-side device. The terminal-side device can be a terminal device, a component in the terminal device (e.g., a processor, chip, circuit, or chip system), or a logic module or software that can implement all or part of the functions of the terminal device. The method includes: determining performance information of a first model, the performance information of the first model including one or more of the following: a predicted value determined based on the first model and a measured value determined by the terminal device, the difference between the predicted value determined by the first model and the measured value determined by the terminal device, system performance information, the accuracy information of the prediction of the first model, or the distribution information of the input data and / or output data corresponding to the first model; and storing the performance information of the first model when the terminal device is in an idle state or an inactive state.

[0044] In the above method, when the terminal device is in an idle or inactive state, the performance information of the first model corresponding to each cell that the terminal device has stayed in will be released, and the performance information of the first model corresponding to each cell that has stayed in will be lost. By storing the performance information of the first model corresponding to each cell that has stayed in the terminal device in the idle state, the loss of information can be avoided.

[0045] In one possible implementation, the method further includes: when the terminal device enters the connected state from the idle state, or enters the connected state from the idle state, sending the performance information of the first model.

[0046] In another possible implementation, the method further includes: receiving first monitoring configuration information from a first node, the first monitoring configuration information being used to configure the terminal device to perform monitoring operations of the first model, the first monitoring configuration information being configured by the second node, the first monitoring configuration information being determined based on first information, the first information including monitoring information of the first model on the first node side.

[0047] In the above method, the terminal device can monitor the first model based on the first monitoring configuration information, thereby ensuring the continuity of the monitoring of the first model.

[0048] In another possible implementation, the monitoring information of the first model on the first node side includes at least one of the following: identification information of the cell where the terminal device has stayed and / or identification information of the cell where the terminal device is currently located; performance information of the first model; or second monitoring configuration information, which is configuration information determined by the first node for monitoring the first model.

[0049] In the above method, the monitoring information of the first model under one or more cells on the first node side can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0050] In another possible implementation, the second monitoring configuration information includes at least one of the following: configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node.

[0051] In the above method, the second monitoring configuration information includes the monitoring configuration information configured by the first node for the terminal device and / or the monitoring configuration information configured by the second node for the terminal device recommended by the first node to the second node. The second monitoring configuration information can be used as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of monitoring of the first model and the performance of the first model.

[0052] In another possible implementation, the first information may further include at least one of the following: the identification information of the first model, or the network-side condition identification information corresponding to the first model.

[0053] In another possible implementation, the method further includes: determining that the first model is available on the second node side; and sending first indication information to the first node, the first indication information being used to indicate that the first model is available on the second node side.

[0054] In the above method, the availability of the first model can be judged, the normal operation of the first model can be guaranteed, and the effectiveness of the application of the first model can be ensured.

[0055] In another possible implementation, determining that the first model is available on the second node side includes: determining that the first node and the second node belong to a first cell group, and that the first model is available in the first cell group; or determining that the first node and the second node are within a first geographic area, and that the first model is available in the first geographic area.

[0056] In another possible implementation, determining that the first model is available on the second node side includes: determining that the network conditions corresponding to the second node side are consistent with or similar to the network conditions corresponding to the network side condition identifier information corresponding to the first model.

[0057] Fourthly, embodiments of this application provide a model monitoring device, which can be a first node. The first node can be an access network device, a component in the access network device (e.g., a processor, chip, circuit, or chip system), or a logic module or software that can implement all or part of the functions of the access network device.

[0058] In one possible implementation, the model monitoring device may include modules, units, or means that correspond one-to-one with the methods / operations / steps / actions described in the first aspect. These modules, units, or means may be hardware circuits, software, or a combination of hardware circuits and software.

[0059] In one possible implementation, the model monitoring device includes a processing unit and a transceiver unit. The transceiver unit is configured to send first information to a second node, the first information including monitoring information of the first model on the first node side. The transceiver unit is also configured to receive first monitoring configuration information from the second node, the first monitoring configuration information being configured by the second node to configure a terminal device to perform the monitoring operation of the first model.

[0060] The first monitoring configuration information is determined based on this first information.

[0061] In one possible implementation, the monitoring information of the first model on the first node side includes at least one of the following: identification information of the cell where the terminal device has stayed and / or identification information of the cell where the terminal device is currently located; performance information of the first model; or second monitoring configuration information, which is configuration information determined by the first node for monitoring the first model.

[0062] In another possible implementation, the second monitoring configuration information includes at least one of the following: configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node.

[0063] In another possible implementation, the performance information of the first model is reported by the terminal device or determined by the first node.

[0064] In another possible implementation, the performance information of the first model is reported by the terminal device, and the performance information of the first model includes at least one of the following: the predicted value determined based on the first model and the measured value determined by the terminal device; the difference between the predicted value determined based on the first model and the measured value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0065] In another possible implementation, the performance information of the first model is determined by the first node, and the performance information of the first model includes at least one of the following: the measurement result determined by the first node in measuring the performance of the first model; the predicted value determined based on the first model and the measurement value determined by the terminal device; the difference between the predicted value determined based on the first model and the measurement value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0066] In another possible implementation, the first information may further include at least one of the following: the identification information of the first model, or the network-side condition identification information corresponding to the first model.

[0067] In another possible implementation, the transceiver unit is further configured to receive model information supported by the second node or model information of interest from the second node; the processing unit is configured to determine, based on the model information supported by the second node or model information of interest, to send the first information to the second node.

[0068] In another possible implementation, the processing unit is also used to determine whether the first model is available on the second node side.

[0069] In another possible implementation, the processing unit is configured to determine that the first node and the second node belong to a first cell group, and that the first model is available in the first cell group; or to determine that the first node and the second node are in a first geographical area, and that the first model is available in the first geographical area.

[0070] In another possible implementation, the processing unit is used to determine that the network conditions corresponding to the second node side and the network conditions corresponding to the network side condition identification information of the first model are consistent or similar.

[0071] In another possible implementation, the transceiver unit is further configured to receive first indication information from the terminal device, the first indication information being used to indicate that the first model is available on the second node side.

[0072] In another possible implementation, the first information is carried in a channel request (HO request) message; the first monitoring configuration information is carried in a channel acknowledgment (HO ACK) message.

[0073] For the technical effects of the fourth aspect or possible implementation, please refer to the introduction of the technical effects of the first aspect or corresponding implementation.

[0074] Fifthly, embodiments of this application provide a model monitoring device, which can be a second node. The second node can be an access network device, a component in the access network device (e.g., a processor, chip, circuit, or chip system), or a logic module or software that can implement all or part of the functions of the access network device.

[0075] In one possible implementation, the model monitoring device may include modules, units, or means that correspond one-to-one with the methods / operations / steps / actions described in the second aspect. These modules, units, or means may be hardware circuits, software, or a combination of hardware circuits and software.

[0076] In one possible implementation, the model monitoring device includes: a processing unit and a transceiver unit. The transceiver unit is configured to receive first information from a first node, the first information including monitoring information of the first model on the first node side; the processing unit is configured to determine first monitoring configuration information based on the first information, the first monitoring configuration information being used to configure a terminal device to perform monitoring operations of the first model; the transceiver unit is further configured to send the first monitoring configuration information to the first node.

[0077] In one possible implementation, the monitoring information of the first model on the first node side includes at least one of the following: identification information of the cell where the terminal device has stayed and / or identification information of the cell where the terminal device is currently located; performance information of the first model; or second monitoring configuration information, which is configuration information determined by the first node for monitoring the first model.

[0078] In another possible implementation, the second monitoring configuration information includes at least one of the following: configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node.

[0079] In another possible implementation, the performance information of the first model is reported by the terminal device or determined by the first node.

[0080] In another possible implementation, the performance information of the first model is reported by the terminal device, and the performance information of the first model includes at least one of the following: the predicted value determined based on the first model and the measured value determined by the terminal device; the difference between the predicted value determined based on the first model and the measured value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0081] In another possible implementation, the performance information of the first model is determined by the first node, and the performance information of the first model includes at least one of the following: the measurement result determined by the first node in measuring the performance of the first model; the predicted value determined based on the first model and the measurement value determined by the terminal device; the difference between the predicted value determined based on the first model and the measurement value determined by the terminal device; system performance information; the accuracy information of the prediction of the first model; or the distribution information of the input data and / or output data corresponding to the first model.

[0082] In another possible implementation, the transceiver unit is also used to send model information supported by the second node or model information of interest to the first node.

[0083] In another possible implementation, the first information may further include at least one of the following: the identification information of the first model, or the network-side condition identification information corresponding to the first model.

[0084] For the technical effects of the fifth aspect or possible implementation, please refer to the introduction of the technical effects of the second aspect or corresponding implementation.

[0085] Sixthly, embodiments of this application provide a model monitoring device, which can be a terminal-side device. The terminal-side device can be a terminal device, a component in the terminal device (e.g., a processor, chip, circuit, or chip system), or a logic module or software that can realize all or part of the functions of the terminal device.

[0086] In one possible implementation, the model monitoring device may include modules, units, or means that correspond one-to-one with the methods / operations / steps / actions described in the third aspect. These modules, units, or means may be hardware circuits, software, or a combination of hardware circuits and software.

[0087] In one possible implementation, the model monitoring device includes: a processing unit and a transceiver unit. The processing unit is configured to determine performance information of a first model, the performance information of which includes one or more of the following: a predicted value determined based on the first model and a measured value determined by the model monitoring device, the difference between the predicted value determined by the first model and the measured value determined by the model monitoring device, system performance information, the accuracy information of the prediction of the first model, or the distribution information of the input data and / or output data corresponding to the first model. The processing unit is configured to store the performance information of the first model when the model monitoring device is in an idle state or an inactive state.

[0088] In one possible implementation, the transceiver unit is also used to send the performance information of the first model when the model monitoring device enters the connected state from the idle state or from the idle state.

[0089] In another possible implementation, the transceiver unit is further configured to receive first monitoring configuration information from the first node, the first monitoring configuration information being configured to configure the model monitoring device to perform monitoring operations of the first model, the first monitoring configuration information being configured by the second node, the first monitoring configuration information being determined based on first information, the first information including monitoring information of the first model on the first node side.

[0090] In another possible implementation, the monitoring information of the first model on the first node side includes at least one of the following: identification information of the cell where the model monitoring device has stayed and / or identification information of the cell where the model monitoring device is currently located; performance information of the first model; or second monitoring configuration information, which is configuration information determined by the first node for monitoring the first model.

[0091] In another possible implementation, the second monitoring configuration information includes at least one of the following: configuration information configured by the first node for the model monitoring device; or configuration information recommended by the first node to the second node.

[0092] In another possible implementation, the first information may further include at least one of the following: the identification information of the first model, or the network-side condition identification information corresponding to the first model.

[0093] In another possible implementation, the processing unit is further configured to determine that the first model is available on the second node side; the transceiver unit is further configured to send first indication information to the first node, the first indication information being used to indicate that the first model is available on the second node side.

[0094] In another possible implementation, the processing unit is configured to determine that the first node and the second node belong to a first cell group, and that the first model is available in the first cell group; or to determine that the first node and the second node are in a first geographical area, and that the first model is available in the first geographical area.

[0095] In another possible implementation, the processing unit is used to determine that the network conditions corresponding to the second node side and the network conditions corresponding to the network side condition identification information of the first model are consistent or similar.

[0096] For the technical effects of the sixth aspect or possible implementation, please refer to the introduction of the technical effects of the third aspect or corresponding implementation.

[0097] In a seventh aspect, embodiments of this application provide a model monitoring device, which includes at least one processor that invokes a computer program or instructions to execute the method described in the first aspect or a possible implementation thereof.

[0098] In one possible implementation, the model monitoring device also includes a memory and a communication interface. Optionally, the memory and processor are integrated together.

[0099] In one possible implementation, the memory is located outside the model monitoring device.

