Operation execution method and apparatus, and network side device

Abstract

The present application relates to the technical field of communications, and discloses an operation execution method and apparatus, and a network side device. The operation execution method in the embodiments of the present application comprises: an information acquisition device receives first information sent by an information source device, wherein the first information is transmission-related information of the information source device; and the information acquisition device performs a first operation of a target AI unit on the basis of the first information, wherein the target AI unit is associated with the information acquisition device, and the first operation comprises at least one of the following: model inference, model monitoring, and training data collection.

Description

Operation execution method, device and network side equipment

[0001] Cross-reference to related applications

[0002] This application claims priority to Chinese Patent Application No. 202510306584.3, filed in China on March 14, 2025, the entire contents of which are incorporated herein by reference. Technical Field

[0003] This application belongs to the field of communication technology, specifically relating to an operation execution method, apparatus, and network-side equipment. Background Technology

[0004] In related technologies, multi-cell joint scheduling can improve the transmission performance of cell edge terminals. Multi-cell joint scheduling requires selection of multiple transmission and reception points (TRPs), user selection, and precoding. Compared to single-cell scheduling, the computational complexity of multi-cell joint decision-making is significantly increased, leading to a higher probability that scheduling latency cannot meet timeline requirements. Summary of the Invention

[0005] This application provides an operation execution method, apparatus, and network-side device that can solve the problem that scheduling latency is likely to fail to meet timeline requirements.

[0006] Firstly, an operation execution method is provided, the method comprising:

[0007] The information acquisition device receives first information sent by the information source device, wherein the first information is transmission-related information of the information source device;

[0008] The information acquisition device performs a first operation on the target AI unit based on the first information;

[0009] The target AI unit is associated with the information acquisition device;

[0010] The first operation includes at least one of the following:

[0011] Model inference, model monitoring, and training data collection.

[0012] Secondly, an operation execution method is provided, the method comprising:

[0013] The information source device sends first information to the information acquisition device, wherein the first information is transmission-related information of the information source device.

[0014] The first information is used to perform the first operation of the target AI unit;

[0015] The target AI unit is associated with the information acquisition device;

[0016] The first operation includes at least one of the following:

[0017] Model inference, model monitoring, and training data collection.

[0018] Thirdly, an operation execution device is provided, comprising:

[0019] The receiving module is used to receive first information sent by the information source device, wherein the first information is transmission-related information of the information source device;

[0020] The processing module is used to perform a first operation on the target AI unit based on the first information;

[0021] The target AI unit is associated with the information acquisition device;

[0022] The first operation includes at least one of the following:

[0023] Model inference, model monitoring, and training data collection.

[0024] Fourthly, an operation execution device is provided, comprising:

[0025] The sending module is used to send first information to the information acquisition device, wherein the first information is transmission-related information of the information source device;

[0026] The first information is used to perform the first operation of the target AI unit;

[0027] The target AI unit is associated with the information acquisition device;

[0028] The first operation includes at least one of the following:

[0029] Model inference, model monitoring, and training data collection.

[0030] Fifthly, an operation execution device is provided, the device being configured to perform the steps of the method described in the first aspect, or to implement the steps of the method described in the second aspect.

[0031] In a sixth aspect, a network-side device is provided, the network-side device including a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the method as described in the first or second aspect.

[0032] In a seventh aspect, a network-side device is provided, which is an information acquisition device, including a processor and a communication interface, wherein...

[0033] A communication interface is used to receive first information sent by an information source device, wherein the first information is transmission-related information of the information source device.

[0034] A processor is configured to perform a first operation on the target AI unit based on the first information;

[0035] The target AI unit is associated with the information acquisition device;

[0036] The first operation includes at least one of the following:

[0037] Model inference, model monitoring, and training data collection.

[0038] Eighthly, a network-side device is provided, which is an information source device, including a processor and a communication interface, wherein...

[0039] A communication interface is used to send first information to an information acquisition device, wherein the first information is transmission-related information of the information source device.

[0040] The first information is used to perform the first operation of the target AI unit;

[0041] The target AI unit is associated with the information acquisition device;

[0042] The first operation includes at least one of the following:

[0043] Model inference, model monitoring, and training data collection.

[0044] A ninth aspect provides a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method described in the first aspect, or implement the steps of the method described in the second aspect.

[0045] In a tenth aspect, a wireless communication system is provided, comprising: an information acquisition device and an information source device, wherein the information acquisition device is configured to perform the steps of the method described in the first aspect, and the information source device is configured to perform the steps of the method described in the second aspect.

[0046] Eleventhly, a chip is provided, the chip including a processor and a communication interface coupled to the processor, the processor being configured to run a program or instructions to implement the steps of the method described in the first aspect, or to implement the steps of the method described in the second aspect.

[0047] In a twelfth aspect, a computer program / program product is provided, the computer program / program product being stored in a storage medium, the computer program / program product being executed by at least one processor to implement the method as described in the first aspect, or to implement the method as described in the second aspect.

[0048] In this embodiment, the information acquisition device receives first information sent by the information source device, the first information being transmission-related information of the information source device; the information acquisition device performs a first operation on a target AI unit based on the first information; wherein the target AI unit is associated with the information acquisition device; the first operation includes at least one of the following: model inference, model monitoring, and training data collection. Thus, the information acquisition device performs a first operation on the transmission-related information of the information source device through the AI ​​unit to determine the transmission strategy of at least one of the information acquisition device and the information source device. Compared to high-performance exhaustive search schemes or greedy algorithms, using an AI unit can reduce computational complexity, thereby increasing the probability that scheduling latency meets timeline requirements; furthermore, by outputting decisions through the AI ​​unit, predictions of future channel state information of the terminal can be considered, which can improve transmission performance. Attached Figure Description

[0049] Figure 1 is a block diagram of a wireless communication system applicable to an embodiment of this application;

[0050] Figure 2a is a schematic diagram of DPS uplink and downlink transmission in related technologies;

[0051] Figure 2b is a schematic diagram of ICBM uplink and downlink transmission in a related technology;

[0052] Figure 2c is a schematic diagram of CJT uplink and downlink transmission in related technologies;

[0053] Figure 2d is a schematic diagram of NCJT uplink and downlink transmission in a related technology;

[0054] Figure 3 is a flowchart of one of the operation execution methods provided in an embodiment of this application;

[0055] Figure 4a is a schematic diagram of information interaction provided in an embodiment of this application;

[0056] Figure 4b is one of the transmission schematic diagrams provided in the embodiments of this application;

[0057] Figure 4c is one of the transmission schematic diagrams provided in the embodiments of this application;

[0058] Figure 5a is one of the transmission schematic diagrams provided in the embodiments of this application;

[0059] Figure 5b is a second flowchart of an operation execution method provided in an embodiment of this application;

[0060] Figure 6a is a second schematic diagram of a transmission method provided in an embodiment of this application;

[0061] Figure 6b is a third flowchart of an operation execution method provided in an embodiment of this application;

[0062] Figure 7a is a third schematic diagram of a transmission method provided in an embodiment of this application;

[0063] Figure 7b is a flowchart of an operation execution method provided in an embodiment of this application;

[0064] Figure 8a is a fourth schematic diagram of a transmission method provided in an embodiment of this application;

[0065] Figure 8b is a flowchart of an operation execution method provided in an embodiment of this application;

[0066] Figure 9a is a fifth schematic diagram of a transmission method provided in an embodiment of this application;

[0067] Figure 9b is a flowchart of an operation execution method provided in an embodiment of this application;

[0068] Figure 10a is a sixth schematic diagram of a transmission method provided in an embodiment of this application;

[0069] Figure 10b is a flowchart of an operation execution method provided in an embodiment of this application;

[0070] Figure 11a is a seventh schematic diagram of a transmission method provided in an embodiment of this application;

[0071] Figure 11b is a flowchart of an operation execution method provided in an embodiment of this application;

[0072] Figure 12a is an eighth schematic diagram of a transmission method provided in an embodiment of this application;

[0073] Figure 12b is a flowchart of an operation execution method provided in an embodiment of this application;

[0074] Figure 13 is a ninth schematic diagram of a transmission method provided in an embodiment of this application;

[0075] Figure 14 is a tenth schematic diagram of a transmission method provided in an embodiment of this application;

[0076] Figure 15 is an eleventh schematic diagram of a transmission method provided in an embodiment of this application;

[0077] Figure 16 is a flowchart of an operation execution method provided in an embodiment of this application;

[0078] Figure 17 is a flowchart of an operation execution method provided in an embodiment of this application;

[0079] Figure 18 is a flowchart of an operation execution method provided in an embodiment of this application;

[0080] Figure 19 is a flowchart of an operation execution method provided in an embodiment of this application;

[0081] Figure 20 is one of the structural schematic diagrams of an operation execution device provided in an embodiment of this application;

[0082] Figure 21 is a second schematic diagram of the structure of an operation execution device provided in an embodiment of this application;

[0083] Figure 22 is a schematic diagram of the structure of a communication device provided in an embodiment of this application;

[0084] Figure 23 is one of the structural schematic diagrams of a network-side device provided in an embodiment of this application;

[0085] Figure 24 is a second schematic diagram of the structure of a network-side device provided in an embodiment of this application. Detailed Implementation

[0086] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0087] The terms "first," "second," etc., used in this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first" and "second" are generally of the same class, not limited in number; for example, the first object can be one or more. Furthermore, "or" in this application indicates at least one of the connected objects. For example, the scope of protection for "A or B" covers at least three scenarios: Scenario 1: including A but not B; Scenario 2: including B but not A; Scenario 3: including both A and B. In addition, the terms "A and / or B," "at least one of A and B," and "at least one of A or B" also cover at least the above three scenarios. The character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0088] The term "instruction" in this application can be either a direct instruction (or explicit instruction) or an indirect instruction (or implicit instruction). A direct instruction can be understood as the sender explicitly informing the receiver of specific information, the required operation, or the requested result in the instruction sent. An indirect instruction can be understood as the receiver determining the corresponding information based on the instruction sent by the sender, or making a judgment and determining the required operation or requested result based on the judgment result.

