Artificial intelligence model associated data reporting method, artificial intelligence model associated data reporting configuration method, communication device, and storage medium

Abstract

The present application provides an artificial intelligence (AI) model associated data reporting method, and an AI model associated data reporting configuration method. The AI model associated data reporting method is executed by a user equipment, and comprises: in a radio resource control (RRC) inactive state, storing AI-associated data; and after a user equipment enters an RRC connected state, reporting the AI-associated data to a first network side entity.

Description

Artificial intelligence model associated data reporting method, reporting configuration method, communication device and storage medium TECHNICAL FIELD

[0001] Embodiments of the present application relate to an artificial intelligence model associated data reporting method, an artificial intelligence model associated data reporting configuration method, a communication device and a storage medium. BACKGROUND

[0002] Data collection is very important in artificial intelligence and communication fusion (AIML-enabled wireless communication) systems. As shown in FIG. 1, the collected data is used for different model management function modules, such as performance monitoring (model monitoring or function monitoring), model training / model fine-tuning / model updating, model inference, etc. The data used for performance monitoring is called "monitoring data", the data used for model training / model fine-tuning / model updating is called "training data", and the data used for model inference is called "inference data". It needs to be considered how to ensure the quality of the collected data.

[0003] In the current use case of wireless network access side, only the data collection of user equipment (UE) in the radio resource control connected state (RRC_Connected) is concerned. However, in the actual data collection process, the RRC state of the UE may change, such as when a link failure or handover occurs, the UE will enter the RRC idle state permanently or temporarily from the RRC connected state. Then, in this case, how to collect data and ensure the quality of the collected data need further study.

[0004] SUMMARY

[0005] Embodiments of the present application provide an artificial intelligence model associated data reporting method, an artificial intelligence model associated data reporting configuration method, a communication device and a storage medium to solve the problems in the prior art.

[0006] In one aspect, the present application provides an artificial intelligence model associated data reporting method, which is executed by a user equipment (UE). The method includes: storing artificial intelligence model (AI) associated data in a radio resource control (RRC) non-connected state; and reporting the AI associated data to a first network side entity after the UE enters an RRC connected state.

[0007] In another aspect, the present application provides an artificial intelligence model associated data reporting configuration method, which is executed by a network side entity. The method includes: sending a storage indication message to a user equipment (UE), wherein the storage indication message is used to indicate whether the UE stores artificial intelligence model (AI) associated data when crossing an RRC connected state, or to indicate whether the UE keeps storing the AI associated data.

[0008] In another aspect, the present application provides an AI correlation data reporting method, executed by a user equipment (UE). The method comprises: storing AI correlation data; and reporting the AI correlation data to a network side entity when a preset condition is met.

[0009] In another aspect, the present application provides a communication device. The communication device comprises a processor and a memory, wherein the memory is configured to store program instructions, which, when executed by the processor, implement any of the above methods.

[0010] In another aspect, the present application provides a readable storage medium configured to store program instructions, which, when executed by a processor, implement any of the above methods. BRIEF DESCRIPTION OF DRAWINGS

[0011] The accompanying drawings, which are included to provide a further understanding of the present application, constitute a part of the specification and illustrate the illustrative embodiments of the present application and together with the description serve to explain the present application. In the drawings:

[0012] FIG. 1 shows a structural schematic diagram of an AI and communication converged system.

[0013] FIG. 2 is a flowchart of an AI correlation data reporting method according to an embodiment of the present application.

[0014] FIG. 3 is a flowchart of an AI correlation data reporting method according to another embodiment of the present application.

[0015] FIG. 4 is a flowchart of an AI correlation data reporting method according to another embodiment of the present application.

[0016] FIG. 5 is a flowchart of an AI correlation data reporting method according to another embodiment of the present application.

[0017] FIG. 6 shows an air interface identifier between a base station and a UE in a correlation technology.

[0018] FIG. 7 is a flowchart of an AI correlation data reporting method according to another embodiment of the present application.

[0019] FIG. 8 is a flowchart of an AI correlation data reporting method according to another embodiment of the present application.

[0020] FIG. 9 is a flowchart of an AI correlation data reporting method according to another embodiment of the present application.

[0021] FIG. 10 is a flowchart of an AI correlation data reporting method according to another embodiment of the present application.

[0022] FIG. 11 is a structural schematic diagram of a communication device according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative work fall within the scope of the present application.

[0024] It should be understood that the term "and / or" herein only describes the association relationship of the associated objects, which means that there can be three relationships, for example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after it.

[0025] The radio resource control (RRC) state of a user equipment (UE) can affect the current data collection process and the quality of collected data. In the traditional data collection mechanism of the NR air interface, such as the CSI framework or the RRM framework, the UE receives a measurement configuration, performs a measurement process, and reports the measurement result to the network side when in the RRC connected state. The UE measurement is instantaneous, that is, each time a new measurement process is performed, the new measurement result will overwrite the original measurement result, and there is no need to save it on the UE side. In the current network side data collection mechanism, the data collection method of the current RRM or CSI architecture is still used, but considering the large amount of AI model training data and low delay, the UE side data logging is supported in the traditional data collection mechanism for artificial intelligence / machine learning (AI / ML) associated data.

[0026] The existing scheme can solve the air interface overhead caused by frequent reporting of measurement data by UE by designing periodic reporting and event-triggered reporting, such as designing a long reporting period interval or defining an event-triggered reporting condition based on a data volume threshold. However, when the RRC state of the UE changes, if the current technology is followed, the stored measurement data will be released, causing waste of data.

[0027] FIG. 2 is a flowchart of an AI associated data reporting method according to an embodiment of the present application. As shown in FIG. 2, the method can include operations S101-S102.

[0028] In S101, the UE stores artificial intelligence model (AI) associated data in the RRC non-connected state.

[0029] The RRC non-connected state can include an RRC idle (RRC_idle) or an RRC inactive (RRC_inactive) state.

[0030] The AI-associated data is data related to AI functions or AI models / services. For example, the AI-associated data can include: measurement data such as reference signal received power (RSRP), reference signal received quality (RSRQ), signal to interference plus noise ratio (SINR), L1-RSRP, L1-RSRQ, L1-SINR, mean of more than two L1-RSRP, mean of more than two L1-RSRQ, mean of more than two L1-SINR, cell ID, beam ID, maximum number of cells, maximum number of beams, precoding matrix indicator (PMI), channel H (ground truth), UE motion trajectory, pattern information, listening accuracy, RSRP difference, mean square error / root mean square error (MSE / RMSE), complexity (e.g., floating-point operations per second), handover probability, handover success rate, etc.

[0031] The UE can start measuring and / or storing the AI-associated data after entering the RRC non-connected state. Alternatively, the UE can start measuring and / or storing the AI-associated data before entering the RRC non-connected state, i.e., while still in the RRC connected state, and continue measuring and / or storing the AI-associated data after entering the RRC non-connected state. The present embodiment mainly focuses on the part of data measured by the UE and failed to be successfully transmitted before the UE enters the non-connected state.

[0032] In S102, after the UE enters the RRC connected state, the UE reports the AI-associated data to the first network-side entity.

[0033] The first network-side entity is a network-side entity to which the UE establishes a connection after entering (or re-entering) the RRC connected state. The network-side entity can be a base station (e.g., gNB), a mobility management entity, a gateway device, a 5G core network device, etc.

[0034] After the UE establishes a connection with the first network-side entity after entering the RRC connected state, the UE reports the data stored by the UE in the RRC non-connected state to the first network-side entity.

[0035] Unlike the prior art, in the present embodiment, the UE continues to store the AI-associated data in the RRC non-connected state without releasing this part of data. Then, after the UE enters the RRC connected state, the UE reports this part of stored AI-associated data to the first network-side device with which the UE establishes a connection. In this way, the loss of AI-associated data when the UE switches RRC states can be avoided or reduced, so that the network side can collect more AI-associated data stored by the UE.

[0036] The behavior of the UE in reporting the AI-associated data after entering the RRC connected state can be pre-defined by the protocol. Alternatively, the UE can also receive a configuration from the network-side device regarding the reporting of the AI-associated data, and report the AI-associated data according to the configuration. For example, the UE can receive a storage indication message from the network-side device. The specific description is as follows.

[0037] In some embodiments, the method shown in FIG. 2 further includes operation S103a or S103b. Wherein, S103a or S103b occurs when the UE is in the RRC connected state.

[0038] In S103a, the first network-side entity sends a storage indication message to the UE, and the UE obtains the storage indication message from the first network-side entity. In S103b, the second network-side entity sends a storage indication message to the UE, and the UE obtains the storage indication message from the second network-side entity.