[0100] Eighthly, embodiments of this application provide a model monitoring device, which includes at least one processor that invokes a computer program or instructions to execute the method described in the second aspect or a possible implementation thereof.

[0101] In one possible implementation, the model monitoring device also includes a memory and a communication interface. Optionally, the memory and processor are integrated together.

[0102] In one possible implementation, the memory is located outside the model monitoring device.

[0103] Ninthly, embodiments of this application provide a model monitoring device, which includes at least one processor that invokes a computer program or instructions to execute the method described in the third aspect or a possible implementation thereof.

[0104] In one possible implementation, the model monitoring device also includes a memory and a communication interface. Optionally, the memory and processor are integrated together.

[0105] In one possible implementation, the memory is located outside the model monitoring device.

[0106] In a tenth aspect, embodiments of this application provide a chip device including at least one processor for executing computer programs or instructions to implement any of the above aspects or possible implementations of any of the above aspects.

[0107] In one possible implementation, the input of the chip device corresponds to the receiving operation in any of the above-mentioned aspects or possible implementations, and the output of the chip device corresponds to the transmitting operation in any of the above-mentioned aspects or possible implementations.

[0108] Optionally, the processor is coupled to the memory via an interface.

[0109] Optionally, the chip device may also include a memory in which computer programs or instructions are stored.

[0110] Eleventhly, embodiments of this application provide a computer-readable storage medium storing a computer program or instructions that, when executed on a processor, implement the method described above.

[0111] In a twelfth aspect, embodiments of this application provide a computer program product that includes a computer program or instructions that, when executed on a processor, implement the method described in any of the above aspects.

[0112] In a thirteenth aspect, embodiments of this application provide a communication system, which includes: the apparatus as described in the seventh aspect, the apparatus as described in the eighth aspect, and the apparatus as described in the ninth aspect. Attached Figure Description

[0113] Figure 1A This is a schematic diagram of the architecture of a communication system provided in an embodiment of this application;

[0114] Figure 1B This is a schematic diagram of the architecture of another communication system provided in the embodiments of this application;

[0115] Figure 2 This is an example diagram of an open RAN architecture (CU-DU separation architecture) provided in an embodiment of this application;

[0116] Figure 3 This is an example diagram of an intelligent controller in an open RAN architecture provided in an embodiment of this application;

[0117] Figure 4 This is a schematic diagram of AI-based beam management;

[0118] Figure 5 This is a schematic diagram of CSI feedback enhancement based on AI / ML;

[0119] Figure 6 This is a schematic diagram illustrating how a network configuration terminal device performs functional or model reasoning and monitoring.

[0120] Figure 7 This is a schematic diagram of a model management system based on network triggering and network decision-making.

[0121] Figure 8 This is a schematic diagram of a model management system where network decisions are triggered by terminal devices.

[0122] Figure 9 This is a schematic diagram of a model management system that uses network triggering and terminal device decision-making.

[0123] Figure 10This is a schematic diagram of a model management system where terminal devices trigger and make decisions.

[0124] Figures 11-14 This is a schematic diagram of a model monitoring method provided in an embodiment of this application;

[0125] Figure 15 This is a schematic diagram of a model monitoring device provided in an embodiment of this application;

[0126] Figure 16 This is a schematic diagram of another model monitoring device provided in the embodiments of this application. Detailed Implementation

[0127] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0128] References to "one embodiment" or "some embodiments" as described in this application mean that one or more embodiments of this application include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0129] In the description of this application, unless otherwise stated, " / " means "or". For example, A / B can mean A or B. "And / or" in this document is merely a description of 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. Furthermore, "at least one" means one or more, and "multiple" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or multiple items. For example, at least one of a, b, or c can represent: a, b, c; a and b; a and c; b and c; or a and b and c. Where a, b, and c can be single or multiple.

[0130] It is understood that in this application, "instruction" can include direct instruction, indirect instruction, explicit instruction, and implicit instruction. When describing a certain instruction information to indicate A, it can be understood that the instruction information carries A, directly indicates A, or indirectly indicates A.

[0131] In this application, the information indicated by the instruction information is called the information to be instructed. In specific implementations, there are many ways to instruct the information to be instructed, such as, but not limited to, directly instructing the information to be instructed, such as the information to be instructed itself or its index; indirectly instructing the information to be instructed by instructing other information, where there is a relationship between the other information and the information to be instructed; or instructing only a part of the information to be instructed, while the other parts are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing instruction overhead to some extent.

[0132] The information to be instructed can be sent as a whole or divided into multiple sub-information messages, and the sending period and / or timing of these sub-information messages can be the same or different. This application does not limit the specific sending method. The sending period and / or timing of these sub-information messages can be predefined, for example, according to a protocol, or configured by the transmitting device by sending configuration information to the receiving device.

[0133] It is understood that "send" and "receive" in this application refer to the direction of signal transmission. For example, "send information to XX" can be understood as the destination of the information being XX, which can include direct transmission via the air interface or indirect transmission via the air interface from other units or modules. "Receive information from YY" can be understood as the source of the information being YY, which can include direct reception from YY via the air interface or indirect reception from YY via the air interface from other units or modules. "Send" can also be understood as the "output" of the chip interface, and "receive" can also be understood as the "input" of the chip interface.

[0134] In other words, sending and receiving can occur between devices, such as between network devices and terminal devices, or within a device, such as between components, modules, chips, software modules, or hardware modules within the device via buses, wiring, or interfaces.

[0135] It is understandable that information may undergo necessary processing, such as encoding and modulation, between the source and destination, but the destination can understand the valid information from the source. Similar statements in this application can be interpreted in a similar way and will not be elaborated further.

[0136] The communication method provided in this application can be applied to cellular communication systems related to the 3rd Generation Partnership Project (3GPP), such as 4th generation (4G) communication systems, such as Long Term Evolution (LTE) communication systems, and also to 5th generation (5G) communication systems, such as 5G New Radio (NR) communication systems, or to various future communication systems and future communication networks. The method provided in this application can also be applied to Bluetooth systems, Wireless Fidelity (WiFi) systems, LoRa systems, or vehicle-to-everything (V2X) systems, communication systems supporting the integration of multiple wireless technologies, and device-to-device (D2D) systems. The method provided in this application can also be applied to satellite communication systems, wherein the satellite communication system can be integrated with the above-mentioned communication systems. The wireless communication systems involved in this application also include, but are not limited to: narrowband internet of things (NB-IoT) systems, global system for mobile communications (GSM), enhanced data rate for GSM evolution (EDGE), wideband code division multiple access (WCDMA) systems, code division multiple access 2000 (CDMA2000) systems, or time division-synchronization code division multiple access (TD-SCDMA) systems.

[0137] Please see Figure 1A , Figure 1A This is a schematic diagram of the architecture of a communication system provided in an embodiment of this application, to Figure 1AThe application scenario used in this application is illustrated using the communication system architecture shown. The communication system includes a terminal device 101, an access network device 102, and an access network device 103. Optionally, the communication system may also include a core network device 104. The terminal device and the access network device are connected, and the terminal device can also access the core network through the access network device. The terminal device can be any of the types described below, the access network device can be any of the types described below, and the core network device can be any of the types described below. It should be understood that a communication system to which the methods of the embodiments of this application can be applied may include more or fewer access network devices or terminal devices. The access network device and the terminal device can be hardware, or software functionally divided, or a combination of both. The access network device and the terminal device can communicate through other devices or network elements. It is understood that... Figure 1A This application only illustrates one possible communication system architecture to which this embodiment can be applied. In other possible scenarios, the communication system architecture may also include other devices. It should be noted that the methods described in this application embodiment can be applied to... Figure 1A The communication system shown.

[0138] Please see Figure 1B , Figure 1B This is a schematic diagram of the architecture of another communication system provided in the embodiments of this application, to Figure 1B The application scenario used in this application is illustrated using the communication system architecture shown. This communication system includes terminal equipment, a master node, secondary nodes, and core network equipment. The network side configures a master cell group (MCG) and a secondary cell group (SCG) for the terminal equipment. For example, the master node can also be called a master station (MN), and the secondary node can also be called a secondary station (SN). Taking a 5th generation core network (5GC) as an example, the NR base station acts as the master station (MN), and the LTE base station acts as the secondary station (SN). The core network equipment and the LTE base station communicate via the NG-U interface, the core network equipment and the NR base station communicate via the NG interface, and the NR base station and the LTE base station communicate via the Xn interface. The master node and secondary node can be any of the access network equipment described below. It should be noted that the embodiments of this application are not limited to the above scenario and can also be applied to various future communication systems and future communication networks, such as the 6th generation (6G) communication system, which includes a 6G core network and 6G base stations. Understandable, Figure 1BThis application only illustrates one possible communication system architecture to which this embodiment can be applied. In other possible scenarios, the communication system architecture may also include other devices. It should be noted that the methods described in this application embodiment can be applied to... Figure 1B The communication system shown.

[0139] (1) Terminal equipment, also known as user equipment (UE), mobile station (MS), mobile terminal (MT), etc., is a device that provides voice or data connectivity to users. Terminal equipment may also include built-in artificial intelligence (AI) modules for training machine learning (ML) models and providing wireless-related data analysis and strategy feedback to the terminal equipment.

[0140] Specifically, this includes devices that provide voice to users, devices that provide data connectivity to users, or devices that provide both voice and data connectivity to users. For example, it may include handheld devices with wireless connectivity or processing devices connected to a wireless modem. The terminal device can communicate with the core network via a radio access network (RAN), exchanging voice or data with the RAN, or interacting with the RAN for both voice and data. Currently, terminal devices can include: mobile phones, tablets, laptops, PDAs, mobile internet devices (MIDs), wearable devices (such as smartwatches, smart bracelets, pedometers, etc.), in-vehicle devices (such as cars, bicycles, electric vehicles, airplanes, ships, trains, high-speed trains, etc.), virtual reality (VR) devices, augmented reality (AR) devices, wireless terminals in industrial control, smart home devices (such as refrigerators, televisions, air conditioners, electricity meters, etc.), intelligent robots, workshop equipment, wireless terminals in autonomous driving, wireless terminals in remote surgery, wireless terminals in smart grids, wireless terminals in transportation safety, wireless terminals in smart cities, or wireless terminals in smart homes, and flying equipment (such as intelligent robots, hot air balloons, drones, airplanes), etc. Terminal devices can also be other devices with terminal functions; for example, a terminal device can also be a device that performs terminal functions in D2D communication. Terminal devices can also include vehicle-to-everything (V2X) terminal devices, machine-to-machine / machine-type communications (M2M / MTC) terminal devices, Internet of Things (IoT) terminal devices, light UEs, reduced capability UEs (REDCAP UEs), subscriber units, subscriber stations, mobile stations, remote stations, access points (APs), remote terminals, access terminals, user terminals, user agents, or user devices, and drone equipment, etc.For example, it can include mobile phones (or "cellular" phones), computers with mobile terminal devices, portable, pocket-sized, handheld, and computer-embedded mobile devices. Examples include personal communication service (PCS) phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, and personal digital assistants (PDAs). It also includes limited devices, such as devices with low power consumption, limited storage capacity, or limited computing power. Examples include information sensing devices such as barcode scanners, radio frequency identification (RFID), sensors, global positioning systems (GPS), and laser scanners. In this application, terminal devices with wireless transceiver capabilities and chips that can be installed in the aforementioned terminal devices are collectively referred to as terminal devices.

[0141] It should be noted that the terminal device may be a device or apparatus with a chip, or a device or apparatus with integrated circuitry, or a chip, module or control unit in the device or apparatus shown above. This application does not limit the specific device.

[0142] (2) Access network equipment is a device deployed in a radio access network to provide wireless communication functions for terminal devices. Access network equipment can also be called a radio access network (RAN) entity, access node, network node, or communication device, etc. Access network equipment may also include built-in AI module functions for training ML models and providing wireless-related data analysis and policy feedback for the access network equipment.