[0089] It is worth noting that the technologies described in this application are not limited to Long Term Evolution (LTE) / LTE-Advanced (LTE-A) systems, but can also be used in other wireless communication systems, such as Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access (OFDMA), Single-carrier Frequency-Division Multiple Access (SC-FDMA), or other systems. The terms "system" and "network" in this application are often used interchangeably, and the described technologies can be used in the systems and radio technologies mentioned above, as well as in other systems and radio technologies. The following description describes New Radio (NR) systems for illustrative purposes, and the term NR is used in most of the following description; however, these technologies can also be applied to systems other than NR systems, such as 6th Generation (6G) communication systems.

[0090] Figure 1 shows a block diagram of a wireless communication system applicable to an embodiment of this application. The wireless communication system includes a terminal 11 and a network-side device 12. The terminal 11 can also be referred to as User Equipment (UE), and can be a mobile phone, tablet computer, laptop computer, notebook computer, personal digital assistant (PDA), handheld computer, netbook, ultra-mobile personal computer (UMPC), mobile internet device (MID), augmented reality (AR), virtual reality (VR) device, robot, wearable device, flight vehicle, vehicle user equipment (VUE), shipboard equipment, pedestrian user equipment (PUE), smart home (home devices with wireless communication capabilities, such as refrigerators, televisions, washing machines, or furniture), game console, personal computer (PC), ATM, or self-service machine, etc. Wearable devices include: smartwatches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart chains, smart rings, smart necklaces, smart anklets, smart anklets, etc.), smart wristbands, smart clothing, etc. Among these, in-vehicle devices can also be referred to as in-vehicle terminals, in-vehicle controllers, in-vehicle modules, in-vehicle components, in-vehicle chips, or in-vehicle units, etc. It should be noted that the specific type of terminal 11 is not limited in this application embodiment. Network-side equipment 12 may include access network equipment or core network equipment, wherein access network equipment may also be referred to as Radio Access Network (RAN) equipment, radio access network function, or radio access network unit. Access network equipment may include base stations, Wireless Local Area Network (WLAN) access points (APs), or Wireless Fidelity (WiFi) nodes, etc.Among them, base stations can be referred to as Node B (NB), Evolved Node B (eNB), Next Generation Node B (gNB), New Radio Node B (NR Node B), Access Point, Relay Base Station (RBS), Serving Base Station (SBS), Base Transceiver Station (BTS), Radio Base Station, Radio Transceiver, Basic Service Set (BSS), Extended Service Set (ESS), Home Node B (HNB), Home Evolved Node B, Transmission and Reception Point (TRP), and Non-Terrestrial Network (NTN) equipment (such as satellite or high altitude platform). The term "base station" can be any suitable term in the field, such as "station" or any other appropriate term in the relevant field, as long as the same technical effect is achieved. The term "base station" is not limited to specific technical terms. It should be noted that the embodiments of this application only use the base station in the NR system as an example for introduction, and do not limit the specific type of base station.

[0091] Core network equipment, also known as core network nodes, core network functions, or core network elements, includes, but is not limited to, at least one of the following: Mobility Management Entity (MME), Access and Mobility Management Function (AMF), Session Management Function (SMF), User Plane Function (UPF), Policy Control Function (PCF), Policy and Charging Rules Function (PCRF), Edge Application Server Discovery Function (EASDF), Unified Data Management (UDM), Unified Data Repository (UDR), Home Subscriber Server (HSS), Centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (L-NEF), and Binding Support. Functions include BSF, Application Function (AF), Location Management Function (LMF), Gateway Mobile Location Centre (GMLC), Network Data Analytics Function (NWDAF), and Non-Terrestrial Network (NTN) equipment (such as satellite or high altitude platform station).It should be noted that the embodiments of this application only use the core network equipment in the NR system as an example for introduction, and do not limit the specific type of core network equipment. If the name of the core network equipment mentioned in the embodiments of this application changes in subsequent protocol versions (e.g., 6G), it is also within the scope of protection of this application.

[0092] Optionally, the core network equipment can be implemented by one or more functional modules in a single device, or by multiple devices working together; this application does not specifically limit this. It is understood that the aforementioned functional modules can be network elements in hardware devices, software functional modules running on dedicated hardware, or virtualized functional modules instantiated on a platform (e.g., a cloud platform).

[0093] For ease of understanding, the following explains some aspects of the embodiments of this application:

[0094] 1. Dynamic point selection (DPS)

[0095] As shown in Figure 2a, DPS refers to the sharing of user data among multiple Transmit / Receive Points (TRPs). In each subframe, uplink and downlink data are sent to the UE by one of the TRPs according to the channel conditions.

[0096] Scenario 1: If multiple TRPs belong to the same cell, the channel conditions can be used to determine which TRP will perform uplink reception or downlink transmission.

[0097] Scenario 2: If multiple TRPs belong to different cells, the UE needs to first switch to a certain cell, and then a certain TRP under that cell will perform uplink and downlink transmission for the UE.

[0098] Downlink transmission includes the transmission of downlink control signaling (such as the Physical downlink control channel (PDCCH)) or downlink data information (such as the Physical downlink shared channel (PDSCH)); uplink transmission includes the transmission of uplink data (such as the Physical uplink shared channel (PUSCH)) or uplink control information (such as channel quality indicator (CQI), Hybrid Automatic Repeat Request (HARQ), Precoding Matrix Indicator (PMI), Reference Signal Received Power (RSRP), etc.).

[0099] 2. Inter-cell beam management (ICBM)

[0100] As shown in Figure 2b, ICBM refers to the sharing of user data among multiple Transmit / Receive Points (TRPs). In each subframe, depending on the channel conditions, one TRP sends downlink data to the UE, and another TRP sends uplink data to the UE. The TRPs sending uplink and downlink data can be the same or different.

[0101] For ICBM, multiple TRPs can belong to different cells. The UE does not need to handover, and uplink or downlink transmission can be performed for the UE by a TRP in a neighboring cell.

[0102] Downlink transmission includes the transmission of downlink control signaling (PDCCH) or downlink data information (such as PDSCH); uplink transmission includes the transmission of uplink data (such as PUSCH) or uplink control information (such as CQI, HARQ, PMI, RSRP, etc.).

[0103] 3. Coherent Joint Transmission (CJT)

[0104] As shown in Figure 2c, CJT is a joint precoding mechanism among various TRPs. Each TRP sends the same data to the UE, coordinating the precoding matrices of different TRPs to ensure coherent superposition of the data streams at the receiving end. CJT has high synchronization requirements among multiple TRPs and is easily affected by factors such as synchronization clock, inaccurate frequency offset estimation, Doppler shift, and path loss estimation, which can have unpredictable impacts on CJT performance.

[0105] Scenario 1: If multiple TRPs belong to the same cell, it can be determined, based on channel conditions, that some TRPs will jointly perform downlink transmission or uplink reception. The uplink and downlink TRPs can be the same or different.

[0106] Scenario 2: If multiple TRPs belong to different cells, the UE does not need to handover and can have joint downlink or uplink transmissions performed for the UE by one or more TRPs in the local cell and neighboring cells.

[0107] Downlink transmission includes the transmission of downlink control signaling (PDCCH) or downlink data information (such as PDSCH); uplink transmission includes the transmission of uplink data (such as PUSCH) or uplink control information (such as CQI, HARQ, PMI, RSRP, etc.).

[0108] It is important to note that in order for downlink CJT to be enabled, the UE needs to perform joint channel feedback using multi-TRP joint precoding when reporting CSI, rather than reporting based on the CSI information of a single TRP. Otherwise, the relative phase information between each TRP will be lost, thus preventing downlink CJT from being enabled.

[0109] 4. Non-Coherent Joint Transmission (NCJT)

[0110] As shown in Figure 2d, NCJT involves independent precoding by each TRP, with each TRP sending different data to the UE. The UE's data rate is equivalent to the sum of the rates of each TRP. NCTJ does not require joint beamforming between TRPs; each TRP can perform precoding independently, and joint calibration of relative phases is not required.

[0111] Scenario 1: If multiple TRPs belong to the same cell, it can be determined, based on channel conditions, that some TRPs will independently perform downlink transmission or uplink reception. The uplink and downlink TRPs can be the same or different.

[0112] Scenario 2: If multiple TRPs belong to different cells, the UE does not need to handover and can have independent downlink or uplink transmissions performed for the UE by one or more TRPs in the local cell and neighboring cells.

[0113] Downlink transmission includes the transmission of downlink control signaling (PDCCH) or downlink data information (such as PDSCH); uplink transmission includes the transmission of uplink data (such as PUSCH) or uplink control information (such as CQI, HARQ, PMI, RSRP, etc.).

[0114] 5. Artificial Intelligence (AI) Unit

[0115] The AI ​​unit described in this application embodiment may also be referred to as an AI model, machine learning (ML) model, ML unit, AI structure, AI function, AI characteristic, machine learning model, neural network, neural network function, neural network functionality, etc. Alternatively, the AI ​​unit may refer to a processing unit capable of implementing specific algorithms, formulas, processing flows, capabilities, etc., related to AI. Or, the AI ​​unit may be a processing method, algorithm, function, module, or unit for a specific dataset. Alternatively, the AI ​​unit may be a processing method, algorithm, function, module, or unit running on AI-related hardware such as a graphics processing unit (GPU), neural network processing unit (NPU), tensor processing unit (TPU), or application-specific integrated circuit (ASIC). This application embodiment does not specifically limit this. Optionally, the specific dataset includes the input or output of the AI ​​unit.

[0116] Optionally, the identifier of the AI ​​unit / AI model may be an AI model identifier, an AI structure identifier, an AI algorithm identifier, or an identifier of a specific dataset associated with the AI ​​unit / AI model, or an identifier of a specific scenario, environment, channel characteristics, or device related to the AI / ML, or an identifier of a function, feature, capability, or module related to the AI / ML. This application embodiment does not specifically limit this.