[0039] As described above, the UE can report the stored AI-associated data to the first network-side device after re-entering the RRC connected state. In this embodiment, the UE also obtains the corresponding configuration before storing these AI-associated data. If the configuration comes from the first network-side device, operation S103a is performed; if the configuration comes from the second network-side device different from the first network-side device, operation S103b is performed. It should be understood that in some cases, the UE can also receive a configuration message regarding the storage of AI-associated data from both the first network-side device and the second network-side device.

[0040] In some embodiments, the storage indication message is used to indicate whether the UE needs to store the AI-associated data across the RRC connected state (i.e., from the RRC connected state to the RRC non-connected state). If the storage indication message indicates that the UE needs to store the AI-associated data across the RRC connected state, the UE can accordingly continue to store the measured or obtained AI-associated data when entering the RRC non-connected state. Conversely, if the storage indication message indicates that the UE does not need to store the AI-associated data across the RRC connected state, the UE can accordingly release or not continue to store the obtained AI-associated data when entering the RRC non-connected state.

[0041] In other embodiments, the storage indication message is used to indicate whether the UE needs to keep storing the AI-associated data. For example, the storage indication message can indicate that the UE needs to keep the obtained AI-associated data until receiving a message that the AI-associated data can be released or the reporting of the AI-associated data is completed. Alternatively, the storage indication message can indicate that the UE needs to keep the obtained AI-associated data for a specified time.

[0042] The storage indication message can be realized by display identification. For example, the storage indication message can include a storage indication identifier. The storage indication identifier is used to indicate that the UE needs to temporarily store measurement data and / or support temporary storage of measurement data across RRC states.

[0043] The storage indication information can be realized by implicit indication. In one way, the storage indication message is directly indicated by the RRC message identifier. That is, the RRC message identifier includes any one of the following: measurement configuration ID, measurement object, measurement ID, reporting configuration ID, reporting content, etc. For example, when the UE receives a measurement configuration message, the measurement configuration message is specifically used for data collection of AIML; in addition, the message can also be used to indicate that the UE can support storage of measurement data or storage of data across RRC states, and then the UE knows that it needs to store measurement data or temporarily store data across RRC states after receiving the configuration message. In another way, one or more information carried in the RRC message or multicast message is used for implicit indication, such as function ID, model ID, data set ID, network side other conditions or association ID information. For example, when the UE receives an RRC message, the message carries indication information related to management functions, such as function ID, model ID, data set ID, network side other conditions or association ID information, or any combination of the above information, and then the UE knows that it needs to store measurement data or temporarily store data across RRC states after receiving the configuration message; in this case, the above information can also be used to indicate that the UE can support temporary storage of data across RRC states.

[0044] In some embodiments, the storage indication message is used to indicate the AI-associated data that the UE needs to store. Alternatively, the storage indication message includes at least one of the following: a measurement configuration identifier corresponding to the AI-associated data; measurement content corresponding to the AI-associated data; a reporting configuration identifier corresponding to the AI-associated data; reporting content corresponding to the AI-associated data; a function identifier corresponding to the AI-associated data; a model identifier corresponding to the AI-associated data; or a data set identifier corresponding to the AI-associated data. In this way, the UE can determine the range of AI-associated data that needs to be stored according to the storage indication message.

[0045] In some embodiments, the storage indication message can be associated with one of the following: a given artificial intelligence (AI) function; a given area; a given cell; or a given AI model. Further, the storage indication message comprises a first indication message and a second indication message; the first indication message is used to indicate the given AI function or the given AI model; the second indication message is used to indicate the given area or the given cell. Wherein, the first indication message is transmitted by unicast, and the second indication message is transmitted by groupcast. In other words, the second indication message can be sent to multiple UEs by groupcast, thereby indicating to these UEs that AI-associated data associated with the given area or the given cell needs to be stored; while the first message is sent to a specific UE by unicast, thereby indicating to the UE that AI-associated data associated with the given AI function or the given AI model needs to be stored. The first indication message and the second indication message can have a corresponding association identifier, so that the UE receiving the corresponding first indication message and the second indication message can know that the two messages are associated, and determine the range of AI-associated data that needs to be stored according to the two messages. Specifically, the association identifier can be a dedicated association ID, or a local index. In the latter case, the UE can determine whether the first indication message and the second indication message belong to the same storage indication message according to the value of the local index in the two messages.

[0046] The first indication message sent by unicast can also indicate measurement-related configurations, such as one of the following: measurement configuration, measurement object, measurement ID, measurement result (i.e., what measurement result needs to be reported, such as RSRP), or reporting configuration. The first indication message can be similar to the configuration of the traditional measurement data collection process, or additionally carry part of the information of the storage indication message on the basis thereof. Accordingly, the second indication message sent by groupcast can include another part of the storage indication message, such as area information, cell information, and storage indication identifier. Wherein, the groupcast message can be used to configure the conditions of a model with better generalization.

[0047] In different scenarios of data collection, the data types of the storage indication identifier are different, which can be understood as the granularity of the storage indication identifier is different, and accordingly, the life cycle of the storage indication identifier is also different. The following explains the different granularities of the storage indication identifier.

[0048] • The storage indication is associated with a given AI function (per AI function)

[0049] In some scenarios, due to the large amount of data collected, since one RRC message can carry at most 9KB of data, the UE needs to send multiple RRC messages to complete the configured data collection requirements. To avoid multiple configuration of storage indication identifiers, reduce the burden of UE managing storage indication identifiers, and reduce air interface signaling overhead, the storage indication identifier can be based on an AI function. The AI function can understand data collection for model listening, data collection for model training, model updating, model fine-tuning, model selection, model switching, model inference data collection, etc. The storage indication identifier based on the AI function can be understood as completing one complete AI function related data collection. In particular, not limited to the RRC state change of the UE. For example, if model fine-tuning requires collecting 50KB of data from the UE, and the UE reports 30KB in the first RRC connected state and enters the idle state, the UE can continue to report the remaining data in the second RRC connected state until all the data is reported.

[0050] • Storage indication is associated with a given area (per area)

[0051] The storage indication identifier associated with a given area can be understood as a unique storage indication identifier within a certain area. UEs entering this area can be assigned the same storage indication identifier, and the network side can collect data from at least one UE based on the storage indication identifier. This is because some models have high generalization requirements. To reduce overhead, the storage indication identifier based on the area can be sent to multiple UEs in the form of a multicast or broadcast message. The multicast message carries area information, and when the UE collects this information, it knows that the network allows it to store measurement data, store measurement data across RRC states, and / or report measurement data. When the UE moves out of the specified area, the storage indication identifier is invalid. The area information can be the actual geographic location, such as longitude, latitude, and altitude, or the area information can consist of one or more cells. In this case, the storage indication identifier can be decoupled from the traditional data collection process.

[0052] In some examples, the storage indication identifier can indicate data collection, i.e., the UE receives the storage indication identifier indicating that the network needs to collect data from the UE receiving the identifier, i.e., the data collection configuration message is sent in a multicast manner. Accordingly, the reporting is based on the received data collection configuration message.

[0053] • Storage indication is associated with a given cell (per cell)

[0054] The storage indication identifier is based on a given cell and can understand that the storage indication identifier is unique in a region. When a UE enters the region, the UE can be assigned the same storage indication identifier. Based on the storage indication identifier, the network side can collect data from at least one UE. This is because some models have high generalization requirements. To reduce overhead, the region-based storage indication identifier can be sent in the form of a multicast and broadcast message to multiple UEs. The cell information is carried in the multicast message. When the UE collects the information, it knows that the network allows it to store measurement data, store measurement data across RRC states, and / or report measurement data. When the UE moves out of the set cell, the storage indication identifier is invalid. The cell information can consist of one or more cells. In this case, the storage indication identifier is decoupled from the traditional data collection process.

[0055] In some examples, the storage indication identifier can represent data collection. When the UE receives the storage indication identifier, it indicates that the network needs to collect data from the UE that receives the identifier, i.e., the data collection configuration message is sent through multicast. Accordingly, the reporting is based on the received data collection configuration message.

[0056] The storage indication is associated with a given AI model (per model)

[0057] The storage indication identifier is based on a given AI model and can understand that the storage indication identifier is unique under an AI model. In this scheme, the storage indication ID and the data collection configuration are independent. As long as the measured data is used for the model, the measured data needs to be stored. When the model-based data collection process is completed, the storage indication identifier is invalid. In particular, in some scenarios, the model identifier can be used instead of the storage indication identifier.

[0058] It can be understood that the granularity of the storage indication identifier can also be a combination of the above-mentioned different granularities. In addition, the storage indication identifier can be unique to the UE, the base station, the AMF, the OAM, the PLMN, the tracking area (TA), or the RNA region.