[0143] Specifically, the access network equipment can be access network equipment for cellular systems related to the 3GPP (3rd Generation Partnership Project). For example, fourth-generation (4G) mobile communication systems or 5G mobile communication systems. The access network equipment can also be access network equipment in open RAN (O-RAN or ORAN) or cloud radioaccess network (CRAN). Alternatively, the access network equipment can also be access network equipment in a communication system resulting from the integration of two or more of the above communication systems.

[0144] Access network equipment includes, but is not limited to: evolved Node B (eNB), radio network controller (RNC), Node B (NB), base station controller (BSC), base transceiver station (BTS), home base station (e.g., home-evolved Node B, or home Node B, HNB), baseband unit (BBU), access point (AP) in Wi-Fi systems, macro base station, micro base station, wireless relay node, donor node, radio controller in CRAN scenarios, wireless backhaul node, transmission point (TP), or transmission reception point (TRP). Access network equipment can also be access network equipment in 5G mobile communication systems. For example, next-generation Node B (gNB), TRP, TP in New Radio (NR) systems, or one or a group of antenna panels (including multiple antenna panels) of a base station in a 5G mobile communication system. Alternatively, access network equipment can also be network nodes constituting a gNB or transmission point. Examples include a central unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU). An optional central unit can also be called a control unit. CUs and DUs can be separate entities or included in the same network element, such as a BBU. RUs can be included in radio equipment or radio units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs). Alternatively, access network equipment can also be servers, wearable devices, vehicles, or in-vehicle equipment. For example, in V2X technology, access network equipment can be roadside units (RSUs).

[0145] It should be noted that CU (or CU-CP and CU-UP), DU, or RU may have different names in different systems, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called an open central unit (O-CU) or an open CU, DU can also be called an open distributed unit (O-DU), central unit control plane (CU-CP) can also be called an open central unit control plane (O-CU-CP) or an open CU-CP, central unit user plane (CU-UP) can also be called an open central unit user plane (O-CU-UP) or an open CU-UP, and RU can also be called an open radio unit (O-RU). This application does not impose any specific limitations. Any of the units CU, CU-CP, CU-UP, DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software and hardware modules.

[0146] In some deployments, the CU and DU implement some of the functions of the gNB. For example, the CU implements the functions of the radio resource control (RRC) and packet data convergence protocol (PDCP) layers, while the DU implements the functions of the radio link control (RLC), media access control (MAC), and physical (PHY) layers. Since RRC layer information ultimately becomes PHY layer information, or is derived from PHY layer information, in this architecture, higher-layer signaling, such as RRC layer signaling or PDCP layer signaling, can also be considered to be sent by the DU, or by the DU+RU. It is understood that access network equipment can be CU nodes, DU nodes, or devices including both CU and DU nodes. Furthermore, the CU can be classified as an access network device in the RAN (RAN) or as an access network device in the CN (CN), without restriction.

[0147] Please see Figure 2 , Figure 2This is an example diagram of an open RAN architecture (CU-DU separated architecture) provided in an embodiment of this application. The open RAN architecture may include other components besides those shown in the diagram. Network elements in the communication system are connected via interfaces (e.g., NG, Xn) or air interfaces (Uu). These network element nodes, such as core network equipment, access network equipment (RAN nodes), and terminal equipment, may also include one or more AI modules (only one is shown in the diagram for clarity). The access network equipment may be a single RAN node or may include multiple RAN nodes, for example, including CU and DU. The CU and / or DU may also be equipped with one or more AI modules. Optionally, the CU may be further divided into CU-CP and CU-UP. One or more AI modules are configured in the CU-CP and / or CU-UP. The AI ​​modules are used to implement corresponding AI functions. The AI ​​modules deployed in different network elements may be the same or different. The AI ​​module's model can implement different functions depending on the different parameter configurations. An AI module may have one or more models. A model can infer an output, which includes one or more parameters. The learning, training, or inference processes of different models can be deployed on different nodes or devices, or they can be deployed on the same node or device.

[0148] Please see Figure 3 , Figure 3 This is an example diagram of an intelligent controller in an open RAN architecture provided in this application embodiment. The RIC architecture communication system includes a RAN intelligent controller (RIC). The RIC includes near-real-time (near-RT) RICs and non-real-time (non-RT) RICs. The near-real-time RIC is used for model training and inference. For example, it is used to train an AI model and then use that AI model for inference. The near-real-time RIC can obtain network-side and / or terminal-side information from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU) and / or terminal devices. This information can be used as training data or inference data. Optionally, the near-real-time RIC can deliver the inference results to the RAN nodes and / or terminal devices. Optionally, inference results can be exchanged between CUs and DUs, and / or between DUs and RUs. For example, the near-real-time RIC delivers the inference results to the DU, and the DU sends them to the RU.

[0149] The non-real-time RIC is used for model training and inference. For example, it is used to train an AI model and then use that model for inference. The non-real-time RIC can obtain network-side and / or terminal-side information from RAN nodes (e.g., CU, CU-CP, CU-UP, DU, and / or RU (Radio Unit)) and / or terminal devices. This information can be used as training data or inference data, and the inference results can be delivered to RAN nodes and / or terminal devices. Optionally, inference results can be exchanged between CU and DU, and / or between DU and RU. For example, the non-real-time RIC delivers the inference results to the DU, which then forwards them to the RU. The near-real-time RIC and non-real-time RIC can also be set up separately as a network element. Optionally, the near-real-time RIC and non-real-time RIC can also be part of other devices. For example, the near-real-time RIC is set in the RAN node (e.g., in the CU, DU), while the non-real-time RIC is set in operations, administration and maintenance (OAM), cloud servers, core network devices, or other network devices.

[0150] It should be noted that the access network device can be the device or apparatus shown above, or a component (e.g., a chip), module, or unit in the device or apparatus shown above; this application does not limit the specifics.

[0151] (3) Core network equipment is the interface provided by the bearer network to the external network. It is mainly responsible for access control, registration management, service management, mobility management, etc., for terminal devices accessing the network. Core network equipment can be core network equipment in 4G evolved packet core network (EPC), fifth-generation core network (5GC), or next-generation core network. Core network equipment may also include built-in AI module functions for ML model training.

[0152] For example, core network equipment may be access and mobility management function (AMF) network elements, user plane function (UPF) network elements, session management function (SMF) network elements, location management function network elements, or policy control function (PCF) network elements.

[0153] It should be noted that the core network equipment can be the equipment or apparatus shown above, or it can be a component (e.g., a chip), module, or unit in the equipment or apparatus shown above. This application does not limit the specific details.

[0154] It should be noted that this application is mainly applied to AI-based wireless communication scenarios, such as AI / ML-based beam management and beam selection, AI / ML-based positioning, AI / ML-based channel state information (CSI) feedback enhancement, and AI / ML-based terminal device mobility management. In the above different scenarios, any one or more devices among the terminal device, access network device, or core network device can have AI / ML capabilities and be configured with AI / ML models or functions for inference. The AI / ML models or functions can be trained internally by the node or passed to the node by other nodes.

[0155] To better understand the solutions provided in the embodiments of this application, some terms, concepts or processes involved in the embodiments of this application will be introduced below.

[0156] I. Artificial Intelligence

[0157] Current mobile networks, supporting increasingly diverse services, require support for ultra-high speeds, ultra-low latency, ultra-high reliability, and massive connectivity, making network planning, configuration, and resource scheduling increasingly complex. Furthermore, the increasing use of higher frequencies in mobile networks places greater demands on base station energy efficiency. These new requirements, scenarios, and characteristics present unprecedented challenges to mobile network planning, operation, and efficient management. Relying on manual experience or simple algorithms for network planning, self-optimization of network configuration, and resource scheduling suffers from drawbacks such as high time consumption, high costs, and poor adaptability of self-optimization and scheduling algorithms, making them inadequate for addressing these new challenges.

[0158] Introducing artificial intelligence (AI) and machine learning into mobile networks can significantly improve the efficiency of network planning, configuration, and resource scheduling, enabling network intelligence. AI can simulate arbitrary nonlinear models, thus effectively adapting to real-world environments and approaching performance limits. AI and machine learning acquire massive amounts of data, and machine learning algorithms use this data for model training and / or decision reasoning, outputting AI models and / or decision results (such as predicting the amount of business data over a certain future period).

[0159] II. Channel State Information

[0160] In wireless communication, Channel Identity (CSI) refers to the known channel properties of a communication link. This information describes how a signal propagates from the transmitter to the receiver and represents the combined effects of scattering, fading, and power attenuation with distance. This method is called channel estimation. CSI enables transmission to adapt to current channel conditions, which is crucial for achieving reliable communication at high data rates in multi-antenna systems. In practical implementation and application, the base station sends a CSI reference signal to the terminal equipment for measurement. The terminal equipment calculates various values ​​through measurement and reports them to the base station for CSI acquisition or beam management, or it may not report these values ​​but instead use them to select the receiving beam.

[0161] III. Dual Connection

[0162] Dual connectivity (DC): A connected terminal device is considered to have dual connectivity when the network configures it with a master cell group (MCG) and a secondary cell group (SCG). Dual connectivity can exist in several combinations:

[0163] 1) When the core network is EPC, the LTE base station acts as the primary station and the NR base station acts as the secondary station.

[0164] 2) When the core network is 5GC, the LTE base station acts as the primary station and the NR base station acts as the secondary station.

[0165] 3) When the core network is 5GC, the NR base station acts as the primary station and the LTE base station acts as the secondary station.

[0166] 4) When the core network is 5GC, both the primary and secondary stations are NR base stations.

[0167] It should be noted that the embodiments of this application are not limited to the above-mentioned scenarios, and can also be applied to various future communication systems and future communication networks, such as 6G communication systems, which include 6G core networks and 6G base stations.

[0168] IV. AI-based beam management and beam selection

[0169] Please see Figure 4 , Figure 4This is a schematic diagram of AI-based beam management. Based on AI prediction, it can predict the beams in set A (SetA) based on the measurement results of beams in set B (SetB). SetB can be a relatively sparse beam set, and SetA can be a beam set that includes SetB, or a beam set different from SetB. After obtaining the measurement results of SetB (as shown in the scanning beam measurement diagram), the beam set in SetA is predicted based on the SetB measurement results. The prediction result can be the Top-K beams in the SetA beam set (the beams corresponding to the Top-K indices in the diagram). The Top-K beams are the K best beams in set A. Specifically, the Top-K beams can be identified by having better signal quality L1-reference signal receiving power (RSRP).

[0170] V. AI / ML-based localization

[0171] AI / ML-based localization can be divided into two types: the first type is direct localization based on AI / ML, and the second type is AI / ML-assisted localization, which are described in detail below:

[0172] The first type: Direct localization based on AI / ML

[0173] 1. Apply terminal device-side models to directly perform positioning based on AI / ML. For example, the terminal device performs signal measurements, and the positioning result is directly predicted based on the measurement results.

[0174] 2. Terminal device assisted positioning management function (LMF) network element positioning, such as LMF network element receiving and processing positioning requests or positioning-related data requests, selecting positioning methods based on the requests and measuring positioning results, LMF network element applying AI / ML models for positioning based on auxiliary information from the terminal device side.

[0175] 3. Network-side assisted LMF network element localization: LMF network elements are located using AI / ML models based on network-side auxiliary information.

[0176] The second type: AI / ML-assisted localization

[0177] 1. Terminal equipment side model-assisted LMF network element localization: The terminal equipment side makes predictions based on the measurement results of the terminal equipment and sends the prediction results to the LMF network element to assist the LMF network element in localization.

[0178] 2. Access network side model-assisted LMF localization: The access network side makes predictions based on the measurement results and sends the prediction results to the LMF network element to assist the LMF in localization.