[0117] The operation execution method, apparatus and related equipment provided in this application will be described in detail below with reference to the accompanying drawings and through some embodiments and application scenarios.

[0118] Referring to Figure 3, which is a flowchart of an operation execution method provided in an embodiment of this application, the operation execution method includes the following steps:

[0119] Step 101: The information acquisition device receives the first information sent by the information source device, wherein the first information is the transmission-related information of the information source device;

[0120] Step 102: The information acquisition device performs a first operation on the target artificial intelligence (AI) unit based on the first information;

[0121] The target AI unit is associated with the information acquisition device;

[0122] The first operation includes at least one of the following:

[0123] Model inference, model monitoring, and training data collection.

[0124] In one embodiment, the information acquisition device performs a first operation on the target AI unit based on the first information to obtain a first operation result, which can be used to determine the transmission strategy of at least one of the information acquisition device and the information source device.

[0125] In one embodiment, after obtaining the first operation result, the method further includes:

[0126] The information acquisition device determines its transmission strategy based on the result of the first operation.

[0127] The information acquisition device transmits information based on the determined transmission strategy.

[0128] In one embodiment, after obtaining the first operation result, the method further includes:

[0129] The information acquisition device determines the transmission strategy of the information source device based on the result of the first operation;

[0130] The information acquisition device performs at least one of the following:

[0131] Send the transmission strategy of the information source device to the information source device;

[0132] Send indication information to the information source device, the indication information being used to indicate whether the information acquisition device supports the determined transmission strategy;

[0133] The transmission strategy of the information acquisition device is adjusted based on the determined transmission strategy.

[0134] In one embodiment, after obtaining the first operation result, the method further includes:

[0135] The information acquisition device sends the first operation result to the information source device.

[0136] It should be noted that, when the first operation includes model inference, the result of the first operation includes the model inference result; when the first operation includes model monitoring, the result of the first operation includes the model monitoring result; when the first operation includes training data collection, the result of the first operation includes the dataset used for training the target AI unit.

[0137] In one embodiment, when the first operation includes model inference, the information acquisition device performs model inference operation on the target AI unit to obtain a model inference result (at this time, the result of the first operation includes the model inference result); the information acquisition device determines the transmission-related scheduling information of the information source device based on the model inference result, and sends the transmission-related scheduling information of the information source device to the information source device, wherein the transmission strategy of the information source device includes the transmission-related scheduling information of the information source device.

[0138] In one implementation, when the first operation includes model monitoring, the information acquisition device performs a model monitoring operation on the target AI unit to obtain a model monitoring result (at this time, the first operation result includes the model monitoring result); the information acquisition device determines, based on the model monitoring result, whether the target AI unit can still be used to determine the transmission strategy of at least one of the information acquisition device and the information source device, or whether to deactivate the target AI unit or update the model.

[0139] In one embodiment, when the first operation includes training data collection, the information acquisition device performs a training data collection operation for the target AI unit to obtain a dataset (at this time, the result of the first operation includes the dataset); the information acquisition device trains the target AI unit based on the dataset, and the trained target AI unit can be used to predict the transmission strategy of at least one of the information acquisition device and the information source device.

[0140] Furthermore, the association between the target AI unit and the information acquisition device can mean that the target AI unit is deployed on the information acquisition device side; that is, the information acquisition device has the target AI unit deployed on it. It should be understood that the target AI unit can be deployed directly on the information acquisition device, or on the cloud server corresponding to the information acquisition device, or on the edge device corresponding to the information acquisition device.

[0141] In this embodiment, the information acquisition device receives first information sent by the information source device, the first information being transmission-related information of the information source device; the information acquisition device performs a first operation on a target AI unit based on the first information, wherein the target AI unit is associated with the information acquisition device; the first operation includes at least one of the following: model inference, model monitoring, and training data collection. Thus, the information acquisition device performs a first operation on the transmission-related information of the information source device through the AI ​​unit to determine the transmission strategy of at least one of the information acquisition device and the information source device. Compared to high-performance exhaustive search schemes or greedy algorithms, using an AI unit can reduce computational complexity, thereby increasing the probability that scheduling latency meets timeline requirements; furthermore, by outputting decisions through the AI ​​unit, predictions of future channel state information of the terminal can be considered, which can improve transmission performance.

[0142] Optionally, when the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following:

[0143] The first terminal has at least one measurement result in the cell corresponding to the information source device;

[0144] The second terminal has at least one measurement result in the cell corresponding to the information source device;

[0145] The second terminal is in at least one serving beam of the cell corresponding to the information source device;

[0146] The second terminal is in at least one TRP in the cell corresponding to the information source device;

[0147] The second terminal has at least one scheduling information in the cell corresponding to the information source device;

[0148] The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device.

[0149] The user throughput rate of the information source device;

[0150] First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device;

[0151] The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device;

[0152] The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device;

[0153] The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device;

[0154] TRP of candidate information source devices;

[0155] Beams of candidate information source devices;

[0156] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0157] In one implementation, when the first operation includes model inference, the model inference-related information used as model input can be information from a first time period, which is a time period corresponding to the acquisition of input data from the target AI unit. For example, the model inference-related information can include at least one measurement result of the first terminal in the cell corresponding to the information source device at the first time period, and / or at least one measurement result of the second terminal in the cell corresponding to the information source device at the first time period. The information predicted by the model inference result is information from a second time period, which is a time period corresponding to the prediction result of the target AI unit. For example, the target AI unit predicts the modulation and coding information of the target terminal at the second time period.

[0158] A1: In one embodiment, at least one measurement result of the first terminal in the cell corresponding to the information source device may include at least one measurement result obtained by the first terminal performing measurements based on measurement resources sent by the information source device and calculating the corresponding measurement quantities. The measurement resources may be associated with a serving beam or a TRP. The measurement quantities include at least one of the following: CQI, SINR, RSRP, PMI.

[0159] Understandably, based on multiple historical data points (i.e., measurement results of the first terminal in the cell corresponding to the information source device), the information acquisition device can predict the future wireless channel state between the first terminal and the information source device, which is beneficial for making transmission strategy decisions in the future. Furthermore, these historical data points can serve as model inputs for the AI ​​unit on the information acquisition device side. Through model inference, the future wireless channel state between the first terminal and the information source device can be predicted, and / or, at least one of the following can be determined: the second terminal, the first terminal, the TRP of the information acquisition device, and the TRP of the information source device.

[0160] A2: In one embodiment, at least one measurement result of the second terminal in the cell corresponding to the information source device may include at least one reported value obtained by the second terminal performing measurements based on measurement resources sent by the information source device and calculating the corresponding measurement quantities. The measurement resources may be associated with a serving beam or a TRP. The measurement quantities include at least one of the following: CQI, SINR, RSRP, PMI.

[0161] Understandably, based on historical data (i.e., measurement results of the second terminal in the cell corresponding to the information source device), the information acquisition device can predict the future wireless channel state between the second terminal and the information source device, which is beneficial for making transmission strategy decisions in the future. Furthermore, the historical data can serve as model input for the AI ​​unit on the information acquisition device side. Through model inference, the future wireless channel state between the second terminal and the information source device can be predicted, and / or at least one of the following can be determined: the second terminal, the first terminal, the TRP of the information acquisition device, and the TRP of the information source device.

[0162] It should be noted that, in the embodiments of this application, the measurement results may include measured reference signal received power (RSRP), and / or measured reference signal received quality (RSRQ), and / or measured signal-to-noise and interference ratio (SINR), and / or measured CQI, and / or measured PMI, and / or measured RI, etc.

[0163] A3: In one embodiment, at least one serving beam of the cell corresponding to the information source device for the second terminal may include a downlink transmission beam and / or an uplink reception beam provided by the cell where the information source device is located for the second terminal.

[0164] Understandably, based on historical data (i.e., at least one serving beam of the second terminal in the cell corresponding to the information source device), the information acquisition device can predict the most suitable transmit beam and / or uplink receive beam for the information source device to serve the second terminal in the future, which is beneficial for making transmission strategy decisions in the future. Furthermore, this historical data can serve as model input for the AI ​​unit on the information acquisition device side; through model inference, the beam selection for the second terminal in the future can be determined.

[0165] A4: In one embodiment, at least one TRP of the cell corresponding to the information source device for the second terminal may include a downlink transmission and / or an uplink reception TRP provided by the cell where the information source device is located for the second terminal.

[0166] Understandably, based on historical data (i.e., the TRP of the second terminal in the cell corresponding to the information source device), the information acquisition device can predict the most suitable transmit beam and / or uplink receive TRP for the information source device to serve the second terminal in the future. This is beneficial for making transmission strategy decisions in the future. In addition, the historical data can be used as model input for the AI ​​unit on the information acquisition device side. Through model inference, the TRP selection of the second terminal in the future can be determined.

[0167] In one implementation, the serving beam and / or TRP in the model inference-related information can be associated with measurement results in the model inference-related information. For example, the measurement results in the model inference-related information can refer to historical measurement results, which include at least one of the following: RSRP; RSRQ; SINR.

[0168] A5: In one embodiment, the scheduling information of the second terminal in the cell corresponding to the information source device can also be described as the scheduling information of the second terminal in the information source device. Based on multiple historical instances of this information (i.e., the scheduling information of the second terminal in the cell corresponding to the information source device), the information acquisition device can predict a suitable scheduling scheme for the second terminal in the future.

[0169] In one embodiment, at least one scheduling information of the second terminal in the cell corresponding to the information source device is associated with at least one serving beam and / or TRP of the second terminal in the cell corresponding to the information source device. While the information source device sends at least one scheduling information of the second terminal in the cell corresponding to the information source device to the information acquisition device, it may also send the serving beam and / or TRP associated with the scheduling information to the information acquisition device.

[0170] For example, scheduling information can be a modulation and coding scheme (MCS) or a precoding matrix.