[0059] In some embodiments, the storage indication identifier can be indicated by 1 bit. For example, the representation of the data collection indication identifier can be absent or present. When the indication identifier is set as absent (i.e. the identifier is not present), it means that the network side does not send the indication identifier, and the UE does not need to store the measurement data across RRC states; when the indication identifier is set as present, it means that the network side sends the indication identifier, and the UE needs to store the measurement data across RRC states. Alternatively, the representation of the data collection indication identifier can be true or false. When the indication identifier is set as true, it means that the network side allows the UE to store the measurement data across RRC states; when the indication identifier is set as false, it means that the UE does not need to store the measurement data across RRC states. In this case, the network side always needs to send the indication identifier.

[0060] In some embodiments, the UE is configured with a plurality of data measurement and / or storage configurations corresponding to the storage indication identifiers; the storage indication message includes a target storage indication identifier; and the AI-associated data is determined according to the target storage indication identifier and the data measurement and / or storage configurations corresponding to the plurality of storage indication identifiers. In other words, the UE can determine a plurality of potential data measurement and / or storage configurations according to the protocol predefinition or pre-received related configurations. After receiving the target storage indication identifier, the UE can determine the currently effective or indicated configuration from the plurality of configurations according to the identifier, so as to perform the storage and / or reporting operation of the AI-associated data according to the selected configuration. Specifically, the storage indication identifier can be indicated by a string or enumeration, for example, 0001, 0010 and 0011 can respectively represent different storage indication identifiers.

[0061] In some embodiments, the storage indication message can include the storage indication identifier and other associated information. Specifically, the storage indication identifier and the related information can be represented by designing the structure of the data storage indication message. For example, the structure can include one or more of the following information: data collection configuration information, measurement content, measurement ID, data set ID, AI model ID, region information and cell information, etc. Tables 1 and 2 respectively show two different structure examples.

[0062] Table 1

[0063] Table 2

[0064] In some scenarios, the storage indication message is accompanied by the data collection procedure, i.e., the storage indication message is equal to the data collection configuration message; or the storage indication identifier is included in the data collection configuration and sent to the UE together with the data collection configuration message. Alternatively, in some scenarios, the storage indication message configuration procedure and the data collection procedure are two procedures, such as the storage indication message is sent to the UE through multicast / broadcast, and the data collection procedure is still performed based on the RRC pair message (RRC reconfiguration / RRC release and measurement reporting message, UEInformationRequest / UEInformationResponse, etc.).

[0065] FIG. 3 is a flowchart of an AI-associated data reporting method according to another embodiment of the present application. As shown in FIG. 3, the method can include operations S201 to S206. The specific process will be described below, and operations that are the same or similar to those shown in FIG. 2 will not be described again. In short, the UE and the network side perform data collection and data storage in the RRC connected state, and when the RRC state changes, such as entering the RRC idle or RRC inactive state, the UE needs to support the storage of measurement data; when entering the RRC connected state again, the UE reports or releases the stored data according to the designed reporting rules or methods.

[0066] In S201a, the UE receives a storage indication message from a network side device (NW). The storage indication message can include but is not limited to a storage indication identifier, a timer Timer N3XX, cell information, area information, a model ID, an AI function ID, a data set ID, etc. The configuration information can be sent to the UE in the form of unicast, multicast, or broadcast.

[0067] In S201b, the UE receives data collection configuration information from the NW. The configuration information can include but is not limited to measurement configuration, reporting configuration, measurement content, model ID, reporting type, etc. The configuration information can be sent to the UE in the form of unicast or multicast.

[0068] It should be understood that the foregoing storage indication message and data collection configuration information can be carried in the same message / configuration signaling or in two independent messages / configuration signaling, and the present application does not limit this.

[0069] It should be understood that the network side device (NW) in the present embodiment can include the first network side device and / or the second network side device shown in FIG. 2. For example, if the base station connected before and after the UE performs RRC connection state switching is the same, the number of network side devices is one; if the base station connected before and after the UE performs RRC connection state switching is different, the number of network side devices is two.

[0070] In some embodiments, before the network side configures the UE with the storage indication message, the UE can inform the network side that it has the capability of supporting the storage of measurement data and / or the storage of measurement data across RRC states. The capability information can be core network capability information or wireless capability information. For example, when the UE registers with the network, it sends a registration request message, in which it reports to the core network its capability of supporting the storage of measurement data and / or the storage of measurement data across RRC states. Alternatively, the base station can send a capability query request message to the UE, which contains the capability of whether the UE stores measurement data and / or stores measurement data across RRC states. If the UE supports it, the UE will inform the base station in the capability reporting message.

[0071] The activation of the storage indication message can have several cases, including but not limited to: 1) When the storage indication message is sent to the UE, it is activated by default; this case is usually when the storage indication message is sent during data transmission, or when the storage indication message is sent in the storage indication message; 2) When the UE receives the sent data collection information and / or produces measurement data, the storage indication message is automatically activated; 3) The network side sends an activation message such as MAC CE, DCI to activate the storage indication message. In addition, there is also a case where the UE does not have a storage indication message, or the storage indication message is implicitly represented by other messages or elements, but the UE supports the storage function (across RRC), at this time, the network side can also directly send an activation message to activate the storage function of the UE.

[0072] In S202, the UE performs measurement and stores the AI-associated data of the measurement. The UE stores the configuration information related to the storage of measurement data across RRC states, and performs the measurement process and stores the measurement data according to the indication.

[0073] In S203, the UE evaluates the reporting condition. The UE can determine the reporting condition according to the measurement data configuration, for example, the reporting condition can include periodic reporting and event-triggered reporting. This operation is optional.

[0074] In S204, the UE reports the stored measurement data. For the case where there is a reporting condition, if the measurement data does not meet the condition, the UE does not report the stored measurement data; if the measurement data meets the condition, the UE reports the stored measurement data. The UE can send multiple RRC messages. It is particularly noted that the UE sends the stored measurement data, but it is also possible that after the sending is not completed, the UE enters the RRC idle or inactive state, at this time, the UE still has a part of the measurement data that has not been successfully sent.

[0075] In S205, the UE continues to store the unreported measurement data. The UE experiences an RRC state transition from RRC connected state to RRC idle state or to RRC inactive state. According to the information configured in step S201, the UE supports storing the measurement data generated in step S202 in RRC non-connected state, including the measurement data that is not successfully reported in step S204.

[0076] In S206, the UE reports the measurement data stored in RRC non-connected state. When the UE experiences an RRC state transition again from RRC non-connected state to RRC connected state. The UE performs the data reporting procedure across RRC states. The reporting scheme can be a UE-based scheme or a network-side entity-based scheme. The data reporting message can include but is not limited to: storage indication identifier, cell information, area information, model identifier, data validity indication, data reporting completion indication (end marker).

[0077] The UE-based scheme can include the following ways:

[0078] Way one, the UE assesses whether the timer related to data validity has expired. If the timer has expired, the UE releases the stored data. If the timer has not expired, the UE mainly initiates the data storage reporting procedure.

[0079] In some embodiments, when the UE enters RRC connected state, it is determined whether the validity timer of the AI-associated data has expired. If the validity timer has not expired, the UE can report the AI-associated data. Otherwise, if the validity timer has expired, the UE releases the AI-associated data and / or reports the invalidity information of the AI-associated data to the first network-side entity. For example, if the value of the validity timer is set to 10 days, if the data does not exceed 10 days, the UE will report the data when it enters RRC state again. If the data exceeds 10 days, the data will be released by the UE.

[0080] Specifically, a validity timer Timer N3XX can be defined. The Timer N3XX is used to indicate the validity of the data. When the UE experiences an RRC state transition, the stored data is still within the defined Timer N3XX interval, the Timer N3XX has not expired, indicating that the stored measurement data is still valid, and therefore the data can be reported to the network side. The value is a constant, and the value unit can be a symbol, a subframe, a slot, a frame, a millisecond, a second, a minute, an hour, a day, a week, etc. The Timer N3XX can be represented in an enumerated form, for example, its value can be set to ENUMERATED{min5, min10, min15, min30}. When the Timer N3XX is set to min5, it means that the Timer N3XX time is 5 minutes.

[0081] The condition for starting the validity timer can be one of the following: receiving a data collection configuration message; the UE starting to perform measurement or storage of the AI-associated data; or receiving an activation message for measurement of the AI-associated data.

[0082] The condition for stopping the validity timer can be one of the following: the UE selecting to another cell; the UE selecting to a cell within a specific area; receiving a message that the AI-associated data has been collected completely; or the AI-associated data has been sent completely. In addition, the validity timer can also be stopped if a time threshold of data collection definition is exceeded or a data reporting request message sent by the NW cannot be received for a long time.