[0179] VI. AI / ML-based CSI Feedback Enhancement

[0180] Currently, communication systems use codebooks as the basic tool for CSI feedback, and multiple schemes, such as Type I and Type II codebooks, have been defined to meet different feedback accuracy requirements. However, these codebooks are all designed for uniformly arranged antenna arrays and are not optimized for special antennas such as 3D antennas, resulting in significant performance limitations. AI-based CSI feedback can overcome these bottlenecks by optimizing for specific channel environments to achieve better feedback performance. The basic principle of AI-based CSI feedback is to treat the high-dimensional channel information feedback task as an end-to-end CSI image compression and reconstruction task. Please refer to [link / reference]. Figure 5 , Figure 5 This diagram illustrates an AI / ML-based CSI feedback enhancement approach. Its basic signal flow resembles an autoencoder: 1) The encoder (typically the terminal side) uses an encoder to extract features from the complete channel information and compress it into a bitstream that meets feedback requirements; 2) This information is fed back to the decoder (typically the base station side) via a feedback link; 3) The decoder uses a decoder to decompress the bitstream and reconstruct its features, ultimately recovering the complete channel information. The encoder includes feature extraction and feature compression modules, while the decoder includes feature decompression and feature recovery modules. Feature extraction and feature recovery can utilize convolutional networks, while feature compression and decompression can use fully connected networks. The encoder and decoder are jointly optimized during end-to-end training to achieve optimal CSI reconstruction performance. In actual deployment, the encoder and decoder need to be paired according to the training process; that is, the compressed CSI output from one encoder needs to be recovered using the corresponding decoder.

[0181] Furthermore, AI / ML-based CSI prediction refers to the technology of obtaining unknown time-frequency resource CSIs based on existing CSIs without increasing new air interface resource overhead. While CSIs at different times / spaces are not entirely identical, they exhibit a certain degree of correlation, making CSI prediction possible. Traditional CSI prediction schemes are limited by prediction accuracy when handling complex data, hindering practical application. AI-based CSI prediction is expected to significantly improve prediction accuracy, thus potentially achieving the goal of obtaining unknown CSIs with low overhead in real-world systems. Based on data correlation categories, AI-based CSI prediction can be divided into four types: 1) The first type considers temporal correlation, i.e., predicting the CSI of the next time moment or the next time moment based on the CSI of the previous period. This is mainly applied in time-varying channels or high-speed mobile scenarios. 2) The second type considers frequency-related correlation, such as predicting and reconstructing downlink CSIs based on the uplink CSI of frequency division duplexing (FDD). 3) The third type considers spatial-related prediction problems. 4) The fourth category considers the channel correlation between adjacent users. For users in the same scenario and within the service range of the same base station, their channels often have a strong correlation, and the channel strength enjoyed by users in different locations may be different.

[0182] VII. AI / ML-based Mobility Management

[0183] During mobility management, access network equipment configures terminal devices to report measurements and makes handover decisions based on the reported results. With the introduction of AI, access network equipment can make predictions based on the limited measurement results from terminal devices (e.g., configuring a small number of measurement beams, or reducing the number of beams and cell measurements by the terminal devices) and select the optimal cell for handover decisions.

[0184] Please see Figure 6 , Figure 6 This is a schematic diagram illustrating how a network configuration terminal device performs functional or model reasoning and monitoring, as shown below:

[0185] Step 1: The network sends instruction information to the terminal device.

[0186] This instruction information is used to instruct the terminal device to report its capability information.

[0187] Step 2: The terminal device sends its capability information to the network.

[0188] The capability information of the terminal device can refer to the AI ​​capabilities that the terminal device supports. For example, it could indicate AI / ML functions or models in beam management scenarios, or AI / ML functions or models in positioning scenarios. In other possible cases, the terminal device can report indications of supporting AI / ML functions.

[0189] Step 3: The network sends RRC configuration information to the terminal device.

[0190] Specifically, based on the AI / ML functions or models supported by the terminal device, the network can send RRC configuration information to the terminal device according to the network's needs, in order to request the terminal device to report available AI / ML functions or models.

[0191] Step 4: The terminal device sends auxiliary information to the network.

[0192] Among them, the auxiliary information of the terminal device can refer to the terminal device reporting available AI / ML functions or models based on the network configuration.

[0193] Step 5: The network sends RRC configuration information to the terminal device.

[0194] The network provides AI / ML configuration to terminal devices based on the information reported by the terminal devices, namely, the reported available AI / ML functions or models.

[0195] Step 6: The terminal device performs functions such as registration, activation, inference, and monitoring.

[0196] Specifically, based on RRC configuration information, the terminal device performs AI inference, monitoring, and other functions.

[0197] When configuring terminal devices to perform AI / ML-related operations, it is necessary to ensure that the AI ​​model or function on the terminal device side is suitable for the current network conditions. For example, the training data used by the AI ​​model / function on the terminal device side is collected under certain network conditions. Therefore, when configuring the terminal device to execute an AI model / function, the configured AI model / function must be suitable for the current network conditions. That is, the network conditions during AI inference should be consistent with the network conditions during model / function training. Only in this way can the model achieve better inference performance.

[0198] Currently, associated IDs can be used to indicate network-side conditions, indicating that the associated ID is valid within a cell. That is, the model trained on the terminal device side is applicable to network-side conditions corresponding to a specific associated ID within a particular cell. If the terminal device switches to another cell, or the network-side conditions change, the current model will no longer be applicable. When associated IDs under a cell group are introduced, meaning a model trained on the terminal device side under multiple cells corresponding to a specific associated ID, it can be applied across all these cells. For example, a cell group includes cells 1 through 10 (these cells can be under one or more base stations), and the terminal device trains a model under this cell group corresponding to network-side condition 1 (identified as associated ID 1). Then, when the terminal device switches between cells 1 through 10 and cells 1 through 10 meet network-side condition 1, the trained model can be applied.

[0199] VIII. Functional Monitoring and Model Management

[0200] In the lifecycle management (LCM) process of AI / ML functions or models, monitoring of the functions or models is necessary. Functionality monitoring is a crucial part of LCM, used to monitor the performance of AI / ML functions. This involves monitoring the performance of AI / ML functions or models based on specific metrics to inform subsequent decision-making. For example, when poor performance is detected, the function / model should be managed, such as switching or deactivating the AI / ML function or model. Taking beam management as an example, in AI-based beam management, the following metrics can be applied for model / function monitoring: key performance indicators (KPIs) related to beam prediction accuracy, such as Top-K / 1 beam prediction accuracy; link quality-related KPIs, such as throughput, L1-RSRP; metrics of AI / ML-based input / output data distribution; and the difference between measured and predicted RSRP values, such as the difference between measured and predicted L1-RSRP values.

[0201] AI / ML model management can be performed based on monitoring of AI / ML models / functions, or based on changes in terminal device / network conditions. This includes model activation, selection, switching, and rollback. Model management includes scenarios where management decisions are made by the network or terminal devices. For network-side decisions, this can be initiated by the network or by the terminal device, which then requests the network. For terminal device-side decisions, this can be triggered by network-configured events that report the terminal device's decision to the network, or it can be a decision made autonomously by the terminal device. The process for each scenario is described in detail below.

[0202] 1. Network triggering, network decision-making

[0203] Please see Figure 7 , Figure 7 This is a schematic diagram of network-triggered, network-decision-based model management. The network sends configuration information, such as measurement configuration and reporting configuration, to the terminal device. The terminal device sends performance or auxiliary information, such as measurement information, to the network. Based on the reported information from the terminal device, i.e., performance or auxiliary information, the network decides whether to perform model management. The network sends management instruction information to the terminal device, which includes management results, such as model activation, selection, switching, and rollback.

[0204] 2. Terminal device triggers network decision.

[0205] Please see Figure 8 , Figure 8 This is a schematic diagram of model management triggered by a terminal device and decided by the network. The network sends configuration information to the terminal device, such as measurement configuration and reporting configuration. The terminal device decides whether to perform model management based on the measurement results. The terminal device sends a management request to the network. The network performs model management based on the management request of the terminal device. The network sends management instruction information to the terminal device. The management instruction information includes management results, such as model activation, selection, switching, and rollback.

[0206] 3. Network triggering, terminal device decision-making

[0207] Please see Figure 9 , Figure 9 This is a schematic diagram of model management that is network-triggered and terminal device makes decisions. The network sends configuration information to the terminal device, which may include measurement configuration, event triggering conditions, etc. After the event triggering conditions are met, the terminal device makes a decision to perform model management and sends a management decision report, i.e., management result, to the network. The management result includes model activation, selection, switching, rollback, etc.

[0208] 4. Terminal device triggers, terminal device makes decisions.

[0209] Please see Figure 10, Figure 10 This is a schematic diagram of model management triggered and decided by a terminal device. The network sends a decision report configuration information, which can also be measurement configuration information, to the terminal device. The terminal device makes decisions on model management based on the measurement results and terminal device-side conditions. The terminal device sends a management decision report, i.e., the management result, to the network. This management result includes model activation, selection, switching, rollback, etc.

[0210] The above describes how to monitor the terminal device-side model within a single network device and manage the model based on the monitoring results. However, the network side may configure the terminal-side AI model to operate across multiple cells or network devices. Therefore, during terminal device mobility, if the source node detects that the performance or prediction accuracy of an AI model is not good, but the terminal device moves from the source node to the target node, the target node, lacking awareness of the model's performance, may still configure and apply the model. This could lead to poor model performance at the target node. Therefore, how to achieve continuous model monitoring and ensure performance when the network side may configure the terminal-side AI model to operate across multiple cells or network devices during terminal device mobility is a technical problem that those skilled in the art are currently solving. To address the above problems, the embodiments of this application propose the following solutions.

[0211] The following is combined Figure 1A The communication system shown provides a detailed description of a model monitoring method provided in this application embodiment. Please refer to... Figure 11 , Figure 11 This is a schematic diagram of a model monitoring method provided in an embodiment of this application. The method includes, but is not limited to, the following steps:

[0212] Step S1101: The first node sends the first information to the second node.

[0213] Correspondingly, the second node receives the first information from the first node.

[0214] The first node can be Figure 1A The access network device 102 in the middle, the second node can be Figure 1A In the access network device 103, the terminal device can be moved from the first node to the second node, that is, the terminal device can be moved from the access network device 102 to the access network device 103. Optionally, the first node can also be called the source node, and correspondingly, the second node can be called the target node.

[0215] The first information includes monitoring information of the first model on the first node side, and further includes at least one of the following: identification information of the first model, or network-side condition identification information corresponding to the first model. Optionally, the first information can be carried in a channel request (HOrequest) message. It should be noted that the model in this application can also be replaced by a model / function.

[0216] The first model is an AI / ML model or an AI / ML function. The monitoring information of the first model on the first node side includes at least one of the following: identification information of the cells the terminal device has stayed in and / or the identification information of the cell the terminal device is currently in, performance information of the first model, and second monitoring configuration information, where the second monitoring configuration information is determined by the first node and used for monitoring the first model. The identification information of the cells the terminal device has stayed in can also be described as the identification information of the cells the terminal device has passed through. The identification information of the cell the terminal device is currently in can include the identification information of the first node. The monitoring information of the first model on the first node side can include monitoring information of the first model under one or more cells on the first node side, with one cell corresponding to one monitoring information. For example, the identification information of one of the one or more cells can correspond to one monitoring information. The one or more cells can be cells the terminal device has stayed in. For example, the cells the terminal device has stayed in include cell 1 and cell 2, where the identification information of cell 1 is identifier 1, and the identification information of cell 2 is identifier 2. Identifier 1 corresponds to monitoring information 1, and identifier 2 corresponds to monitoring information 2. The monitoring information of the first model on the first node side can include: monitoring information 1 corresponding to identifier 1 and monitoring information 2 corresponding to identifier 2. The second monitoring configuration information may include at least one of the following: configuration information configured by the first node for the terminal device and used by the terminal device to perform monitoring operations of the first model; or configuration information recommended by the first node to the second node for the second node to configure for the terminal device to perform monitoring operations of the first model. The configuration information configured by the first node for the terminal device and used by the terminal device to perform monitoring operations of the first model can be referred to as the monitoring configuration information corresponding to the first model configured by the first node. The configuration information recommended by the first node to the second node for the second node to perform monitoring operations of the first model can be referred to as the monitoring configuration information corresponding to the first model recommended by the first node to the second node. The monitoring configuration information corresponding to the first model recommended by the first node to the second node may include configuration information recommended by the first node for the terminal device to perform monitoring of the first model, or, in other words, configuration information recommended by the first node for the second node to send to the terminal device for monitoring the first model after the terminal device moves from the first node to the second node. Of course, the second node may use the monitoring configuration information recommended by the first node to send to the terminal device, or it may not use the recommended monitoring configuration information; this application embodiment does not limit this.