[0171] Understandably, based on historical data (i.e., scheduling information of the second terminal in the cell corresponding to the information source device), the information acquisition device can predict the most suitable scheduling scheme (modulation and coding scheme, frequency resource allocation, precoding scheme, etc.) for the second terminal at a future time. Furthermore, this historical data can serve as model input for the AI ​​unit on the information acquisition device side. Through model inference, the future wireless channel state of the second terminal can be determined, and / or, at least one of the following can be selected: the second terminal, the first terminal, the TRP of the information acquisition device, and the TRP of the information source device.

[0172] A6: In one embodiment, the at least one scheduling feedback information of the second terminal in the cell corresponding to the information source device can also be described as at least one scheduling feedback information of the second terminal in the information source device. The scheduling feedback information, for example, may be HARQ information.

[0173] Understandably, based on historical data (i.e., scheduling feedback information of the second terminal in the cell corresponding to the information source device) and channel quality information such as CQI, the information acquisition device can predict the most suitable modulation and coding scheme (e.g., MCS) for the second terminal at a future time. Furthermore, the historical data can serve as model input for the AI ​​unit on the information acquisition device side. Through model inference, the wireless channel state of the second terminal at a future time can be determined, and / or, at least one of the following can be selected: the second terminal, the first terminal, the TRP of the information acquisition device, and the TRP of the information source device.

[0174] A7: In one embodiment, the user throughput rate of the information source device may include at least one of the following:

[0175] The average user throughput rate of the cell where the information source equipment is located;

[0176] The throughput rate of edge users in the cell where the information source device is located;

[0177] The log average user throughput rate of the cell where the information source device is located;

[0178] The historical average transmission rate of the second terminal in the cell where the information source device is located;

[0179] The historical instantaneous transmission rate of the second terminal in the cell where the information source device is located;

[0180] The historical average transmission rate of the first terminal in the cell where the information source device is located;

[0181] The first terminal's historical instantaneous transmission rate in the cell where the information source device is located.

[0182] Understandably, based on historical data (i.e., the user throughput rate of the information source device), the information acquisition device can predict the future transmission rate or user throughput rate of the information source device, or predict the transmission rate of the first terminal and / or the second terminal in the cell where the information source device is located. Furthermore, this historical data can serve as model input to the AI ​​unit on the information acquisition device side. Through model inference, it can be used to determine at least one of the following: the second terminal, the first terminal, the TRP of the information acquisition device, and the TRP of the information source device.

[0183] A8: In one embodiment, the first indication information is used to indicate whether the information source device fully accepts the scheduling information of the information acquisition device.

[0184] Understandably, based on this information (i.e., the first indication information), the information acquisition device can determine the degree of cooperation of the information source device, thereby influencing subsequent scheduling decisions. For example, if the information source device consistently refuses to accept scheduling information from the information acquisition device, the subsequent information acquisition device can reduce the number of terminals offloaded to the information source device, or reduce the number of requests for the information source device's TRP to serve the first terminal. If the information source device consistently accepts the scheduling information from the information acquisition device, the subsequent information acquisition device can maintain or increase the number of terminals offloaded to the information source device, or maintain or increase the number of requests for the information source device's TRP to serve the first terminal.

[0185] A9: In one implementation, in a scenario where the terminal can switch serving cells and TRPs are not shared, the model inference-related information includes the identifier of the second terminal recommended by the information source device to switch to the cell corresponding to the information acquisition device.

[0186] The scenario where terminals can switch serving cells and TRPs are not shared means that the TRP of the information acquisition device only serves the first terminal and not the second terminal; the TRP of the information source device only serves the second terminal and not the first terminal. For the first terminal to obtain service from the TRP of the information source device, it must first switch to the cell where the information source device is located and become the second terminal. For the second terminal to obtain service from the TRP of the information acquisition device, it must first switch to the cell where the information acquisition device is located and become the first terminal.

[0187] It is understandable that when the model inference-related information includes the identifier of the second terminal recommended by the information source device to switch to the cell corresponding to the information acquisition device, it indicates a request from the information source device. The information source device hopes that the information acquisition device can help offload the second terminal for which the information acquisition device provides services. The prerequisite is that the second terminal needs to first switch to the cell where the information acquisition device is located and become the first terminal.

[0188] It should be noted that the scenario in which the terminal can switch serving cells and the TRP is not shared can be associated with the DPS transmission scenario.

[0189] In one implementation, the model inference-related information under DPS transmission includes: an identifier of a second terminal that the information source device recommends switching to the cell corresponding to the information acquisition device.

[0190] A10: In one embodiment, if the first operation result is used to determine the transmission strategy of the information acquisition device in ICBM transmission mode, and / or the transmission strategy of the information source device, then the model inference related information includes: the identifier of the second terminal recommended by the information source device for separate uplink or downlink transmission by the TRP of the information acquisition device.

[0191] Understandably, this information represents a request from the information source device, which requests that the information acquisition device assist in servicing the second terminal. Under ICBM, if the information acquisition device has already performed downlink transmissions for the second terminal, the information source device cannot perform downlink transmissions for that second terminal.

[0192] This information can be used as model input for the AI ​​unit on the information acquisition device side to determine the transmission strategy of the information acquisition device at future times, and / or the transmission strategy of the information source device.

[0193] A11: In one embodiment, if the first operation result is used to determine the transmission strategy of the information acquisition device under the CJT transmission model, and / or the transmission strategy of the information source device, then the model inference related information includes: the identifier of the second terminal recommended by the information source device for joint uplink or downlink transmission by the TRP of the information acquisition device.

[0194] This information represents a request from the information source device, indicating that the information source device requests assistance from the information acquisition device for the second terminal. Under CJT, at least one TRP of the information acquisition device, and / or at least one TRP of the information source device, will send the same data to the second terminal, configured with the same modulation and coding scheme (e.g., MCS). Joint precoding matrices are required between different TRPs to ensure coherent overlay processing of the data streams at the receiving end.

[0195] This information can be used as model input for the AI ​​unit on the information acquisition device side to determine the transmission strategy of the information acquisition device at future times, and / or the transmission strategy of the information source device.

[0196] A12: In one embodiment, the TRP of the candidate information source device includes: a set of TRPs of information source devices that can provide transmission for the second terminal and / or the first terminal at a second time.

[0197] This information characterizes the TRPs that the information source device can provide; this information can be used to determine the second information. When the first information includes this information, the second information requires that at least one TRP used by the information source device for transmission at the second time belongs to the TRPs of the candidate information source device.

[0198] A13: In one embodiment, the beam of the candidate information source device includes: a set of beams of information source devices that can provide transmission for a second terminal and / or a first terminal at a second time.

[0199] This information characterizes the beams that the information source device can provide; this information can be used to determine the second information. When the first information includes this information, the second information requires that at least one beam used by the information source device for transmission at a second time belongs to the beams of the candidate information source device.

[0200] It should be noted that the aforementioned model inference information includes the measurement results (CQI, PMI, RSRP, etc. reported by the second terminal and / or the first terminal at the information source device), and / or the serving beam, and / or the serving TRP, and / or the modulation and coding scheme, RB scheduling, etc. By using the aforementioned model inference information as input to the target AI unit for model inference, at least one of the following can be determined: the TRP of the second terminal, the first terminal, the information acquisition device, and the TRP of the information source device. This allows for optimization of the transmission rate of the cell where the information acquisition device is located, and / or the cell where the information source device is located.

[0201] In one implementation, the model inference-related information further includes at least one of the following:

[0202] The first terminal is in at least one serving beam of the cell corresponding to the information source device;

[0203] The first terminal is in at least one TRP in the cell corresponding to the information source device;

[0204] The first terminal has at least one scheduling information in the cell corresponding to the information source device;

[0205] The first terminal receives at least one scheduling feedback message from the cell corresponding to the information source device.

[0206] The candidate second terminal set includes: a second terminal that can be served by an information acquisition device at a second time, and / or a second terminal that is served by an information source device.

[0207] Optionally, the beam and / or TRP in the model inference-related information are indicated by at least one of the following:

[0208] Transmission Configuration Indicator (TCI) status identifier; TCI identifier; Synchronization Signal Block (SSB) index; Channel State Information Reference Signal (CSI-RS) index; CSI-RS Resource Indicator (CRI).

[0209] Wherein, the beam and / or TRP in the model inference-related information includes at least one of the following:

[0210] The second terminal is in at least one serving beam or TRP of the cell corresponding to the information source device;

[0211] TRP of candidate information source devices;

[0212] Beam of candidate information source device.

[0213] In one implementation, the CSI-RS index may include a type A index.

[0214] It should be noted that the CRI mentioned above refers to the local CRI, for example, CRI 0 to 7. The TCI mentioned above is the global TCI, which can indicate a larger number of TCIs, for example: TCI ID 0-127. The SSB or CSI-RS index mentioned above is the global SSB or CSI-RS index.

[0215] Optionally, if the information acquisition device collects data for training the target AI unit when the information source device is configured in the first configuration, and then changes the configuration of the information source device to the second configuration when the information acquisition device performs model inference for the target AI unit, this will lead to inconsistencies in the data characteristics or attributes of the information source device during the training data collection process and the inference process of the target AI unit.

[0216] In this case, the above problem can be solved by associating the associated ID with the TCI status ID, and / or the TCI ID, and / or the CSI-RSRP resource indicator, and / or the SS / PBCH block resource indicator, and / or the type A index.

[0217] The information acquisition device can assume that the downlink transmission beams or beam sets / lists of information source devices with the same associated ID have similar characteristics.

[0218] The information acquisition device can assume that the TRPs or sets / lists of information source devices with the same associated ID have similar characteristics.

[0219] Optionally, the scheduling information of the second terminal in at least one cell corresponding to the information source device includes at least one of the following:

[0220] Modulation coding information; precoding matrix; allocated frequency band resources.

[0221] In one implementation, the precoding matrix in the scheduling information may include a joint precoding matrix, an independent precoding matrix, a single-user precoding matrix, or a multi-user precoding matrix, etc.