[0083] When the timer expires, the UE can actively release the recorded stored measurement data, indicating that the stored measurement data is invalid. In addition, if the storage indication identifier is configured, the UE deletes the storage indication identifier. Alternatively, the UE can also report to the network side device that the stored AI-associated data has become invalid and useless data.

[0084] The configuration message of the aforementioned validity timer Timer N3XX can be informed to the UE in the following ways: sent to the UE through a data collection configuration message, sent to the UE through a broadcast message, or sent to the UE through a multicast message. Among them, this configuration message can be periodically sent to the UE, or non-periodically (i.e. in response to a predetermined condition being met) sent to the UE.

[0085] In addition, an indication message about data validity can also be introduced. If the UE-stored measurement data does not exceed the value set by the timer Timer N3XX, the indication message indicates validity; otherwise, the indication message indicates invalidity. The validity indication message can also be included in the data reporting message to inform the network side.

[0086] Method two, the UE evaluates whether the currently accessed destination base station is included in the cell information and / or area information by defining cell information (cell white list) and / or area information, indicating that the stored measurement data is still valid. If included, the UE mainly initiates the process of storing data reporting; if not included, the UE releases the stored data. This scheme can be used in the scenario of UE handover.

[0087] In some embodiments, when the UE enters the RRC connected state, it is determined whether the AI-associated data belongs to a specified range; and in response to the AI-associated data belonging to the specified range, the AI-associated data is reported. Wherein, the determination of whether the AI-associated data belongs to the specified range includes: receiving an indication from a second network side entity whether the AI-associated data belongs to the specified range; or receiving an indication from the first network side entity whether the AI-associated data belongs to the specified range. Alternatively, the determination of whether the AI-associated data belongs to the specified range includes: obtaining an RRC reconfiguration message from a second network side entity; and determining whether the AI-associated data belongs to the specified range according to the area or cell information of the first network side entity in the RRC reconfiguration message.

[0088] This way is suitable for the case where the UE obtains the AI-associated data that needs to be stored in the corresponding range by storing the indication message or other pre-configured ways. The range can be a cell or an area.

[0089] As described above, in some cases, the storage indication message can include cell information or area information. Wherein, the cell information can be composed of one or more cells. If the cell information contains at least one cell, it can be represented by one cell group, which contains multiple cell representations; it can also be represented by a cell list, which contains multiple cell identifiers. In particular, in some cases, the concept of a cell is equivalent to a base station. The area information can be 1) composed of one or more cells, similar to cell information; 2) actual geographic information, which contains one or more of longitude, latitude, and altitude; 3) a combination of cell information and actual geographic location information.

[0090] The cell information can be allocated by the core network or by the base station. The area information can be determined by a model training entity. For example, if the model training entity is a third-party server, the area information is specified by the third-party server and notified to the core network element, or to the base station, or to the core network element, which is then notified to the base station. The core network and / or the base station will notify the UE of any of the above information or combinations of information. Alternatively, if the training model entity is managed by a data plane or an intelligent plane, the area information can be specified by the data plane or the intelligent plane. The data plane and the intelligent plane are logical functional entities that process AI data / traffic / management control / security.

[0091] In some other cases, the cell information or the area information can be configured separately. For example, the base station can inform the UE of the cell information and / or the area information through a broadcast, unicast or groupcast message; or the core network can inform the base station of the cell information and / or the area information through an N2 message, and then the base station informs the UE; or the core network can inform the base station of the cell information and / or the area information through an NAS message, and then the base station informs the UE, which can be carried in an RRC unicast message. In addition, a third-party server or other functional entity can package the area information into a payload and send it to the core network, which carries the message in an NAS or N2 message and sends it to the UE through the base station. Based on the received cell information and area information, the UE can perform a signaling interaction process of data reporting.

[0092] The UE evaluates whether the currently accessed target base station is included in the cell information and / or the area information, and if so, it indicates that the stored measurement data is still valid. Therefore, if the accessed target base station is included in the cell information and / or the area information, the UE mainly initiates the process of stored data reporting; if the accessed target base station is not included in the cell information and / or the area information, the UE releases the stored data. In the data reporting message reported by the UE, one or more of the following can be included: AI associated data, cell information, area information, data reporting completion indication, etc.

[0093] This scheme can be used in the scenario where the UE performs handover. In particular, if handover occurs, the UE can determine whether the target base station is within the set cell information and / or area information by the information of the target base station in the RRC reconfiguration message sent by the original base station. In addition, the original base station can also determine whether the target base station is within the set cell information and / or area information, which can be determined by carrying the cell information and / or area information in the handover request and reply process of the Xn interface. There are two potential processes as follows:

[0094] Process 1: The original base station (such as the second network side device in FIG. 2) carries the cell information or the area information in the handover request message, which is sent to the target base station; the target base station (such as the first network side device in FIG. 2) determines whether the UE stored data is valid, and if the target base station is not within the set cell information and / or area information, an invalid indication can be returned; otherwise, a valid indication is returned.

[0095] Process 2: The original base station carries the UE stored data valid indication information in the handover request message, which is sent to the target base station, and the target base station determines whether the UE stored data is valid, and if the target base station is not within the set cell information and / or area information, an invalid indication can be returned; otherwise, a valid indication is returned.

[0096] If the target base station is not in the set cell information and / or area information, the original base station can inform the UE in the RRC reconfiguration message that the stored measurement data is invalid for the target base station to which the UE is preparing to access; otherwise, the UE can continue to store and report the stored data at an appropriate time.

[0097] In addition, an indication message about data validity can also be introduced. If the measurement data stored by the UE does not exceed the defined cell information or area information, the indication message about validity indicates validity; otherwise, the indication message about validity indicates invalidity. The indication message about validity can also be included in the data reporting message to inform the network side.

[0098] In the third mode, when the UE enters the RRC connected state again, the UE sends a data reporting request message to the network side. If the network side agrees to the UE reporting, the UE can report the stored measurement data; if the network side does not agree to the UE reporting, the UE releases the stored data; or if the UE does not receive information from the network within T time, the UE releases the stored data; the request message can carry 1 bit of request information or buffer size information of the stored measurement data. This scheme can be used in a scenario in which the network side collects data from multiple UEs.

[0099] In some embodiments, when the UE enters the RRC connected state, a data reporting request message is sent to the first network side entity; an agreement message of the data reporting request message is received from the first network side entity; and in response to the agreement message, the AI-associated data is reported.

[0100] The reporting request message can include at least one of the following: a storage indication identifier corresponding to the AI-associated data; cell information corresponding to the AI-associated data; area information corresponding to the AI-associated data; or AI model information corresponding to the AI-associated data. In addition, the reporting request message can also include a buffer size corresponding to the AI-associated data.

[0101] As shown in FIG. 4, the present mode can include operations S301 to S303.

[0102] In S301, when the UE enters the RRC connected state from the RRC non-connected state, a data reporting request message can be sent to the network side device. Specifically, when the UE enters the RRC connected state, a data reporting request message is actively sent. The data reporting request message can be a PHY message, a MAC CE message, or an RRC message, which indicates or includes a 1-bit indication field to inform the network that there is unreported data to be reported to the network side.

[0103] Further, in order to indicate the stored unreported data is valid, the reporting request message can further comprise any one or combination of the following information: storage indication identifier, cell information, area information, model information, etc.; optionally, the UE can also inform the network of the buffer size information of the unreported measurement data.

[0104] In S302, the UE receives a data reporting request reply message from the network side device.

[0105] After receiving the data reporting request message, the NW can send a corresponding data request reply message to the UE, if agreeing to report, a data reporting request acceptance message is sent; if not agreeing to report, a data reporting request rejection message is sent; optionally, the reporting related resource information can be sent to the UE together with the data reporting request acceptance message, or can be sent to the UE separately through other messages.

[0106] In S303, the UE determines whether to perform data reporting according to the reporting reply message.

[0107] Based on the received reply, the UE determines whether to report the stored measurement data, if the network side agrees, the UE performs data reporting process, and the reporting message carries the stored measurement data; if the network side does not agree, the UE does not report the stored measurement data. Optionally, the stored measurement data can be directly released or continue to be stored. Optionally, the data reporting message carries a data reporting completion identifier (end marker).

[0108] In this scheme, when the UE enters the RRC connected state again, if it is found that there is still stored measurement data in its memory, the UE can not know whether the data is still available, then the UE can send a data reporting request message to the network, and based on the reply message from the network side, it is determined whether to report the stored measurement data or release the stored measurement data. In particular, the UE can autonomously determine the timing of sending the request message.

[0109] Method four, when the UE enters the RRC connected state again, it does not need any evaluation, and directly initiates the reporting of the stored measurement data, which can be understood as a kind of predefined rule or configured rule.