[0217] In the above method, the monitoring information of the first model under one or more cells on the first node side can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0218] The performance information of the first model can be reported by the terminal device or determined by the first node, as detailed below:

[0219] When the performance information of the first model is reported by the terminal device, the performance information of the first model may include one or more of the following: predicted values ​​determined based on the first model and measured values ​​determined by the terminal device; the difference between the predicted values ​​determined by the first model and the measured values ​​determined by the terminal device; system performance information; accuracy information of the first model's prediction; or distribution information of the input data and / or output data corresponding to the first model. System performance information may include key performance indicators (KPIs), such as link throughput. The accuracy information of the first model's prediction can be determined based on the predicted values ​​determined by the first model and the measured values ​​determined by the measurement, and may be expressed as percentage, absolute error, root mean square error, or confidence interval, for example. The distribution information of the input data and / or output data corresponding to the first model may also be referred to as the measurement information of the distribution of the input data and / or output data corresponding to the first model. This distribution information of the input data and / or output data can be determined by the following measurement methods, such as mean, variance, and standard deviation.

[0220] Optionally, the performance information of the first model can be reported by the terminal device once or multiple times; this embodiment does not limit this. The performance information of the first model corresponds to one or more cells. For example, the terminal device reports the performance information of the first model once for each cell it stays in. The performance information of the first model can be obtained by the terminal device applying the first model for inference and monitoring based on the second monitoring configuration information configured by the first node, and by the terminal device performing measurements. Alternatively, it can be obtained by the terminal device applying the first model for inference and monitoring based on the monitoring configuration information corresponding to the first model configured by other nodes in the cells that the terminal device has stayed in, excluding the first node, and by the terminal device performing measurements. Optionally, before the terminal device applies the first model for inference and monitoring based on the second monitoring configuration information configured by the first node, the first node performs model selection and inference configuration based on at least one model reported by the terminal device. The at least one model is a model available on the terminal device side. For example, the first node selects the first model from the at least one model and determines the inference configuration of the first model. Accordingly, the first node instructs the terminal device to apply the first model for inference. Then, the first node sends the second monitoring configuration information to the terminal device. The second monitoring configuration information includes configuration information configured by the first node for the terminal device and used by the terminal device to perform the monitoring operation of the first model.

[0221] For example, when the first model is applied to beam management, the performance information of the first model may include the accuracy information of the first model's predictions, and the difference between the predicted values ​​determined based on the first model and the measured values ​​determined by the terminal device. The accuracy information of the first model's predictions may include the accuracy of the Top K beam predictions (where K>=1). For example, the accuracy of the Top K beam predictions is determined by the following methods: the terminal device obtains measured Top K beam information (measured values ​​determined by the terminal device) based on beam measurements, and simultaneously obtains predicted Top K beam information (predicted values ​​determined based on the first model) based on inference from the first model; the accuracy of the predictions is determined based on the measured Top K beam information and the predicted Top K beam information. The difference between the predicted value determined by the first model and the measured value determined by the terminal device can include the difference between the predicted value and the measured value of the Top K beam quality. For example, the predicted value is determined based on the beam signal quality predicted by the first model (the predicted value determined by the first model), and the measured value is determined by the L1-RSRP of the beam measured by the terminal device (the measured value determined by the terminal device), thereby determining the difference between the predicted value and the measured value. The difference between the predicted value determined by the first model and the measured value determined by the terminal device can also include the difference between the predicted value and the measured value of the signal quality of beams in a specific set. For example, a network device can configure a set of beams for monitoring, and obtain the measured value (the measured value determined by the terminal device) by measuring the beams in the set, and determine the predicted value (the predicted value determined by the first model) based on the beam signal quality predicted by the first model, thereby obtaining the difference between the predicted value and the measured value of the signal quality of the beams in the set.

[0222] For example, when the performance information of the first model is determined by the first node, the performance information of the first model includes at least one of the following: a measurement result determined by the first node in measuring the performance of the first model; a predicted value determined based on the first model and a measurement value determined by the terminal device; the difference between the predicted value determined by the first model and the measurement value determined by the terminal device; system performance information; accuracy information of the first model's prediction; or distribution information of the input data and / or output data corresponding to the first model. The measurement result determined by the first node in measuring the performance of the first model may include an indication of the first node's performance measurement of the first model. Optionally, the measurement result or performance measurement indication may include a quality level of the first model, for example, the quality level of the first model includes excellent, good, average, or poor.

[0223] In the above method, the performance information of the first model can be used to determine the historical performance or historical monitoring status of the first model. The historical performance or historical monitoring status of the first model serves as a reference for the second node to determine the first monitoring configuration information, thus ensuring the continuity of the monitoring of the first model and the performance of the first model.

[0224] The identification information of the first model refers to information used to characterize the identity of the first model, which may include the model's name, version, creator, or creation time. The network-side condition identification information corresponding to the first model may refer to the identification information of the network-side conditions applicable to the first model, the network conditions corresponding to the first node, or the network-side conditions under which the training data used by the first model was collected. Since the training data used by the first model is collected under certain network-side conditions, when configuring the terminal device to execute the first model, the configured first model needs to be applicable to the current network-side conditions. That is, the network-side conditions during inference should be consistent with the network-side conditions when the training data of the first model was determined, so that the first model can obtain better inference performance. The network-side condition identification information corresponding to the first model can be an associated ID. The network-side conditions corresponding to this identification information may include one or more of the following: spatial arrangement of physical beams, beam gain, antenna height, antenna downtilt angle, or environmental parameters. For example, environmental parameters may include building density or terrain.

[0225] In one possible implementation, the method further includes: the terminal device determining the performance information of the first model, and storing the performance information of the first model when the terminal device is in an idle state or an inactive state.

[0226] Specifically, each time a terminal device switches to a new cell, transitioning from an idle state to a connected state, or from an inactive state to a connected state, the terminal device sends performance information for the first model. Optionally, the terminal device sends the performance information for the first model to the access network device or the core network device. This can be understood as follows: before switching to a new cell, or while in an idle or inactive state, the terminal device stores the performance information for the first model corresponding to the cells it has previously visited. After switching to a new cell, transitioning from an idle state to a connected state, or from an inactive state to a connected state, the terminal device can report the stored performance information for the first model corresponding to the cells it has previously visited to the access network device or the core network device all at once. Optionally, when the terminal device is in a connected state, the terminal device stores the performance information for the first model.

[0227] For example, the cells where the terminal device has stayed include cell 1, cell 2, and cell 3. The identification information of cell 1 is identifier 1, the identification information of cell 2 is identifier 2, and the identification information of cell 3 is identifier 3. The performance information of the first model determined by the terminal device in cell 1 is information 1. Identifier 1 and information 1 are stored in list 1. The performance information of the first model determined in cell 2 is information 2. Identifier 2 and information 2 are stored in list 1. The terminal device is in an idle state in cell 1 and cell 2 and in a connected state in cell 3. When the terminal device enters the connected state from the idle state, the terminal device can send list 1 to the access network device or core network device. List 1 includes: identifier 1 and information 1, as well as identifier 2 and information 2.

[0228] In the above method, when the terminal device is in an idle or inactive state, the performance information of the first model corresponding to each cell that the terminal device has stayed in is released, and the performance information of the first model corresponding to each cell that has stayed in is lost. By storing the performance information of the first model corresponding to each cell that has stayed in the terminal device in the idle or inactive state, the loss of information can be avoided.

[0229] In another possible implementation, the method further includes: the core network device receiving the performance information of the first model, and the core network device storing the performance information of the first model.

[0230] The performance information of the first model received by the core network device can be reported to the core network device by the terminal device in the connected state through the access network device, or it can be reported by the access network device.

[0231] Since the terminal device releases the performance information of the first model corresponding to each cell it has stayed in when it is in an idle or inactive state, and this performance information is lost, the terminal device obtains the performance information of the first model corresponding to each cell it has stayed in from the core network device when it enters the connected state from the idle state or from the active state.

[0232] In the above method, when the terminal device enters the idle state from the connected state, the performance information of the first model corresponding to each cell that the terminal device has stayed in is released, and the performance information of the first model corresponding to each cell that has stayed in is lost. By storing the performance information of the first model corresponding to each cell that has stayed in the core network device in the idle state or the inactive state, the loss of information can be avoided.

[0233] In one possible implementation, the method further includes: the first node determining whether the first model is available on the second node side.

[0234] Optionally, before sending the first information to the second node, the first node determines whether the first model is available on the second node's side.

[0235] In one possible implementation, the first node determines that the first model is available on the second node's side, including: the first node and the second node belong to a first cell group, and the first model is available in the first cell group; or, the first node and the second node are within a first geographical area, and the first model is available in the first geographical area. In another possible implementation, the source cell in the first node and the target cell in the second node are within the same cell group or within the same geographical area. This can be understood as follows: if the first node and the second node belong to the same cell group, or are within the same geographical area, or correspond to the same group of base stations, the network-side conditions corresponding to the first model are valid within that same cell group, the same physical area, or the same group of base stations. When the terminal device moves out of that cell group, the same physical area, or the same group of base stations, the first model is no longer applicable. For example, the first cell group includes cell 1 and cell 2. Cell 1 is under the first node and cell 2 is under the second node. The terminal device is under cell 1. The first model is trained according to network-side condition 1 (identified as associated ID 1). When the terminal device switches between cell 1 and cell 2 (e.g., the terminal device moves from the first node to the second node) and cell 1 and cell 2 meet network-side condition 1, the first model can be applied.

[0236] In another possible implementation, the first node determines that the first model is available on the second node side, including: the network conditions corresponding to the second node side are consistent with or similar to the network conditions corresponding to the network condition identifier information of the first model. This can be understood as the network conditions corresponding to the second node side being consistent with or similar to the network conditions corresponding to the first node side, or the network conditions corresponding to the second node side being consistent with or similar to the network conditions when collecting the training data applied to the first model. In yet another possible implementation, the first node determines that the first model is available on the second node side, including: the first node belongs to the first cell group, the second node belongs to the second cell group, the first model is available in both the first and second cell groups, and the network conditions corresponding to the second node side are consistent with or similar to the network conditions corresponding to the network condition identifier information of the first model. In yet another possible implementation, the first node determines that the first model is available on the second node side, including: the first node is within the first physical region, the second node is within the second physical region, the first model is available in both the first and second physical regions, and the network conditions corresponding to the second node side are consistent with or similar to the network conditions corresponding to the network condition identifier information of the first model. Optionally, the first node may determine the availability of the first model in the first cell group and the second cell group based on implementation, or the first node may determine the availability of the first model in the first physical area and the second physical area based on implementation. Implementation-based determination may include through network management or inter-site interaction. The above process can be understood as follows: when the network-side conditions corresponding to the network-side condition identifier of the first model are valid in at least one cell group (e.g., including the first cell group and the second cell group) or at least one physical area (e.g., including the first physical area and the second physical area), the first node needs to determine whether the network-side conditions of the first cell group where the first node is located (the network-side conditions corresponding to the network-side condition identifier of the first model) are applicable in the second cell group where the second node is located. This applicability may include the network-side conditions of the first cell group where the first node is located (the network-side conditions corresponding to the network-side condition identifier of the first model) being similar to or consistent with the network-side conditions of the second node. In one example, the first model is available under a first cell group and a second cell group. The first cell group includes cell 1 and cell 2, and the second cell group includes cell 3. Cell 1 and cell 2 are under a first node, and cell 3 is under a second node. The terminal device is under cell 1. The first model is trained according to network-side condition 1 (identified as associated ID 1). Then, when the terminal device switches between cell 1 and cell 3 (e.g., the terminal device moves from the first node to the second node) and cell 1 and cell 3 meet network-side condition 1, the first model can be applied.