[0222] Optionally, if the first operation includes model inference, the method further includes:

[0223] The information acquisition device sends at least one of second information and second indication information to the information source device. The second indication information is used to indicate that the second information is obtained based on AI unit prediction. The second information is scheduling information related to the transmission of the information source device. The second information is obtained based on the model inference result.

[0224] B1: In this embodiment, the information acquisition device sends a second instruction information to the information source device, so that the information source device can know that the second information is obtained based on AI unit prediction, rather than based on non-AI methods. This helps the information source device to determine whether to adopt the scheduling information corresponding to the second information, and whether to run relevant algorithms on the information source device side to obtain the corresponding scheduling information.

[0225] B2: In this embodiment, the information acquisition device sends second information to the information source device, so that the information source device can learn about the transmission-related scheduling information of the information source device recommended by the information acquisition device through the second information, without having to run an algorithm on the information source device side to obtain the corresponding scheduling information, thereby saving the computing power overhead on the information source device side.

[0226] Optionally, the second information includes the model inference result, which includes at least one of the following:

[0227] The identifier of the target terminal; the TRP corresponding to the target terminal; the beam corresponding to the target terminal; the precoding matrix corresponding to the target terminal; the modulation and coding information corresponding to the target terminal; the predicted channel quality information corresponding to the target terminal; and the first predicted measurement.

[0228] The target terminal includes at least one of the following:

[0229] It is recommended to switch to the first terminal in the cell corresponding to the information source device; it is recommended to switch to the second terminal in the cell corresponding to the information acquisition device; it is recommended that the second terminal be scheduled by the information source device; it is recommended that the first terminal be scheduled by the information acquisition device.

[0230] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0231] In one embodiment, the target terminal can be considered a selected terminal, and the terminal to be scheduled can be determined by predicting the target terminal; the TRP corresponding to the target terminal can be considered a selected TRP, and the TRP transmitted with the target terminal can be determined by predicting the TRP corresponding to the target terminal; the beam used by the target terminal can be determined by predicting the beam corresponding to the target terminal; the precoding matrix used for transmission by the target terminal can be determined by predicting the precoding matrix corresponding to the target terminal; the modulation and coding information used for transmission by the target terminal can be determined by predicting the modulation and coding information used for transmission by the target terminal; and the transmission strategy of at least one of the information acquisition device and the information source device can be determined by predicting the channel quality information corresponding to the target terminal.

[0232] In one implementation, the second information may include a predicted first measurement. Sending the predicted first measurement to the information source device via the information acquisition device helps the information source device determine relevant scheduling information based on this information. For example, if the predicted first measurement is the CQI predicted by the target AI unit as reported by the second terminal to the information source device at a second time, then the information source device can determine the MCS of the second terminal at the second time based on this information. As another example, if the predicted first measurement is the PMI predicted by the target AI unit as reported by the first terminal to the information source device at a second time, then the information source device can determine the PMI of the first terminal at the second time based on this information. The terminal associated with the predicted first measurement may include the candidate second terminals in the first information, and / or the first terminal that the information acquisition device hopes the information source device will assist in servicing.

[0233] Optionally, the second information further includes at least one of the following:

[0234] The information acquisition device can provide an independently pre-encoded identifier for the second terminal;

[0235] The information acquisition device can provide a jointly pre-coded identifier for the second terminal;

[0236] The identifier of the first terminal is recommended to be independently pre-encoded by the information source device;

[0237] The identifier of the first terminal is recommended to be jointly pre-coded by the information source device.

[0238] In one implementation, in a scenario where TRP can be shared, the second information includes the aforementioned information.

[0239] In one implementation, under CJT transmission mode, the second information further includes at least one of the following:

[0240] At a second time, the information acquisition device may provide a jointly precoded identifier of the second terminal;

[0241] At the second time, it is recommended that the identifier of the first terminal be jointly pre-coded by the information source device.

[0242] In one implementation, in NCJT transmission mode, the second information further includes at least one of the following:

[0243] At a second time, the information acquisition device can provide an independently pre-coded identifier of the second terminal;

[0244] In the second time, it is recommended that the identifier of the first terminal be independently pre-encoded by the information source device;

[0245] C1: Wherein, the information acquisition device can provide an independently pre-coded identifier of the second terminal; it can also be understood that at a second time, the information acquisition device can provide an independently pre-coded identifier of the second terminal; the second information includes the identifier indicating that at a second time, the information acquisition device can provide data transmission to the second terminal of the information source device in an NCTJ manner.

[0246] C2: Wherein, the information acquisition device can provide a jointly precoded identifier of the second terminal; it can also be understood that at a second time, the information acquisition device can provide a jointly precoded identifier of the second terminal; the second information includes the identifier indicating that at a second time, the information acquisition device can provide data transmission to the second terminal of the information source device in a CTJ manner.

[0247] C3: Wherein, the identifier of the first terminal recommended to be independently pre-encoded by the information source device; it can also be understood as the identifier of the first terminal recommended to be independently pre-encoded by the information source device at the second time; the second information includes the identifier indicating that at the second time, the information acquisition device requests the information source device to provide data transmission to the first terminal in the manner of NCJT.

[0248] C4: Wherein, the identifier of the first terminal recommended to be jointly precoded by the information source device; it can also be understood as the identifier of the first terminal recommended to be jointly precoded by the information source device at a second time; the second information includes the identifier indicating that at a second time, the information acquisition device requests the information source device to provide data transmission to the first terminal in the CJT manner.

[0249] In one implementation, the second information further includes: modulation and coding information; a precoding matrix; and allocated frequency band resources. For example, the second information also includes the MCS of the UE to be scheduled, the allocated RBs, and the PMI.

[0250] In this embodiment of the application, by interacting with the second information, the target first terminal and the target second terminal at the second time can be determined by the information acquisition device, thereby saving the computing power overhead of the information source device for determining the target terminal; in addition, by interacting with the second information, joint scheduling between multiple cells is enabled, which can better improve the throughput at the cell edge compared with independent scheduling of a single cell.

[0251] Optionally, when the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following:

[0252] The first quantifiable quantity measured; the first key performance indicator (KPI); the parameters associated with the first KPI; and the scheduling information of the second terminal in the cell corresponding to the information source device.

[0253] The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device.

[0254] The serving cell of the second terminal is the cell corresponding to the information source device.

[0255] D1: The first measured quantity is a value calculated by the second terminal or the first terminal based on the measurement resource measurement sent by the information source device. The first measured quantity characterizes the radio channel state of the cell where the information source device is located, including at least one of the following: CQI, SINR, PMI, RI, RSRP, RSRQ, RSSI.

[0256] The first measured quantity can be used as a label or ground truth for the first measured quantity to be predicted, thereby generating training samples and improving the prediction accuracy of the target AI unit.

[0257] D2: The first key performance indicator (KPI) can be used by the information acquisition device to judge the inference performance of the target AI unit, thereby for model supervision, and then to determine whether the target AI unit should revert to a non-AI algorithm, activate, deactivate, switch, or fine-tune, etc. on the information acquisition device side.

[0258] D2-1: The prediction accuracy of the first measurement, such as the CQI prediction accuracy or SINR prediction accuracy, can be used by the information acquisition device to judge the inference performance of the target AI unit, thereby for model supervision, and further to determine whether the target AI unit should fall back to a non-AI algorithm, or be activated, or deactivated, or switched, or fine-tuned, etc., on the information acquisition device side.

[0259] Furthermore, during reinforcement learning, this information can serve as environmental feedback to determine the reward for reinforcement learning.

[0260] D2-2: In one embodiment, the transmission rate information of the second terminal may include at least one of the following:

[0261] The instantaneous transmission rate of the second terminal;

[0262] The cumulative transmission rate of the second terminal;

[0263] The average transmission rate of at least one second terminal within the information source device.

[0264] The average transmission rate of at least one second terminal within the information source device can refer to the average transmission rate of at least one second terminal within the cell corresponding to the information source device, i.e., the average transmission rate over a period of time, and can be characterized by at least one of the following: ∑ k log2r k ;

[0265] Where, r kIt is the instantaneous transmission rate of the kth second terminal in the cell corresponding to the information source device at a certain moment;

[0266] It is the historical transmission rate of the kth second terminal in the cell corresponding to the information source device before a certain moment;

[0267] α: Weighted value of instantaneous velocity;

[0268] β: Weighted value of historical rates.

[0269] The information acquisition device can use the aforementioned information to monitor the reliability of the inference results of the target AI unit. For example, if the accuracy of the predicted first measurement (e.g., CQI, SINR) is high, but the throughput rate of the information source device remains low for an extended period, it may indicate a problem with terminal scheduling or TRP scheduling. Furthermore, during reinforcement learning, this information can serve as environmental feedback to determine the reward for reinforcement learning.

[0270] D2-3: The scheduling fairness of the information source device is an indicator for evaluating the scheduling fairness among multiple users of the information source device. A common indicator is the proportional fairness factor (PF), which is an algorithm used in wireless communication systems for scheduling and resource allocation, aiming to balance system throughput and user fairness.

[0271] The information acquisition device obtains the aforementioned information and can use it to monitor the reliability of the inference results of the target AI unit. Furthermore, during reinforcement learning, this information can serve as environmental feedback to determine the reward for reinforcement learning.

[0272] D3: In one implementation, the parameters associated with the first KPI may include: the duration of observation, the transmission time interval (TTI) of the observation, the number of samples observed, the weighted value of the instantaneous rate, and / or the weighted value of the historical rate, etc.

[0273] D3-1: The observation duration is the duration for calculating the first KPI. Different observation durations affect the frequency and accuracy of the information acquisition device acquiring the first KPI. A longer observation duration results in a larger sample size, leading to higher prediction accuracy and a higher probability of the information acquisition device acquiring the first KPI more frequently. Conversely, a shorter observation duration results in a smaller sample size, leading to lower prediction accuracy and a lower probability of the information acquisition device acquiring the first KPI less frequently. Based on this information, the information acquisition device can determine the reliability of the first KPI.