[0110] In this scheme, when the UE enters the RRC connected state again, the UE detects its memory content and determines whether the previous measurement data that was not reported in the RRC idle state is still stored in the memory. If the UE finds that the previous measurement data is still stored in the memory, the UE directly performs the data reporting process. This scheme is applicable to two cases: 1) the UE directly performing the data reporting process is based on the implementation behavior of the UE; and 2) the UE directly performing the data reporting process is based on a predefined default rule, which is described as follows: “When the UE enters the RRC connected state, if the previous measurement data is still stored in the memory, the UE directly reports the stored measurement data”. Optionally, the data reporting message carries an end marker.

[0111] The network-side entity-based scheme can include the following manner: the NW actively sends a data collection request message, and the UE reports the stored measurement data based on the content of the data request message.

[0112] In some embodiments, the UE receives a data collection request message from the first network-side entity; and in response to the data collection request message, reports the AI-associated data. As shown in FIG. 4, this scheme can include operations S401 to S403.

[0113] In S401, the UE receives a data reporting request message from the NW. The message contains at least one of the following: a storage indication identifier, a UE identifier, cell information, area information, and the like.

[0114] In S402, the UE compares the stored information with the information carried in the request message. That is, the UE detects its memory content and compares the stored information with the information carried in the request message.

[0115] In S403, the UE reports data to the NW based on the comparison result. If the comparison result is consistent, the UE performs the data reporting process, and the reporting message carries data. Optionally, the data reporting message also carries an end marker. Otherwise, the UE does not perform the data reporting process.

[0116] In this scheme, since the data collection is initiated by the network side, the NW knows whether the collected data meets the data collection rules set by the NW. Therefore, if the collected data is insufficient, the NW can actively initiate a data reporting request message to the UE.

[0117] In one case, if the NW knows the identity of the UE, which does not change with the RRC state change (such as IMSI), the NW can actively send a data reporting request message to the specific UE, i.e. the data collection request message is transmitted by unicast or groupcast. This case can be applied to the UE-specific model, the cell-specific model, the area-specific model, etc.

[0118] Correspondingly, the application also provides a temporary identifier of the UE, which is used for reporting the AI-associated data. Correspondingly, the method shown in FIG. 3 can further include: obtaining the temporary identifier; and releasing the temporary identifier when the AI-associated data reporting is completed or invalid.

[0119] In another case, if the NW does not know the identity of the UE, but knows the cell information and / or the area information. The NW can actively send a data reporting request message to the UE in the area or the cell, the data collection request message is transmitted by groupcast, and includes the cell or area information corresponding to the AI-associated data. This case is applicable to the cell-specific model, the area-specific model, etc.

[0120] The temporary identifier of the UE provided by the application will be described below.

[0121] On the air interface between the base station and the UE, a radio network temporary identifier (RNTI) is commonly used to identify the UE. When the UE accesses the network, the base station allocates a cell radio network temporary identifier (C-RNTI) to the UE, which is used for the base station to identify the UE. In addition, the RNTI can also be used to scramble the cyclic redundancy check (CRC). After the UE enters the connected state, the MAC layer of the base station allocates the C-RNTI to the UE, which has a fixed length of 16 bits, and has a mapping relationship with the logical channel. The logical channel identifier (LCID) corresponding to the C-RNTI is 53, and its structure is shown in FIG. 6. However, since the identification on the air interface of the UE, such as the C-RNTI, is temporary, each time the UE disconnects from the network and then connects to the network again, the allocated C-RNTI is different. Therefore, in the above-mentioned cross-RRC state data collection process, the base station cannot find the same UE at different RRC connections through the existing air interface identification C-RNTI. Therefore, the NW cannot identify the data collected from which UE, and cannot complete the related functions of the UE-level model, such as data collection for model training. To solve the above problem, in the present embodiment, a semi-permanent identifier of the UE on the air interface is mainly designed to ensure that the network side can still identify the same UE across the RRC state, and continuously collect data from the UE, thereby improving the quality of the model function. In addition, the length of the identifier is small, which can further reduce the air interface overhead.

[0122] The UE temporary identifier provided in the present application can also be referred to as a "semi-permanent identifier" of the UE. The semi-permanent identifier can be used to identify the UE for data collection between the UE and the network. The semi-permanent identifier is allocated by the base station and has a time limit, is valid in some scenarios, changes in state, does not change with the RRC state, and is invalid after the completion of the relevant process.

[0123] The semi-permanent identifier has a small length, which can be a fixed length such as 2, 4 or 8 bits, or can be variable. The base station evaluates how many semi-permanent identifiers need to be allocated and how much length is needed according to the actual need to simultaneously complete data collection from how many UEs at the same time. If the NW needs to collect data from 100 UEs at the same time, then the semi-permanent identifier allocated only needs to occupy 5 bits to be distinguished. The semi-permanent identifier can be composed of a binary string.

[0124] The time limit of the semi-permanent identifier can be closely related to the data collection process. Specifically, the semi-permanent identifier is invalid when the data collection process is completed, or the semi-permanent identifier is invalid when the stored measurement data of the UE becomes useless. The semi-permanent identifier is invalidated by the base station, such as when the base station finds that the data collection is completed, or when the UE reports that the data is invalid, the base station can choose to release the identifier. Specifically, the NW sends an invalidation indication signaling to the UE, informing the UE that the previously allocated semi-permanent identifier is not available, and the signaling contains an invalidation indication field and a semi-permanent identifier, and the UE can release the stored semi-permanent identifier. Alternatively, if the base station evaluates that the quality of the measurement data collected by the UE is poor, it will also carry an indication about the UE measurement data, which indicates that the data is not available. When the UE still has stored data that has not been sent, according to the indication, the UE can choose to release the stored data. The above-mentioned signaling can be DCI, or MAC CE, or RRC.

[0125] The semi-permanent identifier and the storage indication identifier described above can have a one-to-many mapping relationship, a one-to-one mapping relationship, or a many-to-one mapping relationship.

[0126] The semi-permanent identifier can be sent to the UE by the base station through an RRC message. In some scenarios, the base station needs to collect data from the UE for AI-related functions, such as model training and model monitoring. However, the base station cannot guarantee that the same UE can be selected in different RRC states to ensure that the base station continuously collects data from the same UE. Therefore, the base station can define a semi-permanent identifier for the UE to perform the data collection process based on the semi-permanent identifier. In some scenarios, the OAM needs to collect data from the UE for AI-related functions, such as model training and model monitoring. However, the base station cannot guarantee that the same UE can be selected in different RRC states to ensure that the OAM continuously collects data from the same UE. Therefore, the base station can design a mapping relationship between the semi-permanent identifier of the UE and the network-side identifier (such as IMSI) to perform data collection based on the defined semi-permanent identifier of the UE over the air interface. In the above-mentioned data collection process, the semi-permanent identifier of the UE for the Uu air interface can be included in the data collection configuration message or in the data reporting message.

[0127] When the AI-associated data transmission is completed, the corresponding storage indicator identifier can be considered invalid. In addition, there are several other possible cases for the invalidation of the storage indicator identifier, which can be initiated by the UE to release the storage indicator identifier, or indicated by the network side to invalidate the storage indicator identifier.

[0128] For example, when the data collection process is completed, the storage indicator identifier is invalidated. Specifically, when the UE successfully transmits the last stored measurement data, the UE considers that the data collection process has been completed, and the UE considers that the identifier is invalid and can choose to release the storage identifier. Alternatively, when the NW collects sufficient data to complete the relevant AI function, the NW determines that the data collection process has been completed, and the NW sends an invalidation indication to the UE, which can be sent through unicast or groupcast. In some embodiments, in response to all of the AI-associated data being completely transmitted, the UE sends a transmission completion identifier of the AI-associated data to the first network-side entity. In this way, both the UE and the network-side entity can consider that the storage indicator identifier corresponding to the AI-associated data is invalid. The transmission completion identifier (end marker) is used to indicate that the transmitted data is the last data that needs to be uploaded by the UE, and the successful transmission of the data indicates that there is no more data information that needs to be reported by the UE. The network side determines whether to configure data collection configuration again based on the reporting completion indication.

[0129] In addition, if the collected data has not been reported within a defined threshold, it can be understood that the data has expired and become invalid. Assuming that the UE is required to complete data reporting within 10 days, but the UE has not had the opportunity to enter the connected state, in this case, the corresponding storage identifier is also invalidated.