[0237] In the above method, the availability of the first model can be judged, the normal operation of the first model can be guaranteed, and the effectiveness of the application of the first model can be ensured.

[0238] In another possible implementation, the method further includes: a first node receiving first indication information from a terminal device, the first indication information being used to indicate that a first model is available on the second node side.

[0239] Optionally, before the first node sends the first information to the second node, the terminal device sends the first indication information to the first node. Accordingly, the first node receives the first indication information from the terminal device, and then the first node can send the first information to the second node based on the first indication information.

[0240] In one possible implementation, the terminal device determines that the first model is available on the second node side, including: the terminal device determining that the first node and the second node belong to a first cell group, and the first model is available in the first cell group; or the terminal device determining that the first node and the second node are within a first geographical area, and the first model is available in the first geographical area. The specific process can be referred to in the process of the first node determining that the first model is available on the second node side, and will not be repeated here.

[0241] In another possible implementation, the network-side conditions corresponding to the first model are valid within the first cell group. The terminal device determines the configuration information of the first cell group, which can also be understood as the terminal device being able to determine that the first model is available within the first cell group. Therefore, when the terminal device reports the performance information of the first model to the first node, it needs to indicate whether the second cell is within the first cell group, thereby further instructing whether the first node needs to send the first information to the second node. For example, if the second cell is within the first cell group, the first node sends the first information to the second node; or if the second cell is within the first cell group, the first node does not need to send the first information to the second node.

[0242] In another possible implementation, the network-side conditions corresponding to the first model are valid within the first cell group. The configuration information of the first cell group can be configured by the network management system. The first node and the second node can determine the configuration information of the first cell group. Accordingly, when the terminal device trains the first model, it can obtain the association information between the first model and the first cell group. For example, if the first model is available within the first cell group, that is, the network-side conditions corresponding to the first model are valid within the first cell group, then when the first node and the second node belong to the first cell group, the first node can send the first information to the second node. Accordingly, the second node determines the first monitoring configuration information based on the first information.

[0243] In another possible implementation, the terminal device determines that the first model is available on the second node side, including: the network conditions corresponding to the second node side are consistent with or similar to the network conditions corresponding to the network side condition identifier information of the first model. The specific process can be referred to the process by which the first node determines that the first model is available on the second node side, and will not be repeated here. In yet another possible implementation, the terminal device can determine that the first model is available in the first cell group and the second cell group based on implementation, or the terminal device can determine that the first model is available in the first physical area and the second physical area based on implementation. When the terminal device reports the performance information of the first model to the first node, it needs to indicate whether the network side conditions of the first cell group where the first node is located (the network side conditions corresponding to the network side condition identifier of the first model) are similar to or consistent with the network side conditions of the second node side, thereby further instructing the first node whether it needs to send the first information to the second node.

[0244] In the above method, the availability of the first model can be judged, the normal operation of the first model can be guaranteed, and the effectiveness of the application of the first model can be ensured.

[0245] In another possible implementation, before the first node sends the first information to the second node, the method further includes: the first node receiving model information supported by the second node or model information of interest from the second node; the first node determining to send the first information to the second node based on the model information supported by the second node or model information of interest; the first information including monitoring information of the first model on the first node side.

[0246] For example, when the model information supported by the second node or the model information of interest includes the first model, the first node determines to send the monitoring information of the first model on its side to the second node; when the model information supported by the second node or the model information of interest does not include the first model, the first node determines not to send the monitoring information of the first model on its side to the second node. Optionally, the first node may also send the model information supported by the first node or the model information of interest to the second node. This process can be understood as the first node and the second node being able to interact with each other using their supported model information or the model information of interest. The first node can determine whether to send the monitoring information of the second node's supported model information or the model information of interest to the second node based on the model information supported by the second node or the model information of interest on its side.

[0247] Step S1102: The second node determines the first monitoring configuration information based on the first information.

[0248] The first monitoring configuration information is used to configure the terminal device to perform monitoring operations of the first model. This can be understood as follows: when the terminal device moves from the first node to the second node, it monitors the first model based on the first monitoring configuration information issued by the second node. Optionally, the first monitoring configuration information can be the same as or different from the second monitoring configuration information. The second monitoring configuration information includes configuration information recommended by the first node to the second node for configuring the terminal device to perform monitoring operations of the first model. This embodiment of the application does not limit this. For example, the second node determines that the first model performs better on the first node side based on the first information, and the second node determines that the first monitoring configuration information can be relatively sparse monitoring.

[0249] Optionally, the second node can also determine the inference configuration information of the first model based on the first information. The second node can also send the inference configuration information of the first model to the terminal device through the first node. Accordingly, when the terminal device moves from the first node to the second node, the terminal device performs inference of the first model based on the inference configuration information of the first model.

[0250] Step S1103: The second node sends the first monitoring configuration information to the first node.

[0251] Accordingly, the first node receives the first monitoring configuration information from the second node. Optionally, the first monitoring configuration information can be carried in a channel acknowledgment (HOACK) message.

[0252] Step S1104: The first node sends the first monitoring configuration information to the terminal device.

[0253] Accordingly, the terminal device receives first monitoring configuration information from the first node. Based on this first monitoring configuration information, the terminal device performs monitoring operations using the first model to determine monitoring information, which may include predicted values ​​determined by the first model and / or measured values ​​determined by measurement. Optionally, the terminal device may send the monitoring information to the second node.

[0254] Optionally, the first monitoring configuration information can be carried in RRC signaling.

[0255] Optionally, the first node can also send handover configuration information to the terminal device, which is used to configure the terminal device to switch from the first node to the second node.

[0256] It should be noted that the above process is illustrated by the example of a first node sending first information to a second node, the first information including monitoring information of a first model on the first node side, the second node determining first monitoring configuration information based on the first information, and the second node sending the first monitoring configuration information to the terminal device through the first node. This application may also include the following schemes: for example, the first node may send monitoring information of at least one model on the first node side to the second node, the second node may select one or more models from at least one model, for example, based on the model's performance, application scenario requirements, etc., the second node determines the monitoring configuration information of each of the one or more models based on the monitoring information of at least one model on the first node side, and the second node sends the monitoring configuration information of each model to the terminal device through the first node, thereby enabling the terminal device to perform the corresponding monitoring operation based on the monitoring configuration information of each model.

[0257] In another possible approach, the terminal device-side model may be available within a single cell. In this case, there is no continuity in model monitoring because the terminal device must use a new model for each new cell it enters.

[0258] exist Figure 11 In the described method, during the process of the terminal device moving from the first node to the second node, monitoring information of the first model on the first node side is sent to the second node. The second node can determine the historical performance or monitoring status of the first model based on this monitoring information. Therefore, the second node can select a function / model (i.e., the first model) for the terminal device based on the historical performance or monitoring status of the first model, and configure first monitoring configuration information. This first monitoring configuration information is used by the terminal device to monitor the first model, thus ensuring the continuity of model monitoring. For example, when the second node determines that the historical performance of the first model is good based on the monitoring information on the first node side, the first monitoring configuration information determined by the second node can be a relatively sparse model monitoring configuration; when the second node determines that the historical performance of the first model is poor based on the monitoring information on the first node side, the first monitoring configuration information determined by the second node can be a relatively dense model monitoring configuration. Furthermore, when the terminal device moves from the first node to the second node, the second node can determine the historical performance or historical monitoring status of the first model based on the monitoring information of the first model on the first node side. This avoids the problem that the second node may still configure and apply the first model due to a lack of performance awareness of the first model during the process of the terminal device moving from the first node to the second node, which would lead to poor performance of the first model on the second node side, thus ensuring the performance of the model.

[0259] Please see Figure 12 , Figure 12 This is a schematic diagram of a model monitoring method provided in an embodiment of this application. The method includes, but is not limited to, the following steps:

[0260] Step S1201: The second node sends the model information supported by the second node or the model information of interest to the first node.

[0261] Accordingly, the first node receives model information supported by the second node or model information of interest from the second node. The model information supported by the second node or model information of interest is used to determine whether to send the first information to the second node. For details, please refer to the relevant description in S1101.

[0262] This step is optional.

[0263] Step S1202: The first node selects a model and configures inference based on at least one model reported by the terminal device.

[0264] Wherein, the at least one model is a model available on the terminal device side. For example, the first node selects a first model from the at least one model and determines the inference configuration of the first model. Accordingly, the first node instructs the terminal device to apply the first model for inference.

[0265] Step S1203: The first node sends the second monitoring configuration information to the terminal device.

[0266] The second monitoring configuration information includes configuration information configured by the first node for the terminal device and used by the terminal device to perform monitoring operations of the first model. See the relevant description in step S1101 for details.

[0267] Step S1204: The terminal device monitors the first model based on the second monitoring configuration information to determine the performance information of the first model.

[0268] The performance information of the first model may include one or more of the following: the predicted value determined by the first model and the measured value determined by the terminal device; the difference between the predicted value determined by the first model and the measured value determined by the terminal device; system performance information; the accuracy information of the first model's prediction; or the distribution information of the input data and / or output data corresponding to the first model. Refer to the relevant description in step S1101 for details.

[0269] Step S1205: The terminal device sends the performance information of the first model to the first node.

[0270] Accordingly, the first node receives performance information of the first model from the terminal device, as detailed in the relevant description in step S1101.

[0271] Step S1206: The first node determines the handover decision for the terminal device.

[0272] The switching decision of the terminal device includes deciding to switch the terminal device from the first node to the second node.

[0273] Step S1207: The first node sends the first message to the second node.

[0274] Correspondingly, the second node receives the first information from the first node.

[0275] The first information includes monitoring information of the first model on the first node side, and further includes at least one of the following: identification information of the first model, or network-side condition identification information corresponding to the first model. Refer to the relevant description in step S1101 for details.

[0276] Step S1208: The second node determines the first monitoring configuration information based on the first information.

[0277] Please refer to the relevant description in step S1102 for details.

[0278] Step S1209: The second node sends the first monitoring configuration information to the first node.

[0279] Accordingly, the first node receives the first monitoring configuration information from the second node. See the relevant description in step S1103 for details.

[0280] Step S1210: The first node sends the first monitoring configuration information to the terminal device.

[0281] Accordingly, the terminal device receives the first monitoring configuration information from the first node. See the relevant description in step S1104 for details.

[0282] exist Figure 12In the described method, during the process of the terminal device moving from the first node to the second node, monitoring information of the first model on the first node side is sent to the second node. The second node can determine the historical performance or monitoring status of the first model based on this monitoring information. Therefore, the second node can select a function / model (i.e., the first model) for the terminal device based on the historical performance or monitoring status of the first model, and configure first monitoring configuration information. This first monitoring configuration information is used by the terminal device to monitor the first model, thus ensuring the continuity of model monitoring. For example, when the second node determines that the historical performance of the first model is good based on the monitoring information on the first node side, the first monitoring configuration information determined by the second node can be a relatively sparse model monitoring configuration; when the second node determines that the historical performance of the first model is poor based on the monitoring information on the first node side, the first monitoring configuration information determined by the second node can be a relatively dense model monitoring configuration. Furthermore, when the terminal device moves from the first node to the second node, the second node can determine the historical performance or historical monitoring status of the first model based on the monitoring information of the first model on the first node side. This avoids the problem that the second node may still configure and apply the first model due to a lack of performance awareness of the first model during the process of the terminal device moving from the first node to the second node, which would lead to poor performance of the first model on the second node side, thus ensuring the performance of the model.