[0274] D3-2: Observation TTI: Similar to the duration of observation, but using TTI as the unit of time.

[0275] D3-2: Number of Observed Samples: The number of observed samples affects the frequency and accuracy at which the information acquisition device acquires the first KPI. A larger number of observed samples and a longer statistical timeframe result in higher prediction accuracy, and the information acquisition device is likely to acquire the first KPI more frequently. Conversely, a smaller number of observed samples and a shorter statistical timeframe result in lower prediction accuracy, and the information acquisition device is likely to acquire the first KPI less frequently. Based on this information, the information acquisition device can determine the reliability of the first KPI.

[0276] D3-3: Weighted value of instantaneous rate / weighted value of historical rate, reflecting the relative importance of the information source device to the user's instantaneous rate and user fairness.

[0277] The information acquisition device can use this information to optimize the training process, such as optimizing the design of the loss function. Therefore, customized weights can be configured for different information source devices to match the needs of different information source devices.

[0278] D4: The scheduling information of the second terminal in the cell corresponding to the information source device may or may not match the second information. When the information matches the second information, it indicates that the information source device has fully accepted the scheduling decision recommended by the information acquisition device; when the information does not partially match the second information, it indicates that the information source device has only accepted part of the scheduling decision recommended by the information acquisition device; when the information does not match the second information at all, it indicates that the information source device has not accepted the scheduling decision recommended by the information acquisition device.

[0279] Based on this information, the information acquisition device can determine the degree of cooperation of the information source device. This allows it to decide whether to continue joint multi-TRP transmission with the information source device, or whether to continue providing TRP transmission strategy recommendations to the information source device.

[0280] In one implementation, the first operation includes model inference and / or model monitoring, and the first information includes model inference-related information and / or model monitoring-related information. The model inference-related information and / or model monitoring-related information can be sent via different signaling methods. The information acquisition device performs model inference operations on the target AI unit based on the model inference-related information to obtain model inference results; the information acquisition device determines transmission-related scheduling information of the information source device based on the model inference results and sends the transmission-related scheduling information of the information source device to the information source device, wherein the transmission strategy of the information source device includes the transmission-related scheduling information of the information source device. After performing the model inference operation, the information acquisition device performs model monitoring operations on the target AI unit based on the model monitoring-related information to obtain model monitoring results; the information acquisition device determines, based on the model monitoring results, whether the target AI unit should revert to a non-AI algorithm, activate, deactivate, switch, or fine-tune, etc., on the information acquisition device side, and perform subsequent operations.

[0281] It should be noted that the aforementioned model monitoring information includes the measured first measurement quantity, the first key performance indicator (KPI), the parameters associated with the first KPI, and / or the scheduling information of the second terminal in the cell corresponding to the information source device. Based on the aforementioned model monitoring information, the model monitoring result of the target AI unit is determined. Based on the model monitoring result, the information acquisition device determines on the information acquisition device side whether the target AI unit should revert to a non-AI algorithm, activate, deactivate, switch, or fine-tune, etc., to achieve better transmission performance.

[0282] Optionally, when the first operation includes training data collection, the first information includes at least one of model inference-related information and model monitoring-related information;

[0283] The information acquisition device performs a first operation on the target AI unit based on the first information, including any one of the following:

[0284] The information acquisition device generates a dataset based on the model inference-related information and the model monitoring-related information;

[0285] The information acquisition device performs model inference of the target AI unit based on the model inference-related information, and determines the environmental feedback or reward value of the target AI unit based on the model monitoring-related information; the information acquisition device generates a dataset based on the model inference result of the target AI unit and the environmental feedback or reward value.

[0286] E1: The information acquisition device generates a dataset based on the model inference-related information and the model monitoring-related information. This can also be understood as the information acquisition device generating a supervised learning training dataset based on the model input and ground truth / labels of the target AI unit. The model input of the target AI unit corresponds to the model inference-related information, and the ground truth / labels of the target AI unit correspond to the model monitoring-related information.

[0287] The information acquisition device can simultaneously acquire the model input and truth value / label of the target AI unit; or acquire the model input and truth value / label of the target AI unit separately using different first information.

[0288] E2: The information acquisition device performs model inference for the target AI unit based on the model inference-related information, and determines the environmental feedback or reward value of the target AI unit based on the model monitoring-related information; the information acquisition device generates a dataset based on the model inference result of the target AI unit and the environmental feedback or reward value. This can also be understood as, in reinforcement learning, the agent determines the action based on the state and obtains the environmental feedback or reward value. Then, a training dataset for reinforcement learning is generated based on the state, action, reward, and next state. Here, the state corresponds to the model inference-related information in a first piece of information obtained from the information source device, and the reward is determined based on the model monitoring-related information in another first piece of information. The action is sent from the information acquisition device to the information source device through a second piece of information.

[0289] Based on the above process, the information acquisition device can acquire the reinforcement learning training data of the target AI unit.

[0290] Optionally, the method further includes:

[0291] The information acquisition device sends third information to the information source device;

[0292] Alternatively, the information acquisition device may send third information to the information source device and receive confirmation information of the third information sent by the information source device;

[0293] The third information is used to request the first information;

[0294] The confirmation information is used to instruct the information source device to agree to the request corresponding to the third information.

[0295] In this embodiment, the information acquisition device sends third information to the information source device, the third information being used to request the first information, thereby enabling the information source device to send the first information to the information acquisition device based on the request of the third information.

[0296] Optionally, the third information includes at least one of the following: the service type that the information acquisition device can provide to the second terminal, and the service type that the information acquisition device requests from the information source device;

[0297] Wherein, the serving cell of the second terminal is the cell corresponding to the information source device;

[0298] The service type includes at least one of the following: DPS; ICBM; CJT; NCJT.

[0299] F1: The service type that the information acquisition device can provide to the second terminal is DPS. This can be understood as the information acquisition device providing assistance to the information source device in DPS transmission mode. That is, if the information source device wants the second terminal to be served by the information acquisition device in DPS mode, then the second terminal needs to switch to the cell where the information acquisition device is located and become the first terminal before the second time, in order to be served by the information acquisition device in DPS mode.

[0300] F2: The service type requested by the information acquisition device from the information source device is DPS. This can be understood as the information acquisition device wanting the first terminal to be served by the information source device in the second time using the DPS transmission method. That is, the first terminal will switch to the information source device in the second time and become the second terminal, so that the data transmission is carried out by the TRP of the information source device.

[0301] F3: The information acquisition device can provide the second terminal with an ICBM service type. This can be understood as the information acquisition device assisting the information source device in the ICBM transmission mode. That is, if the information source device wants the second terminal to be served by the information acquisition device in the ICBM manner, then the second terminal does not need to become the first terminal in the cell where the information acquisition device is located at a second time to be served by the information acquisition device in the ICBM manner.

[0302] F4: The service type requested by the information acquisition device from the information source device is ICBM. This can be understood as the information acquisition device wanting the first terminal to be served by the information source device in the second time using ICBM transmission mode. That is, the first terminal does not need to switch to the information source device in the second time, and the data can be transmitted by the TRP of the information source device.

[0303] F5: The information acquisition device can provide the second terminal with a service type of CJT. This can be understood as the information acquisition device providing assistance to the information source device in the CJT transmission mode. That is, if the information source device wants the second terminal to be served by the information acquisition device in the CJT manner, then at the second time, the information acquisition device and the information source device need to perform collaborative precoding and send the same data, and perform clock synchronization to ensure that the data stream is coherently superimposed on the second terminal.

[0304] F6: The service type requested by the information acquisition device from the information source device is CJT. This can be understood as the information acquisition device wanting the first terminal to be served by the information source device in the second time using the CJT transmission method. That is, in the second time, the information acquisition device and the information source device need to perform collaborative precoding, send the same data, and synchronize their clocks to ensure that the data stream is coherently superimposed on the first terminal.

[0305] F7: The information acquisition device can provide the second terminal with a service type of NCJT. This can be understood as the information acquisition device providing assistance to the information source device in the NCJT transmission mode. That is, if the information source device wants the second terminal to be served by the information acquisition device in the NCJT manner, then at the second time, the information acquisition device and the information source device can perform independent precoding and send different data to the second terminal.

[0306] F8: The service type requested by the information acquisition device from the information source device is NCJT. This can be understood as the information acquisition device wanting the first terminal to be served by the information source device in the second time using the NCJT transmission method. That is, in the second time, the information acquisition device and the information source device can perform independent precoding and can send different data to the first terminal.

[0307] Based on this information, the information source device can know the TRP transmission model that the information acquisition device can provide, and / or, the information source device can know the TRP transmission mode that the information acquisition device expects the information source device to provide when cooperating.

[0308] In one implementation, when the first operation result is used to determine the transmission strategy of at least one of the information acquisition device and the information source device under DPS transmission, the service type includes DPS.

[0309] In one embodiment, when the first operation result is used to determine the transmission strategy of at least one of the information acquisition device and the information source device under ICBM transmission, the service type includes ICBM;

[0310] In one implementation, when the first operation result is used to determine the transmission strategy of at least one of the information acquisition device and the information source device under DPS transmission, the service type includes DPS.

[0311] In this embodiment, the information acquisition device sends third information to the information source device, which facilitates the information source device to provide feedback to the information acquisition device based on the third information, and provides model inference-related information corresponding to the service type that the information acquisition device can provide or request, which helps the target AI unit predict the transmission strategy.

[0312] In this embodiment of the application, the first information to the third information, as well as the confirmation information of the third information, can be directly interacted between the information acquisition device and the information source device, or forwarded through the target device.

[0313] The target device can be a gNB-CU or a UE.

[0314] G1. As shown in Figure 4a, when the information acquisition device is a first network device, the information source device is a second network device, and the target AI unit is a first AI unit, the relevant operations for the first AI unit on the first network device side are as follows:

[0315] G1-1: Sending confirmation messages for the first and third messages.