[0130] Alternatively, the UE does not receive the data reporting request message sent by the NW for a long time, the NW has collected useful data from other UEs or other nodes, and the AI function execution is completed, so it is not necessary to send a data reporting request message from some UEs, and in this case, the corresponding storage identifier is also invalid. It should be noted that the aforementioned threshold can be a time-based threshold or a data volume-based threshold. The aforementioned "long time" can correspond to a predefined time length or a predefined transmission number. For example, if the UE receives multiple data collection configurations or multiple sets of data collection configuration messages sent by the network side, and none of them includes a request for measurement data of one or more storage identifiers, the UE can consider that the storage identifier is invalid.

[0131] FIG. 7 is a flowchart of an AI-associated data reporting method according to another embodiment of the present application. As shown in FIG. 7, the method includes operations S501 to S502.

[0132] In S501, the UE stores AI-associated data.

[0133] Similarly, the AI-associated data is data related to an AI function or an AI model / service.

[0134] In S502, when a preset condition is met, the UE reports the AI-associated data to a network side entity.

[0135] The present embodiment can not involve RRC state switching of the UE, for example, the present embodiment can be applied to a case where the UE is in an RRC connected state. By designing the corresponding conditions for data reporting, the UE can report the stored AI-associated data at an appropriate time, thereby achieving "lossless" reporting of this part of data to some extent.

[0136] When the preset condition is met, the UE reports the AI-associated data to the network side entity, which can avoid congestion caused by resource conflicts between these data and other control information, thereby improving system efficiency and reducing or avoiding loss of important data.

[0137] FIG. 8 illustrates an AI-associated data reporting method based on a preset condition according to another embodiment of the present application. As shown in FIG. 8, the method can include operations S601 to S604. Simply put, the UE and the network side perform data collection and data storage in the RRC connected state, and when the data reporting condition is met, the UE reports the stored measurement data in advance according to the designed reporting rules or ways.

[0138] In S601, the UE collects data collection configuration from the network side device. The data collection configuration is used to indicate the content of the AI associated data and / or the preset condition, for example, the data collection configuration can include the rule of lossless reporting, the condition of data reporting, the priority of data reporting condition, etc., in addition, the data collection configuration can also include measurement configuration, measurement object, storage indication information, etc. Optionally, the data collection configuration can also include the UE semi-persistent identifier described in the foregoing.

[0139] In S602, the UE performs measurement and stores the AI associated data.

[0140] In S603, the UE evaluates whether the condition of lossless reporting is met.

[0141] In S604, when the condition of reporting is met, the UE performs data reporting to the network side device. The content of reporting can include AI associated data, data validity indication, etc., and a transmission completion identifier can also be reported when the data transmission is completed.

[0142] In some embodiments, the preset data reporting condition is that the radio resource control (RRC) state of the UE is about to change from a connected state to a non-connected state. That is, before the UE enters the RRC idle state or the RRC inactive state from the RRC connected state, the UE can initiate a data reporting process. When the RRC state of the UE changes, the measurement data temporarily stored by the UE can be reported before the RRC state of the UE changes. In this way, the AI associated data can be avoided to be discarded when the RRC state of the UE changes, thereby improving the collection rate of the AI associated data.

[0143] In one way, the UE can determine whether the RRC state of the UE is about to change according to the quality of the current wireless link. Specifically, the UE can determine that the RRC state of the UE can change by detecting that the channel / link quality is poor through measurement methods, sensing methods, AI prediction, or other ways. For example, a threshold based on RSRP, RSRP, or SINR can be defined. In some cases, when the obtained data of the UE, such as RSRP, is less than the defined threshold of RSRP, or when the obtained data of the UE, such as RSRQ, is less than the defined threshold of RSRQ, or when the obtained data of the UE, such as SINR, is less than the defined threshold of SINR, the data reporting process can be initiated.

[0144] In some cases, the thresholds of RSRQ, RSRP, and SINR are the same as the handover threshold defined at the time of handover. It can be understood that when the measurement value obtained by the UE meets the condition configured at the time of handover, it can be known that the channel quality of the serving cell of the UE is not good, or the channel quality of the neighboring cell is better, and the UE can initiate handover, which can cause the RRC state of the UE to change, but whether the UE can finally hand over depends on the implementation of the base station. At this time, the UE can initiate the data reporting process.

[0145] In one embodiment, the UE can determine whether the RRC state of the UE is about to change according to whether the current radio link is about to fail. Specifically, when the UE monitors that the radio link monitoring is about to fail, the UE can initiate the data reporting procedure. The UE tracks the radio link condition through the radio link monitoring, and if the UE monitors that the radio link monitoring fails, it can determine that its RRC state can change. Specifically, the RRC layer receives N310 consecutive out-of-sync indications, or does not receive N311 consecutive in-sync indications within T310, or before T310 is about to expire, the UE knows that it can have a radio link failure. The about to expire can be understood as any time before the value defined by T310, for example, T310 is set to 10s, the about to expire can be any time before the 10s, which can be configured by the base station, for example, the base station configures a bias value of 2s, which means that the UE can report at the 8s; it can also be autonomously determined by the UE.

[0146] In one embodiment, the UE can determine whether the RRC state of the UE is about to change according to whether the current radio link is about to fail. Specifically, when the UE monitors that the radio link monitoring is about to fail, the UE can initiate the data reporting procedure. The UE tracks the radio link condition through the radio link monitoring, and if the UE monitors that the radio link monitoring fails, it can determine that its RRC state can change. Specifically, the RRC layer receives N310 consecutive out-of-sync indications, or does not receive N311 consecutive in-sync indications within T310, or before T310 is about to expire, the UE knows that it can have a radio link failure. The about to expire can be understood as any time before the value defined by T310, for example, T310 is set to 10s, the about to expire can be any time before the 10s, which can be configured by the base station, for example, the base station configures a bias value of 2s, which means that the UE can report at the 8s; it can also be autonomously determined by the UE.

[0147] In other embodiments, the preset data reporting condition can also be that the current power consumption of the UE is less than a preset threshold, the current memory of the UE reaches a preset threshold, or the validity timer of the AI-associated data expires. In these cases, whether the RRC state of the UE is about to change can not be considered.

[0148] For the case where the preset condition is that the current power consumption is less than a preset threshold, when the UE detects that the current power consumption is poor, such as below a power consumption threshold, the UE sends a data reporting request message to the network side. The power consumption threshold can be predefined or configured by the network, and the power consumption threshold can be the minimum power consumption threshold or the maximum power consumption threshold.

[0149] For the case that the preset condition is that the current memory reaches a preset threshold, such as the current memory resource of the UE is about to be full, such as less than a minimum defined memory threshold, the UE sends a data reporting request message to the network side. The memory threshold can be predefined or configured by the network, and the memory threshold can be a minimum memory threshold or a maximum memory threshold. Optionally, there can also be a memory threshold offset value, such as when the memory resource is less than the sum of the memory threshold and the memory threshold offset value, the UE sends a data reporting request message to the network side. Optionally, there can also be a time information, such as when the memory resource is less than the sum of the memory threshold and the memory threshold offset value, the UE sends a data reporting request message to the network side.

[0150] For the case that the preset condition is that the validity timer is expired, the UE can also perform the data reporting process based on the validity timer, for example, the UE can be configured with a validity timer Timer N3XX as described above, which is used to indicate the validity of the data. When the Timer N3XX is about to expire but has not expired, the data reporting process can be initiated.

[0151] It should be understood that the above-mentioned preset data reporting conditions can be used in combination. In addition, the above-mentioned conditions can be protocol predefined or obtained by network side configuration.

[0152] In some embodiments, the aforementioned reporting conditions of AI-related data can be multiple, or can be stored simultaneously with other data reporting conditions, for example, the reporting conditions of AI-related data can include a first condition and a second condition, and the reporting configuration (such as reporting content, reporting resource, etc.) corresponding to the first condition and the reporting configuration corresponding to the second condition can be different. In this case, the priority of the first condition and the second condition can be set.

[0153] For example, the priority of the first condition can be set to be higher than the second condition. In this case, if the first condition and the second condition are satisfied at the same time, the data reporting can be performed according to the reporting configuration corresponding to the first condition with higher priority.

[0154] The priority of each reporting condition can be predefined. Alternatively, the priority can be configured by the network side, in particular, in some examples, the priority can be changed.

[0155] FIG. 9 is a flow diagram of an AI-related data reporting method according to another embodiment of the present application. In this embodiment, the data reporting request can be initiated by the UE. As shown in FIG. 9, the method includes operations S701-S706.

[0156] In S701, the UE receives the configuration of data collection from the network side device.

[0157] In S702, the UE performs measurement and stores AI associated data.

[0158] In S703, the UE evaluates the condition for lossless reporting (optional).

[0159] The condition for the UE to initiate the reporting request message can be the same as the condition for initiating lossless reporting described above. Alternatively, the condition for the UE to initiate the reporting request message can be different from the condition for the UE to initiate data reporting, and both can exist at the same time.