[0283] The following is combined Figure 1B The communication system shown can be applied to dual-connectivity scenarios, such as scenarios involving the addition or updating of secondary nodes. A detailed description of a model monitoring method provided in this application embodiment is given below. Please refer to... Figure 13 , Figure 13 This is a schematic diagram of a model monitoring method provided in an embodiment of this application. The method includes, but is not limited to, the following steps:

[0284] Step S1301: The secondary node sends the model information supported by the secondary node or the model information of interest to the primary node.

[0285] This step is optional.

[0286] Among them, the auxiliary node is Figure 1B The auxiliary node in the middle, the primary node is Figure 1B The primary node in the process. Optionally, the primary node can be called the first node, and the corresponding secondary node can also be called the second node, as detailed in the relevant description in step S1101.

[0287] Steps S1302-S1305 can be referred to the relevant descriptions in steps S1202-S1205.

[0288] Step S1306: The master node decides whether to add or change a secondary node.

[0289] Step S1307: The master node sends a request message to the slave node.

[0290] Correspondingly, the secondary node receives request messages from the primary node.

[0291] This request message is used to request the addition or modification of a secondary node. The request message may include first information. A description of the first information can be found in the relevant description in step S1101.

[0292] Step S1308: The auxiliary node determines the first monitoring configuration information based on the first information.

[0293] Please refer to the relevant description in step S1102 for details.

[0294] Step S1309: The secondary node sends a response message to the primary node.

[0295] Correspondingly, the master node receives a response message from the slave node in response to the addition or modification of the slave node. This response message includes first monitoring configuration information, which is used for model monitoring between the terminal device and the second node. A description of the first monitoring configuration information can be found in step S1103.

[0296] Step S1310: The master node sends configuration information to the terminal device.

[0297] This configuration information may include primary monitoring configuration information, and may also include information for configuring the addition or modification of secondary nodes. Refer to the relevant description in step S1104 for details.

[0298] exist Figure 13In the described method, when auxiliary nodes are added or updated, by sending the monitoring information of the first model on the master node side to the auxiliary nodes, the auxiliary nodes can determine the historical performance or monitoring status of the first model based on this information. Therefore, the auxiliary nodes can select a function / model (i.e., the first model) for the terminal device based on this historical performance or monitoring status, and configure first monitoring configuration information. This first monitoring configuration information is then used by the terminal device to monitor the first model, ensuring the continuity of model monitoring. For example, when the auxiliary node determines that the historical performance of the first model is good based on the monitoring information on the master node side, the first monitoring configuration information determined by the auxiliary node can be a relatively sparse model monitoring configuration; when the auxiliary node determines that the historical performance of the first model is poor based on the monitoring information on the master node side, the first monitoring configuration information determined by the auxiliary node can be a relatively dense model monitoring configuration. Furthermore, when a terminal device moves from the master node to the slave node, the slave node can determine the historical performance or historical monitoring status of the first model based on the monitoring information of the first model on the master node side. This avoids the problem that the slave node may still configure and apply the first model due to a lack of performance awareness of the first model during the process of the terminal device moving from the master node to the slave node, which would lead to poor performance of the first model on the slave node side, thus ensuring the performance of the model.

[0299] The following is combined Figure 2 The communication system shown illustrates a model monitoring method provided in this application embodiment. It should be noted that CU and DU in the following description can simultaneously refer to the CU and DU in the first node, or simultaneously refer to the CU and DU in the second node. Please refer to... Figure 14 , Figure 14 This is a schematic diagram of a model monitoring method provided in an embodiment of this application. The method includes, but is not limited to, the following steps:

[0300] S1401: CU performs model selection.

[0301] The CU's model selection process includes: managing and determining the models configured on the terminal device side. For example, the CU can perform operations such as model selection / switching and activation / deactivation based on the model monitoring status or the needs of the terminal device side.

[0302] S1402: CU sends historical model monitoring information to DU.

[0303] Correspondingly, DU receives historical model monitoring information from CU.

[0304] Optionally, the CU can also send selected model information to the DU, such as the model's input and output data. In an AI-based beam management scenario, the selected model information sent by the CU to the DU includes a set A (SetA) and a set B (SetB) of models. Model monitoring history information refers to the historical monitoring information of the selected model. In one possible implementation, the model monitoring history information can be the historical monitoring information reported by the terminal device in the remaining cells of the CU, or it can be the historical monitoring information of the terminal device-side model received by the CU from the source node. For example, when the CU and DU are the CU and DU in the second node, the historical monitoring information of the terminal device-side model received by the CU from the source node includes: the first information received by the CU from the first node, as described above.

[0305] S1403: DU determines the model monitoring configuration information.

[0306] Among them, DU determines the model monitoring configuration information based on the selected model and model monitoring history information indicated by CU.

[0307] S1404: DU sends model monitoring configuration information to CU.

[0308] Correspondingly, the CU receives model monitoring configuration information from the DU, wherein the model monitoring configuration information includes configuration information for the terminal device to perform model monitoring operations. When the CU and DU are the CU and DU in the first node, and the selected model is the first model, the model monitoring configuration information can be the second monitoring configuration information; when the CU and DU are the CU and DU in the second node, and the selected model is the first model, the model monitoring configuration information can be the first monitoring configuration information.

[0309] In one possible implementation, S1403 and S1404 can also be replaced by the following steps: DU can send the auxiliary information required for monitoring to CU, and CU determines the model monitoring configuration information based on the selected model, the auxiliary information required for monitoring, and the model monitoring history information.

[0310] S1405: The CU sends model monitoring configuration information to the terminal device.

[0311] Optionally, the model monitoring configuration information can be carried in an RRC message.

[0312] S1406: The terminal device performs model monitoring based on the model monitoring configuration information to determine the monitoring information.

[0313] The monitoring information may include predicted values ​​determined by model prediction and / or measured values ​​determined by measurement. When the CU and DU are the CU and DU in the first node, and the selected model is the first model, the monitoring information may be the performance information of the first model reported by the terminal device.

[0314] S1407: The terminal device sends monitoring information to the DU.

[0315] Correspondingly, the DU receives monitoring information from the terminal device.

[0316] Optionally, the monitoring information can be carried in Layer 1 (L1) signaling.

[0317] For example, after the DU receives monitoring information from the terminal device, there are two possible implementation methods, namely Method 1 and Method 2, as follows:

[0318] Method 1 includes steps S14081 and S14091, as detailed below:

[0319] S14081: DU sends monitoring information to CU.

[0320] S14091: CU manages models based on monitoring information.

[0321] Method 2 includes steps S14082, S14092, and S14102, as detailed below:

[0322] S14082: DU determines auxiliary information for model management based on monitoring information.

[0323] The auxiliary information for model management may include the accuracy of model inference.

[0324] S14092: DU sends auxiliary information for model management to CU.

[0325] S14102: CU makes model management decisions based on auxiliary information for model management.

[0326] In one possible implementation, the model monitoring configuration information and the monitoring information reported by the terminal device are carried in Layer 2 (L2) or Layer 3 (L3) signaling, such as RRC messages. Optionally, in the signaling interaction between the terminal device and the CU, the DU plays a transparent role, that is, the terminal device can interact with the CU through the DU.

[0327] In another possible implementation, the model monitoring configuration information and the monitoring information reported by the terminal device are carried in L1 signaling. Optionally, the model monitoring configuration information is generated by the DU. The terminal device can send the monitoring information to the DU through L1 signaling. Correspondingly, the DU can send the monitoring information to the CU for the CU to make model management decisions, such as model switching, model deactivation, etc.

[0328] exist Figure 14 In the described method, the historical performance or historical monitoring status of the model is determined based on historical model monitoring information. Therefore, model monitoring configuration information can be configured for the terminal device based on the historical performance or historical monitoring status of the model, so that the terminal device can monitor the model based on the model monitoring configuration information.

[0329] The methods of the embodiments of this application have been described in detail above, and the apparatus of the embodiments of this application is provided below.

[0330] Please see Figure 15 , Figure 15 This is a schematic diagram of the structure of a model monitoring device 1500 provided in an embodiment of this application. The model monitoring device 1500 may include modules, units, or means that correspond one-to-one with the methods / operations / steps / actions executed by the first node, the second node, or the terminal device in the above method embodiments. The modules, units, or means may be hardware circuits, software, or a combination of hardware circuits and software.

[0331] In one possible implementation, the model monitoring device 1500 may include a processing unit 1501 and a transceiver unit 1502, the specific details of which are as follows:

[0332] The processing unit 1501 is used for data processing. The transceiver unit 1502 can implement corresponding communication functions. The transceiver unit 1502 can also be called a communication interface or a communication module.

[0333] Optionally, the model monitoring device 1500 may further include a storage unit, which can be used to store instructions and / or data. The processing unit 1501 can read the instructions and / or data in the storage module to enable the implementation of the aforementioned method embodiments.

[0334] Optionally, the transceiver unit 1502 may include a sending unit and a receiving unit. The sending unit is used to perform the sending operation in the above method embodiments. The receiving unit is used to perform the receiving operation in the above method embodiments.

[0335] It should be noted that the model monitoring device 1500 may include a transmitting unit but not a receiving unit. Alternatively, the model monitoring device 1500 may include a receiving unit but not a transmitting unit. Specifically, it depends on whether the above-described scheme executed by the model monitoring device 1500 includes both transmitting and receiving actions.

[0336] Optionally, the model monitoring device 1500 is used to perform the above. Figure 11 , Figure 12 , Figure 13 or Figure 14 The action performed by the first node in the illustrated embodiment. See the above for details. Figure 11 , Figure 12 , Figure 13 or Figure 14 The relevant descriptions in the illustrated embodiments are not elaborated here. For example, the model monitoring device 1500 is used to execute the following scheme:

[0337] The transceiver unit 1502 is used to send first information to the second node, the first information including monitoring information of the first model on the first node side; the transceiver unit 1502 is also used to receive first monitoring configuration information from the second node, the first monitoring configuration information is used to configure the terminal device to perform the monitoring operation of the first model, and the first monitoring configuration information is configured by the second node.

[0338] It should be noted that the implementation and beneficial effects of each module can be found by referring to [the relevant documentation / reference]. Figure 11 , Figure 12 , Figure 13 or Figure 14 The corresponding description of the method embodiments shown.

[0339] Optionally, the model monitoring device 1500 is used to perform the above. Figure 11 , Figure 12 , Figure 13 or Figure 14 The action performed by the second node in the illustrated embodiment. See the above for details. Figure 11 , Figure 12 , Figure 13 or Figure 14 The relevant descriptions in the illustrated embodiments are not elaborated here. For example, the model monitoring device 1500 is used to execute the following scheme:

[0340] The transceiver unit 1502 is configured to receive first information from the first node, the first information including monitoring information of the first model on the first node side; the processing unit 1501 is configured to determine first monitoring configuration information based on the first information, the first monitoring configuration information being used to configure the terminal device to perform the monitoring operation of the first model; the transceiver unit 1502 is also configured to send the first monitoring configuration information to the first node.

[0341] It should be noted that the implementation and beneficial effects of each module can be found by referring to [the relevant documentation / reference]. Figure 11 , Figure 12 , Figure 13 or Figure 14 The corresponding description of the method embodiments shown.

[0342] Optionally, the model monitoring device 1500 is used to perform the above. Figure 11 , Figure 12 , Figure 13 or Figure 14 The actions performed by the terminal device in the illustrated embodiment are shown above. For details, please refer to the above. Figure 11 , Figure 12 , Figure 13 or Figure 14 The relevant descriptions in the illustrated embodiments are not elaborated here. For example, the model monitoring device 1500 is used to execute the following scheme:

[0343] The processing unit 1501 is configured to determine the performance information of the first model, which includes one or more of the following: a predicted value determined based on the first model and a measured value determined by the model monitoring device, the difference between the predicted value determined by the first model and the measured value determined by the model monitoring device, system performance information, the accuracy information of the first model's prediction, or the distribution information of the input data and / or output data corresponding to the first model; the processing unit 1501 is configured to store the performance information of the first model when the model monitoring device is in an idle state or an inactive state.