[0316] The information source device ID contained in the above information (i.e., the confirmation information of the first information and the third information), or Source gNB-DU ID = the cell ID where the second network device is located, for example, the base station distributed unit ID (DU ID) corresponding to the second network device;

[0317] The information obtained above includes the device ID, which is equal to the cell ID where the first network device is located.

[0318] The first terminal = the UE in the cell where the information acquisition device is located, that is, the UE in the cell where the first network device is located;

[0319] The second terminal is the UE in the cell where the information source device is located, that is, the UE in the cell where the second network device is located;

[0320] For example, the information source device (in this case, the second network device) first sends the aforementioned information to the base station centralized unit (gNB-CU), and then the gNB-CU forwards the information to the information acquisition device (in this case, the first network device).

[0321] For example, the information source device (in this case, the second network device) first sends the aforementioned information to the first terminal, and then the first terminal forwards the information to the information acquisition device (in this case, the first network device).

[0322] For example, the information source device (in this case, the second network device) first sends the aforementioned information to the second terminal, and then the second terminal forwards the information to the information acquisition device (in this case, the second network device).

[0323] G1-2: Regarding the second and third pieces of information:

[0324] The information source device ID contained in the above information (i.e., the second information and the third information), or Source gNB-DU ID = the cell ID where the second network device is located, for example, the base station distributed unit ID (DU ID) corresponding to the second network device;

[0325] The information obtained above includes the device ID, which is equal to the cell ID where the first network device is located.

[0326] The first terminal = the UE in the cell where the information acquisition device is located, that is, the UE in the cell where the first network device is located;

[0327] The second terminal is the UE in the cell where the information source device is located, that is, the UE in the cell where the second network device is located;

[0328] For example, the information acquisition device (in this case, the first network device) first sends the aforementioned information to the gNB-CU, and then the gNB-CU forwards the information to the information source device (in this case, the second network device).

[0329] For example, the information acquisition device (in this case, the first network device) first sends the aforementioned information to the first terminal, and then the first terminal forwards the information to the information source device (in this case, the second network device).

[0330] For example, the information acquisition device (in this case, the first network device) first sends the aforementioned information to the second terminal, and then the second terminal forwards the information to the information source device (in this case, the second network device).

[0331] G2. As shown in Figure 4a, when the information acquisition device is the second network device, the information source device is the first network device, and the target AI unit is the second AI unit, the relevant operations for the second AI unit on the second network device side are as follows:

[0332] G2-1: Sending confirmation messages for the first and third messages.

[0333] The information source device ID contained in the above information (i.e., the confirmation information of the first information and the third information), or Source gNB-DU ID = the cell ID where the first network device is located, for example, the base station distributed unit ID (DU ID) corresponding to the first network device;

[0334] The information obtained above includes the device ID, which is equal to the cell ID where the second network device is located.

[0335] The first terminal is the UE in the cell where the information acquisition device is located, which is the UE in the cell where the second network device is located;

[0336] The second terminal is the UE in the cell where the information source device is located, that is, the UE in the cell where the first network device is located;

[0337] For example, the information source device (in this case, the first network device) first sends the aforementioned information to the gNB-CU, and then the gNB-CU forwards the information to the information acquisition device (in this case, the second network device).

[0338] For example, the information source device (in this case, the first network device) first sends the aforementioned information to the first terminal, and then the first terminal forwards the information to the information acquisition device (in this case, the second network device).

[0339] For example, the information source device (in this case, the first network device) first sends the aforementioned information to the second terminal, and then the second terminal forwards the information to the information acquisition device (in this case, the second network device).

[0340] G1-2: Regarding the second and third pieces of information:

[0341] The information source device ID contained in the above information (i.e., the second information and the third information), or Source gNB-DU ID, is the cell ID where the first network device is located. For example, the Distributed Unit Identifier (DU ID) of the base station corresponding to the first network device.

[0342] The information obtained above includes the device ID, which is equal to the cell ID where the second network device is located.

[0343] The first terminal is the UE in the cell where the information acquisition device is located, which is the UE in the cell where the second network device is located;

[0344] The second terminal is the UE in the cell where the information source device is located, that is, the UE in the cell where the first network device is located;

[0345] For example, the information acquisition device (in this case, the second network device) first sends the aforementioned information to the gNB-CU, and then the gNB-CU forwards the information to the information source device (in this case, the first network device).

[0346] For example, the information acquisition device (in this case, the second network device) first sends the aforementioned information to the first terminal, and then the first terminal forwards the information to the information source device (in this case, the first network device).

[0347] For example, the information acquisition device (in this case, the second network device) first sends the aforementioned information to the second terminal, and then the second terminal forwards the information to the information source device (in this case, the first network device).

[0348] Optionally, the first terminal includes at least one of the following:

[0349] The serving cell is the cell corresponding to the information acquisition device, and the network device corresponding to the downlink transmission is the terminal of the information acquisition device;

[0350] The serving cell is the cell corresponding to the information acquisition device, and the network device corresponding to the downlink transmission is the terminal of the information source device;

[0351] And / or,

[0352] The second terminal includes at least one of the following:

[0353] The serving cell is the cell corresponding to the information source device, and the network device corresponding to the downlink transmission is the terminal of the information source device;

[0354] The serving cell is the cell corresponding to the information source device, and the network device corresponding to the downlink transmission is the terminal of the information acquisition device.

[0355] In one implementation, the first terminal satisfies the following conditions: the serving cell is the cell where the information acquisition device is located, and uplink and downlink transmissions are performed only by the TRP of the information acquisition device.

[0356] In one implementation, the second terminal satisfies the following conditions: the serving cell is the cell where the information source device is located, and uplink and downlink transmissions are performed solely by the TRP of the information source device; or,

[0357] The second terminal meets the following conditions: the serving cell is the cell where the information acquisition device is located, and uplink and downlink transmission is performed only by the TRP of the information source device.

[0358] Optionally, the first operation result is used to determine the transmission strategy of at least one of the information acquisition device and the information source device under Dynamic Node Selection (DPS) transmission; or,

[0359] The first operation result is used to determine the transmission strategy of at least one of the information acquisition device and the information source device under inter-cell beam-managed ICBM transmission; or,

[0360] The first operation result is used to determine the transmission strategy of at least one of the information acquisition device and the information source device under Coherent Joint Transmission (CJT); or,

[0361] The first operation result is used to determine the transmission strategy of at least one of the information acquisition device and the information source device under non-coherent joint transmission (NCJT).

[0362] Currently, the complexity of multi-user pairing and precoding technologies within a single cell is a significant challenge. Multi-TRP transmission technologies can be understood as multi-user pairing and precoding technologies on a larger scale, further increasing complexity and making implementation more difficult. Multi-TRP transmission technologies include DPS, ICBM, CJT, and NCJT. Through multi-cell joint scheduling, multi-TRP technology can improve the transmission performance of UEs at the cell edge. However, issues such as multi-TRP selection, user selection, and precoding in multi-TRP transmission suffer from excessive complexity.

[0363] When performing multiple TRP transmissions, network devices can follow a primary-secondary mode or a peer-to-peer mode, and / or use a peer-to-peer mode for transmission.

[0364] As shown in Figures 4b and 4c, in the master-slave mode, the cells are divided into master cells and slave cells. The master cell simultaneously considers the user and TRP states within both the master and slave cells to determine the scheduling status of UEs and / or TRPs within each cell, and then instructs the slave cell on the scheduling results. The master-slave mode can also be called the master-slave mode, or the team leader-team member mode. In this mode, the master cell can combine information from multiple cells for joint scheduling, thereby improving the throughput of users at the cell edge; while the slave cell, not needing to perform related scheduling calculations, reduces its computing power requirements. The master-slave mode is well-suited for scenarios where base station deployment computing power is uneven.

[0365] In peer-to-peer mode, it is assumed that each of the multiple cells deploys an intelligent agent (or AI unit). Each intelligent agent performs inference from the AI ​​unit to obtain the selection result for multiple TRPs. Peer-to-peer mode can also be called equal mode, game mode, or alternating mode. Peer-to-peer mode is more suitable for multi-TRP transmission among multiple cells with evenly distributed computing power, or for multi-TRP transmission between different vendors.

[0366] This application proposes a multi-agent-based multi-TRP transmission method, which can reduce the complexity of multi-TRP transmission and improve the transmission rate based on AI.

[0367] The following examples provide further explanation:

[0368] Example 1: Primary-Secondary Mode

[0369] Example 0:

[0370] This example illustrates the scheduling functions included in the leader cell and member cell in the leader-member mode.

[0371] This example assumes that the multi-cell TRP selection is a centralized reasoning process, where a certain cell is the leader cell (or master cell), which determines the selection results of the multi-cell TRP for itself and the remaining cells, and notifies the remaining cells (member cells, or auxiliary cells, or subordinate cells). This mode (such as Mode 1 and Mode 2 in Table 1 below) avoids the inability of multiple cells to select TRPs separately, which would prevent effective coordination of inter-cell interference, thus obtaining the optimal solution over a larger range.

[0372] Table 1

[0373] Example 1-1: Supervised Learning - Captain-Member Pattern - Model Inference Flow - DPS

[0374] Example 1-1 is based on the assumptions of Example 0, further assuming that DPS transmission is used between multiple cells.

[0375] As shown in Figures 5a and 5b, the operation execution method includes the following process:

[0376] Step (1a) (optional): The first network device (in this example, an information acquisition device) sends the first target information (i.e., the third information) to the second network device (in this example, an information source device). The first target information is used to request relevant information required for AI unit inference.

[0377] Optionally, the relevant information required for AI unit inference associated with the first target information includes at least one of the following:

[0378] At least one measurement result of UE1 in the cell corresponding to the second network device (hereinafter referred to as at least one measurement result of UE1 in the second network device), wherein the serving cell of UE1 is the cell where the first network device is located. For example, under DPS, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0379] At least one measurement result of UE2 in the cell corresponding to the second network device (hereinafter referred to as at least one measurement result of UE2 in the second network device), for example, this option is carried under DPS;

[0380] UE2 carries at least one serving beam and / or TRP of the cell corresponding to the second network device (hereinafter referred to as at least one serving beam and / or TRP of UE2 in the second network device), for example, under DPS.