[0160] In S704, the UE sends a reporting request message.

[0161] The reporting request message described above can be used to inform the network that it needs to urgently report the stored measurement data.

[0162] In one way, the reporting request message is a dedicated indication message that informs the network that it needs to urgently report the stored measurement data. The message can be a UCI dedicated message or a MAC CE dedicated message, and when the network receives the dedicated indication message, it knows that the UE wants to report the stored measurement data. Alternatively, the reporting request message can also be an RRC message.

[0163] In one way, the reporting request message includes a 1-bit indication field to indicate whether the UE needs to report the AI associated data. For example, when the indication field is 1, it means that the UE wants to report the stored measurement data; when the indication field is 0, it means that the UE wants to report the stored measurement data. When the network receives the reporting request message, it knows that the UE wants to report the stored measurement data according to the setting of the indication field in the message; the reporting request message can be a UCI, MAC CE, or RRC message.

[0164] In some embodiments, the reporting request message can be used to indicate the reason why the UE needs to report the AI associated data. The reason identifier is used to indicate the reason why the UE needs to report the stored measurement data on the NW side. For example, reason #1: radio link failure; reason #2: radio link monitoring failure; reason #3: channel quality deterioration; reason #4: UE power deterioration. Alternatively, reason #1: continuous N310 synchronization loss indication (i.e. radio link problem); reason #2: T310 running; reason #3: T310 timeout; reason #4: T311 running; reason #5: continuous access random access N times, N is an integer, N is less than or equal to the maximum random access number, N is predefined or configured, etc. When the network receives the reporting request message, it knows that the UE wants to report the stored measurement data according to the reason identifier for the UE to report the data. The reporting request message can be a UCI, MAC CE, or RRC message.

[0165] In some embodiments, the reporting request message can further comprise a buffer size of the AI-associated data.

[0166] It should be understood that the reporting request message can comprise a combination of the above-mentioned manners. For example, the UE can send a dedicated reporting request message for reporting stored measurement data, wherein at least a 1-bit indication field or / and a reason for reporting measurement data is included. Alternatively, the UE can send a reporting request message for reporting stored measurement data, wherein at least a 1-bit indication field or / and a reason for reporting measurement data is included. Alternatively, the UE can send a dedicated / non-dedicated reporting request message for reporting stored measurement data, wherein at least a 1-bit indication field or / and a reason for reporting measurement data is included. Optionally, a buffer size of the stored data to be reported can also be included.

[0167] In S705, the network-side device sends a reporting request message (reporting request feedback message) to the UE, and the UE receives the reporting request feedback message.

[0168] The network side sends a corresponding reply message to agree or refuse the UE to report data, and if agreed, the resource information for transmitting data can also be carried.

[0169] In S706, the UE decides whether to perform data reporting according to the reporting request feedback message. If it is decided to report, the UE reports the stored AI-associated data to the network side.

[0170] FIG. 10 is a flowchart of an AI-associated data reporting method according to another embodiment of the present application. In this embodiment, the network-side device can initiate the request for data reporting. As shown in FIG. 10, the method can comprise steps S801 to S803. The present embodiment mainly designs a scheme of data reporting request initiated by the NW.

[0171] In S801, the network-side entity sends an RRC release message to the UE.

[0172] Before the NW intends to release the UE from the RRC connected state to the RRC non-connected state, the NW can send data reporting request information to the UE. The reporting request information can be used to indicate the AI-associated data. For example, the network-side entity can carry the aforementioned reporting request information in the RRC release message, so that the UE can know that it needs to perform AI-associated data reporting according to the reporting request information.

[0173] Specifically, the request reporting information can include at least one of the following: a storage indication identifier corresponding to the AI associated data; an artificial intelligence (AI) model identifier corresponding to the AI associated data; a measurement identifier corresponding to the AI associated data; a function identifier corresponding to the AI associated data; a region identifier corresponding to the AI associated data; or a cell identifier corresponding to the AI associated data.

[0174] Optionally, the RRC release message can also carry time-frequency resource information of the transmission resource, and the time-frequency resource information can be estimated by the base station itself.

[0175] In S802, the UE sends a data reporting message to the network side device.

[0176] Based on the received data request reporting information, the UE sends a data reporting message to the NW. The reporting message carries the stored measurement data. The reporting message is a new RRC message for carrying data, or if the amount of measured data is not very large, the stored measurement data can also be carried by a MAC message.

[0177] In this example, the data reporting message is sent by the UE to the NW before the transmission of the traditional RRC release complete message. It should be understood that when the UE receives the data request reporting request information, the corresponding data reporting message can be sent immediately, or the data reporting message can be sent after waiting for a period of time.

[0178] In some other cases, the UE can carry the stored measurement data in the RRCReleasecomplete message sent to the NW, that is, step S802 and step S803 can be combined.

[0179] In S803, the UE sends an RRC release complete message (RRCReleaseComplete) to the network side device.

[0180] The UE sends the RRCReleasecomplete message and enters the RRC idle state or the RRC inactive state.

[0181] FIG. 11 is a schematic block diagram of a communication device 900 according to an embodiment of the present application. As shown in FIG. 11, the communication device 900 includes a processor 901 and a memory 902, and the processor 901 is communicatively connected with the memory 902. The communication device 900 can be, for example but not limited to, a user equipment or a network side device (such as a base station). In some embodiments, the communication device 900 can further include a transceiver for transmitting / receiving data, or only include a transmitting circuit for transmitting data, or only include a receiving circuit for receiving data. The memory 902 of the communication device 900 is used to store program instructions, which can be executed by the processor 901 to implement the AI-associated data reporting method or the reporting configuration method described in any of the foregoing embodiments.

[0182] It should be understood that the processor in the embodiments of the present application can be an integrated circuit chip having a processing capability of signals. In the implementation process, each step of the method embodiments described above can be completed by integrated logic circuits in the hardware or instructions in the form of software in the processor.

[0183] It can be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. It should be noted that the memory of the system and method described herein is intended to include but not limited to these and any other suitable types of memory. The embodiments of the present application also provide a computer readable storage medium for storing a computer program.

[0184] Optionally, the computer readable storage medium can be applied to the communication device in the embodiments of the present application, and the computer program makes the computer execute the corresponding processes realized by the communication device in the various methods of the embodiments of the present application. For the sake of brevity, details are not repeated here. Alternatively, the computer readable storage medium can be applied to the UE or the network side device in any of the embodiments of the present application, and the computer program makes the computer execute the processes realized by the UE or the network side device in the various methods of the embodiments of the present application. For the sake of brevity, details are not repeated here.

[0185] The embodiments of the present application also provide a computer program product including computer program instructions.

[0186] Optionally, the computer program product can be applied to the communication device in the embodiments of the present application, and the computer program instructions make the computer execute the corresponding processes realized by the communication device in the various methods of the embodiments of the present application. For the sake of brevity, details are not repeated here.

[0187] Those skilled in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on specific application and design constraints of the technical solution. Those skilled in the art can use different methods to realize the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0188] The above description is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or replacements within the technical scope disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A method for reporting artificial intelligence model associated data, performed by a user equipment (UE), comprising: storing artificial intelligence (AI) associated data in a radio resource control (RRC) non-connected state; and reporting the AI associated data to a first network side entity after the UE enters an RRC connected state. Before the storing the AI associated data, further comprising: obtaining a storage indication message from the first network side entity or a second network side entity; wherein the reporting the AI associated data to the first network side entity after the UE enters the RRC connected state comprises: reporting the AI associated data to the first network side entity according to the storage indication message.

3. The method of claim 2, further comprising: sending a transmission completion identifier of the AI associated data to the first network side entity in response to the AI associated data being completely sent.

4. The method of claim 2, wherein: the storage indication message is used to indicate whether the UE stores the AI associated data across RRC connection states; or the storage indication message is used to indicate whether the UE keeps storing the AI associated data.

2. The method of claim 1, wherein, 5. The method of claim 2, wherein: the storage indication message is used to indicate the AI associated data that the UE needs to store.

6. The method of claim 5, wherein: the storage indication message comprises a storage indication identifier; the storage indication message comprises at least one of: a measurement configuration identifier corresponding to the AI associated data; a measurement content corresponding to the AI associated data; a reporting configuration identifier corresponding to the AI associated data; a reporting content corresponding to the AI associated data; a function identifier corresponding to the AI associated data; a model identifier corresponding to the AI associated data; or a data set identifier corresponding to the AI associated data. The storage indication message is associated with one of: a given artificial intelligence (AI) function; a given area; a given cell; or a given AI model.

9. The method of claim 8, wherein: the storage indication message comprises a first indication message and a second indication message; the first indication message is used to indicate the given AI function or the given AI model; and the second indication message is used to indicate the given area or the given cell.