[0344] It should be noted that the implementation and beneficial effects of each module can be found by referring to [the relevant documentation / reference]. Figure 11 , Figure 12 , Figure 13 or Figure 14 The corresponding description of the method embodiments shown.

[0345] The module division in this embodiment is illustrative and represents only one logical functional division; in actual implementation, other division methods may be used. The processing unit 1501 in the above embodiment can be implemented by at least one processor or processor-related circuitry. The transceiver unit 1502 can be implemented by a transceiver or transceiver-related circuitry. The transceiver unit 1502 can also be referred to as a communication module or communication interface. The storage module can be implemented using at least one memory.

[0346] Please see Figure 16 , Figure 16 This is a schematic diagram of the structure of a model monitoring device 1600 provided in an embodiment of this application. The model monitoring device 1600 may include modules, units, or means corresponding to the methods / operations / steps / actions executed by the first node, the second node, or the terminal device in the above method embodiments. The modules, units, or means may be hardware circuits, software, or a combination of hardware circuits and software.

[0347] The model monitoring device 1600 includes at least one processor 1601. Optionally, it also includes a communication interface 1603 and a memory 1602. The processor 1601, memory 1602, and communication interface 1603 are interconnected via a bus 1604. Optionally, the processor 1601 and memory 1602 can be integrated together.

[0348] The memory 1602 includes, but is not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), or compact disc read-only memory (CD-ROM), and is used for related computer programs and data. The communication interface 1603 is used for receiving and sending data.

[0349] Processor 1601 can be one or more central processing units (CPUs). When processor 1601 is a CPU, the CPU can be a single-core CPU or a multi-core CPU.

[0350] The processor 1601 in the model monitoring device 1600 is used to read the computer program or instructions stored in the memory 1602 to implement the functions of the above-mentioned processing unit, and the communication interface 1603 in the model monitoring device 1600 is used to implement the functions of the above-mentioned transceiver unit.

[0351] This application also provides a chip device including at least one processor for executing computer programs or instructions to cause the processor to perform the methods provided in the above embodiments.

[0352] In one possible implementation, the input of the chip device corresponds to the receiving operation in any of the above embodiments, and the output of the chip device corresponds to the sending operation in any of the above embodiments.

[0353] Optionally, the processor is coupled to the memory via an interface.

[0354] Optionally, the chip device may also include a memory in which computer programs or instructions are stored.

[0355] This application also provides a computer-readable storage medium storing a computer program or instructions that, when executed on a processor, implement the method described in the above method embodiments, which is executed by a first node, a second node, or a terminal device.

[0356] This application also provides a computer program product, which includes a computer program or instructions that, when run on a processor, implement the method executed by the first node, the second node, or the terminal device in the above method embodiments.

[0357] This application also provides a communication system, which includes a first node, a second node, and a terminal device as described in the above embodiments. The first node is used to execute some or all of the operations of the first node in the above method embodiments, the second node is used to execute some or all of the operations of the second node in the above method embodiments, and the terminal device is used to execute some or all of the operations executed by the terminal device in the above method embodiments.

[0358] It is understood that the processor in the embodiments of this application can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. A general-purpose processor can be a microprocessor or any conventional processor.

[0359] The method steps in the embodiments of this application can be implemented in hardware or by a processor executing software instructions. The software instructions can consist of corresponding software modules, which can be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disks, portable hard disks, CD-ROMs, or any other form of storage medium known in the art. An exemplary storage medium is coupled to a processor, enabling the processor to read information from and write information to the storage medium. Of course, the storage medium can also be a component of the processor. The processor and storage medium can reside in an ASIC. Alternatively, the ASIC can reside in a base station or terminal. Of course, the processor and storage medium can also exist as discrete components in the base station or terminal.

[0360] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer programs or instructions. When the computer program or instructions are loaded and executed on a computer, the processes or functions described in the embodiments of this application are performed entirely or partially. The computer can be a general-purpose computer, a special-purpose computer, a computer network, a network device, a user equipment, or other programmable device. The computer program or instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another. For example, the computer program or instructions can be transferred from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium, such as a floppy disk, hard disk, or magnetic tape; it can also be an optical medium, such as a digital video optical disc; or it can be a semiconductor medium, such as a solid-state drive. The computer-readable storage medium may be a volatile or non-volatile storage medium, or may include both types of storage media.

[0361] In the various embodiments of this application, unless otherwise specified or in case of logical conflict, the terminology and / or descriptions of different embodiments are consistent and can be referenced by each other. The technical features of different embodiments can be combined to form new embodiments according to their inherent logical relationship.

[0362] In the description of this application, terms such as "first", "second", "S1101" or "S1102" are used only for the purpose of distinguishing descriptions and for the convenience of context. Different sequence numbers do not have specific technical meanings themselves and should not be construed as indicating or implying relative importance, nor should they be construed as indicating or implying the order of execution of operations. The order of execution of each process should be determined by its function and internal logic.

Claims

1. A model monitoring method characterized by, Comprising: sending first information to a second node, the first information comprising monitoring information of a first model at a first node side; receiving first monitoring configuration information from the second node, the first monitoring configuration information being used for configuring a terminal device to perform monitoring of the first model, the first monitoring configuration information being configured by the second node.

2. The method of claim 1, wherein, The monitoring information of the first model at the first node side comprises at least one of: identification information of a cell where the terminal device has stayed and / or identification information of a cell where the terminal device is currently located; performance information of the first model; Or second monitoring configuration information, the second monitoring configuration information being configuration information for monitoring of the first model and being determined by the first node.

3. The method of claim 2, wherein, The second monitoring configuration information comprises at least one of: configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node.

4. The method according to claim 2 or 3, characterized in that, The performance information of the first model is reported by the terminal device or determined by the first node.

5. The method of claim 4, wherein, The performance information of the first model is reported by the terminal device, and the performance information of the first model comprises at least one of: a predicted value determined based on the first model and a measured value determined based on measurement of the terminal device; a difference between the predicted value determined based on the first model and the measured value determined based on measurement of the terminal device; system performance information; accuracy information of prediction of the first model; or distribution information of input data and / or output data corresponding to the first model.

6. The method of claim 4, wherein, The performance information of the first model is determined by the first node, and the performance information of the first model comprises at least one of: a measurement result determined by the first node based on measurement of performance of the first model; a predicted value determined based on the first model and a measured value determined based on measurement of the terminal device; a difference between the predicted value determined based on the first model and the measured value determined based on measurement of the terminal device; system performance information; accuracy information of prediction of the first model; or distribution information of input data and / or output data corresponding to the first model.

7. The method according to any one of claims 1 to 6, characterized in that, The first information further comprises at least one of: identification information of the first model, or network side condition identification information corresponding to the first model.

8. The method according to any one of claims 1-7, characterized in that, The method further comprises: receiving model information supported by the second node or model information of interest from the second node; determining to send the first information to the second node based on the model information supported by the second node or the model information of interest.

9. The method according to claim 7 or 8, characterized in that, The method further comprises: determining whether the first model is available at the second node side.

10. The method of claim 9, wherein, The determining that the first model is available at the second node side comprises: determining that the first node and the second node belong to a first cell group, and that the first model is available at the first cell group; or determining that the first node and the second node are within a first geographical area, and that the first model is available at the first geographical area.

11. The method according to claim 9 or 10, characterized in that, The determining that the first model is available at the second node side comprises: The network condition corresponding to the second node side is consistent or similar to the network condition corresponding to the network side condition identification information of the first model.

12. The method according to any one of claims 1-11, characterized in that, The method further comprises: receiving first indication information from the terminal device, the first indication information being used to indicate that the first model is available at the second node side.

13. A model monitoring method characterized by, comprises: receiving first information from the first node, the first information comprising monitoring information of the first model at the first node side; determining first monitoring configuration information based on the first information, the first monitoring configuration information being used to configure the terminal device to perform monitoring of the first model; sending the first monitoring configuration information to the first node.

14. The method of claim 13, wherein, The monitoring information of the first model at the first node side comprises at least one of: identification information of a cell where the terminal device has stayed and / or identification information of a cell where the terminal device is currently located; performance information of the first model; or second monitoring configuration information, the second monitoring configuration information being configuration information for monitoring of the first model determined by the first node.

15. The method according to claim 13 or 14, characterized in that, The second monitoring configuration information comprises at least one of: configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node.

16. The method according to any one of claims 13-15, characterized by, The performance information of the first model is reported by the terminal device or determined by the first node.

17. The method of claim 16, wherein, The performance information of the first model is reported by the terminal device, and the performance information of the first model comprises at least one of: a predicted value determined based on the first model and a measured value determined based on measurement of the terminal device; a difference between the predicted value determined based on the first model and the measured value determined based on measurement of the terminal device; system performance information; accuracy information of prediction of the first model; or distribution information of input data and / or output data corresponding to the first model.

18. The method of claim 16, wherein, The performance information of the first model is determined by the first node, and the performance information of the first model comprises at least one of: a measurement result determined by the first node based on measurement of performance of the first model; a predicted value determined based on the first model and a measured value determined based on measurement of the terminal device; a difference between the predicted value determined based on the first model and the measured value determined based on measurement of the terminal device; system performance information; accuracy information of prediction of the first model; or distribution information of input data and / or output data corresponding to the first model.

19. The method according to any one of claims 13-18, characterized by, The method further comprises: sending model information supported by the second node or model information of interest to the first node.

20. The method of any one of claims 13-19, wherein, The first information further comprises at least one of: identification information of the first model or network side condition identification information corresponding to the first model.

21. A model monitoring method characterized by, comprises: determining performance information of the first model, the performance information of the first model comprising one or more of: a predicted value determined based on the first model and a measured value determined based on a terminal device measurement, a difference between the predicted value determined based on the first model and the measured value determined based on the terminal device measurement, system performance information, accuracy information of the first model, or distribution information of input data and / or output data corresponding to the first model; storing the performance information of the first model when the terminal device is in an idle state or an inactive state.

22. The method of claim 21, wherein, The method further comprises: sending the performance information of the first model when the terminal device enters a connected state from an idle state or enters a connected state from an idle state.

23. The method of claim 21 or 22, wherein, The method further comprises: receiving first monitoring configuration information from a first node, the first monitoring configuration information being used to configure the terminal device to perform monitoring of the first model, the first monitoring configuration information being configured by the second node, the first monitoring configuration information being determined based on first information, the first information comprising monitoring information of the first model at the first node side.

24. The method of claim 23, wherein, The monitoring information of the first model at the first node side comprises at least one of: identification information of a cell in which the terminal device has stayed and / or identification information of a cell in which the terminal device currently stays; performance information of the first model; or second monitoring configuration information, the second monitoring configuration information being configuration information for monitoring of the first model determined by the first node. The second monitoring configuration information comprises at least one of:

25. The method of claim 24, wherein, configuration information configured by the first node for the terminal device; or configuration information recommended by the first node to the second node. The first information further comprises at least one of:

26. The method of any one of claims 23-25, wherein, identification information of the first model or network side condition identification information corresponding to the first model. The method further comprises:

27. The method of any one of claims 23-26, wherein, determining that the first model is available at a second node side; sending first indication information to a first node, the first indication information being used to indicate that the first model is available at the second node side. The determining that the first model is available at the second node side comprises:

28. The method of claim 27, wherein, determining that the first node and the second node belong to a first cell group, the first model being available at the first cell group; or determining that the first node and the second node are within a first geographical area, the first model being available at the first geographical area. The determining that the first model is available at the second node side comprises:

29. The method of claim 27 or 28, wherein, determining that a network condition corresponding to the second node side and a network condition corresponding to network side condition identification information corresponding to the first model are consistent or similar. ​

Images