10. The method of claim 9, wherein: the first indication message is transmitted by unicast, and the second indication message is transmitted by groupcast.

11. The method of claim 10, wherein: the first indication message and the second indication message have corresponding association identifiers.

12. The method of claim 5, wherein: the UE is configured with data measurement and / or storage configurations corresponding to a plurality of storage indication identifiers; the storage indication message comprises a target storage indication identifier; and the AI associated data is determined according to the target storage indication identifier and the data measurement and / or storage configurations corresponding to the plurality of storage indication identifiers. The reporting the AI associated data comprises: determining whether a validity timer of the AI associated data is expired when the UE enters the RRC connected state; and ​ ​ ​ ​ ​ 7. The method of claim 5, wherein, ​ ​ ​ ​ ​ ​ ​ ​ 8. The method of claim 5, wherein, ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ 13. The method of claim 1, wherein, ​ ​ reporting the AI-associated data in response to the validity timer not expiring. 14.The method of claim 13, further comprising: releasing the AI-associated data and / or reporting invalidation information of the AI-associated data to the first network-side entity in response to the validity timer expiring.

15. The method of claim 13, wherein, starting the validity timer in response to one of the following conditions: receiving a data collection configuration message; the UE starting to perform measurement or storage of the AI-associated data; or receiving an activation message for measurement of the AI-associated data.

16. The method of claim 13, wherein, stopping the validity timer in response to one of the following conditions: the UE selecting to another cell; the UE selecting to a cell within a specific area; receiving a message that the AI-associated data has been collected completely; or the AI-associated data has been sent completely.

17. The method of claim 1, wherein, the reporting the AI-associated data comprises: determining whether the AI-associated data belongs to a specified range when the UE enters an RRC connected state; and reporting the AI-associated data in response to the AI-associated data belonging to the specified range.

18. The method of claim 17, wherein, the determining whether the AI-associated data belongs to the specified range comprises: receiving an indication from a second network-side entity whether the AI-associated data belongs to the specified range; or receiving an indication from the first network-side entity whether the AI-associated data belongs to the specified range.

19. The method of claim 17, wherein, the determining whether the AI-associated data belongs to the specified range comprises: obtaining an RRC reconfiguration message from a second network-side entity; determining whether the AI-associated data belongs to a specified range according to area or cell information of the first network-side entity in the RRC reconfiguration message.

20. The method of claim 1, wherein, the reporting the AI-associated data comprises: sending a data reporting request message to the first network-side entity when the UE enters an RRC connected state; receiving an approval message of the data reporting request message from the first network-side entity; and reporting the AI-associated data in response to the approval message.

21. The method of claim 20, wherein, the reporting request message comprises at least one of: a storage indication identity corresponding to the AI-associated data; cell information corresponding to the AI-associated data; area information corresponding to the AI-associated data; or AI model information corresponding to the AI-associated data. 22.The method of claim 21, wherein the reporting request message further comprises: a cache size corresponding to the AI-associated data.

23. The method of claim 1, wherein, the reporting the AI-associated data comprises: receiving a data collection request message from the first network-side entity; and reporting the AI-associated data in response to the data collection request message. 24.The method of claim 23, wherein the data collection request message is transmitted by unicast and comprises a temporary identity of the UE, wherein the temporary identity is used for reporting of the AI-associated data. 25.The method of claim 24, further comprising: obtaining the temporary identity; and releasing the temporary identity when the AI-associated data reporting is completed or invalid. 26.The method of claim 23, wherein ​ The data collection request message is transmitted in a multicast manner and includes cell or area information corresponding to the AI-associated data.

27. A configuration method for reporting AI-associated data, performed by a network-side entity, comprising: sending a storage indication message to a user equipment (UE), wherein the storage indication message is used to indicate whether the UE stores AI-associated data in a cross-RRC connection state or whether the UE keeps storing the AI-associated data.

28. The method of claim 27, wherein the storage indication message is used to indicate the AI-associated data that the UE needs to store.

29. The method of claim 28, wherein the storage indication message includes a storage indication identifier.

30. The method of claim 28, wherein, The storage indication message includes at least one of the following: a measurement configuration identifier corresponding to the AI-associated data; a measurement content corresponding to the AI-associated data; a reporting configuration identifier corresponding to the AI-associated data; a reporting content corresponding to the AI-associated data; a function identifier corresponding to the AI-associated data; a model identifier corresponding to the AI-associated data; or a dataset identifier corresponding to the AI-associated data.

31. The method of claim 28, wherein, The storage indication message is associated with one of the following: a given AI function; a given area; a given cell; or a given AI model.

32. The method of claim 31, wherein the storage indication message includes a first indication message and a second indication message; the first indication message is used to indicate the given AI function or the given AI model; the second indication message is used to indicate the given area or the given cell.

33. The method of claim 32, wherein the first indication message is transmitted in a unicast manner, and the second indication message is transmitted in a multicast manner.

34. The method of claim 33, wherein the first indication message and the second indication message have corresponding association identifiers.

35. A method for reporting AI-associated data, performed by a user equipment (UE), comprising: storing AI-associated data; and reporting the AI-associated data to a network-side entity when a preset condition is met.

36. The method of claim 35, wherein, The reporting of the AI-associated data to the network-side entity when the preset condition is met includes: determining whether a radio resource control (RRC) state of the UE is about to change from a connected state to a non-connected state; and in response to the RRC state of the UE being about to change from the connected state to the non-connected state, reporting the AI-associated data to the network-side entity.

37. The method of claim 36, wherein, The determination of whether the RRC state of the UE is about to change from the connected state to the non-connected state includes: obtaining a quality of a current wireless link; and when the quality of the current wireless link is less than a preset threshold, determining that the RRC state of the UE is about to change from the connected state to the non-connected state.

38. The method of claim 36, wherein, The determination of whether the RRC state of the UE is about to change from the connected state to the non-connected state includes: determining that a radio resource control (RRC) state of the UE is about to change from a connected state to an unconnected state when a current radio link monitoring of the UE fails.

39. The method of claim 36, wherein, The determining whether the RRC state of the UE is about to change from the connected state to the unconnected state comprises: determining that a radio resource control (RRC) state of the UE is about to change from a connected state to an unconnected state when the UE detects that a current radio link is about to fail.

40. The method of claim 35, wherein, The preset condition is that a current power consumption of the UE is less than a preset threshold.

41. The method of claim 35, wherein, The preset condition is that a current memory of the UE reaches a preset threshold.

42. The method of claim 35, wherein, The preset condition is that a validity timer of the AI-associated data expires.

43. The method of claim 35, before the storing the AI-associated data, further comprising: obtaining a data collection configuration, the data collection configuration being used to indicate a content of the AI-associated data and / or the preset condition.

44. The method of claim 35, before the reporting the AI-associated data to the network-side entity, further comprising: sending a reporting request message to the network-side entity; and obtaining a reporting request feedback message from the network-side entity; wherein the reporting the AI-associated data to the network-side entity comprises reporting the AI-associated data to the network-side entity according to the reporting request feedback.

45. The method of claim 44, wherein, The reporting request message comprises a 1-bit indication field to indicate whether the UE needs to report the AI-associated data.

46. The method of claim 44, wherein, The reporting request message is used to indicate a reason why the UE needs to report the AI-associated data.

47. The method of claim 44, wherein, The reporting request message is a dedicated indication message for the UE to report the AI-associated data, and the dedicated indication message is a uplink control information (UCI) dedicated message, a medium access control element (MAC-CE) dedicated message, or a radio resource control (RRC) dedicated message.

48. The method of claim 44, wherein, The reporting request message comprises a buffer size occupied by the AI-associated data.

49. The method of claim 35, wherein, The preset condition is that a release message of a radio resource control (RRC) state is received from the network-side entity.

50. The method of claim 49, further comprising: obtaining a reporting request information from the network-side entity, the reporting request information being used to indicate the AI-associated data.

51. The method of claim 50, wherein, The reporting request information comprises at least one of: a storage indication identity corresponding to the AI-associated data; an artificial intelligence (AI) model identity corresponding to the AI-associated data; a measurement identity corresponding to the AI-associated data; a function identity corresponding to the AI-associated data; a region identity corresponding to the AI-associated data; or a cell identity corresponding to the AI-associated data.

52. The method of claim 50, wherein, The reporting the AI-associated data to the network-side entity comprises: including the AI-associated data in a release completion message of the RRC state.

53. A communication device comprising a processor and a memory, the memory being configured to store program instructions which, when executed by the processor, implement the method described in any one of claims 1 to 52.

54. A readable storage medium for storing program instructions, wherein, The program instructions, when executed by the processor, implement the method described in any one of claims 1 to 52.

Images