[0381] At least one scheduling information of UE2 in the cell corresponding to the second network device (hereinafter referred to as at least one scheduling information of UE2 in the second network device), for example, this option is carried under DPS;

[0382] At least one scheduling feedback information of UE2 in the cell corresponding to the second network device (hereinafter referred to as at least one scheduling feedback information of UE2 in the second network device), such as HACK feedback. For example, if the first network device cannot obtain HACK feedback under DPS, this option is carried.

[0383] Timestamp.

[0384] Wherein, at least one serving beam and / or TRP of UE2 in the second network device can be indicated by at least one of the following:

[0385] TCI state ID; TCI ID; SSB index; CSI-RS index; Type A index; CRI.

[0386] It should be noted that the CRIs mentioned above are local CRIs, for example, CRIs 0-7. The TCIs mentioned above are global TCIs, indicating a wider range of TCIs, for example: TCI IDs 0-127. The network can translate CRIs into TCIs, meaning there is a mapping relationship between CRIs and TCIs. CRIs are not global CRIs, and the UE is aware of this mapping relationship. The SSB or CSI-RS indexes mentioned above are global SSB or CSI-RS indexes.

[0387] Wherein, at least one of the scheduling information of UE2 in the second network device includes at least one of the following:

[0388] Allocated MCS; precoding matrix; allocated frequency band resources (e.g., RB information).

[0389] At least one scheduling information of UE2 in the second network device can be associated with at least one serving beam and / or TRP of UE2 in the second network device.

[0390] Additionally, the measurement results in the AI ​​unit inference information can refer to historical measurement results, which include at least one of the following: RSRP; RSRQ; SINR.

[0391] Step (1b): The first network device receives second target information (i.e., first information) from the second network device. The second target information includes AI unit inference-related information, including at least one of the following:

[0392] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under DPS, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0393] UE2 carries at least one measurement result from the second network device, for example, under DPS;

[0394] UE2 carries at least one serving beam and / or TRP of the second network device, for example, under DPS;

[0395] UE2 carries at least one scheduling information from the second network device, for example, under DPS;

[0396] UE2 carries at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, if the first network device cannot obtain HACK feedback under DPS, this option is carried.

[0397] Timestamp.

[0398] Step (2): The first network device performs inference using its AI unit to obtain third target information. The third target information includes scheduling information from the first network device and at least one second network device, and this scheduling information includes at least one of the following:

[0399] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp; UE switching to the second network device; UE switching from the second network device to the first network device.

[0400] For example, the model inputs of the AI ​​unit include at least one of the following: cell ID; TRP ID / beam ID (such as historical service TRP and / or beam (BEAM)); measurement (such as RSRP, RSRQ, SINR, CQI, PMI, MCS and / or HARQ).

[0401] The TRP ID / beam ID can be represented by TCI ID, SSB index, CSI-RS ID (type A index), or CRI.

[0402] Step (3a): The first network device sends fourth target information to UE1, the fourth target information including a handover command. The handover command is determined based on the third target information.

[0403] For example, if the selected UE includes UE1, and the selected TRP belongs to the second network device, then the handover target cell of the fourth target information is the corresponding second network device.

[0404] Step (3b): The first network device sends fifth target information (i.e., second information) to the second network device. The fifth target information is determined based on the third target information and includes at least one of the following:

[0405] Switching command; sixth target information; first target indication (i.e., second indication information).

[0406] The first target indication is used to indicate the sixth target information and / or switching command, which is obtained based on AI prediction.

[0407] The sixth target information is scheduling information related to the second network device, including at least one of the following:

[0408] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp.

[0409] Step (4): The second network device sends seventh target information to UE2, the seventh target information including a handover command. The handover command is determined based on the fifth target information.

[0410] For example, if the selected UE includes UE2, and the selected TRP belongs to the first network device, then the handover target cell of the seventh target information is the first network device.

[0411] Example 1-2: Supervised Learning - Captain-Member Pattern – Model Inference Flow - ICBM

[0412] Example 1-2 is based on the assumptions of Example 0, and further assumes that ICBM is used for transmission between multiple cells.

[0413] As shown in Figures 6a and 6b, the operation execution method includes the following process:

[0414] Step (1a) (optional): The first network device (in this example, an information acquisition device) sends the first target information (i.e., the third information) to the second network device (in this example, an information source device). The first target information is used to request relevant information required for AI unit inference.

[0415] Optionally, the relevant information required for AI unit inference associated with the first target information includes at least one of the following:

[0416] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under ICBM, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0417] At least one measurement result of UE2 in the second network device. For example, under ICBM, the first network device can obtain at least one measurement result of UE2 in the second network device. In this case, this option may not be carried.

[0418] UE2 has at least one serving beam and / or TRP in the second network device;

[0419] At least one scheduling information of UE2 in the second network device;

[0420] UE2 receives at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, under ICBM, the first network device can obtain the scheduling feedback message, in which case this option may not be carried.

[0421] Timestamp.

[0422] Step (1b): The first network device receives second target information (i.e., first information) from the second network device. The second target information includes relevant information required for AI unit inference, including at least one of the following:

[0423] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under ICBM, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0424] At least one measurement result of UE2 in the second network device. For example, under ICBM, the first network device can obtain at least one measurement result of UE2 in the second network device. In this case, this option may not be carried.

[0425] UE2 has at least one serving beam and / or TRP in the second network device;

[0426] At least one scheduling information of UE2 in the second network device;

[0427] UE2 receives at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, under ICBM, the first network device can obtain the scheduling feedback message, in which case this option may not be carried.

[0428] Timestamp.

[0429] Step (2): The first network device performs inference using its AI unit to obtain third target information. The third target information includes scheduling information from the first network device and at least one second network device, and this scheduling information includes at least one of the following:

[0430] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp; UE switching to the second network device; UE switching from the second network device to the first network device.

[0431] Step (3): The first network device sends fifth target information (i.e., second information) to the second network device. The fifth target information is determined based on the third target information and includes at least one of the following (the difference from step (3b) in Example 1-1 is that it does not include the handover command):

[0432] Sixth objective information; First objective indication.

[0433] The first target indication is used to indicate the sixth target information and / or switching command, which is obtained based on AI prediction.

[0434] The sixth target information is scheduling information related to the second network device, including at least one of the following:

[0435] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp.

[0436] Example 1-3: Supervised Learning - Captain-Member Pattern - Model Inference Flow - CJT

[0437] Examples 1-3 are based on the assumptions of Example 0, further assuming that CJT transmission is used between multiple cells.

[0438] As shown in Figures 7a and 7b, the operation execution method includes the following process:

[0439] Step (1a) (optional): The first network device (in this example, an information acquisition device) sends first target information (i.e., third information) to the second network device (in this example, a data source device). The first target information is used to request relevant information required for AI unit inference.

[0440] Optionally, the relevant information required for AI unit inference associated with the first target information includes at least one of the following:

[0441] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under CJT, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0442] UE2 carries at least one measurement result from the second network device, for example, under CJT;

[0443] UE2 carries at least one serving beam and / or TRP of the second network device, for example, under CJT;

[0444] UE2 has at least one scheduling information in the second network device. For example, under CJT, the first network device can obtain the MCS allocated to UE2 in the second network device. In this case, the first target information may not request the allocated MCS.

[0445] UE2 carries at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, if the first network device cannot obtain HACK feedback under CJT, this option is carried.

[0446] Timestamp.

[0447] Step (1b): The first network device receives second target information (i.e., first information) from the second network device. The second target information includes relevant information required for AI unit inference, including at least one of the following:

[0448] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under CJT, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0449] UE2 carries at least one measurement result from the second network device, for example, under CJT;

[0450] UE2 carries at least one serving beam and / or TRP of the second network device, for example, under CJT;

[0451] UE2 has at least one scheduling information in the second network device. For example, under CJT, the first network device can obtain the MCS assigned to UE2 in the second network device. In this case, the second target information may not carry the assigned MCS.

[0452] UE2 carries at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, if the first network device cannot obtain HACK feedback under CJT, this option is carried.

[0453] Timestamp.

[0454] Step (2): The first network device performs inference using its AI unit to obtain third target information. The third target information includes scheduling information from the first network device and at least one second network device, and this scheduling information includes at least one of the following:

[0455] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp; UE switching to the second network device; UE switching from the second network device to the first network device.

[0456] Step (3): The first network device sends fifth target information (i.e., second information) to the second network device. The fifth target information is determined based on the third target information and includes at least one of the following (the difference from step (3b) in Example 1-1 is that it does not include the handover command):

[0457] Sixth objective information; First objective indication.

[0458] The first target indication is used to indicate the sixth target information and / or switching command, which is obtained based on AI prediction.

[0459] The sixth target information is scheduling information related to the second network device, including at least one of the following:

[0460] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp.

[0461] Example 2-1: Supervised Learning - Captain-Member Pattern - Model Monitoring Process - DPS

[0462] Example 2-1 is a model monitoring process under the assumptions of Example 0, further assuming that multiple cells use DPS transmission mode. Optionally, the model inference of Examples 1-1 to 1-3 can be used as the basic process for model monitoring.

[0463] As shown in Figures 8a and 8b, the operation execution method includes the following process:

[0464] Step (1a) (optional): The first network device (in this example, an information acquisition device) sends the eighth target information (i.e., the third information) to the second network device (in this example, a data source device), the eighth target information being used to request information required for model monitoring by the AI ​​unit.

[0465] Optionally, the information required for model monitoring of the AI ​​unit includes at least one of the following:

[0466] The first quantity measured;

[0467] First KPI;

[0468] The parameters associated with the first KPI include, for example, the duration of observation, the time to time (TTI) of observation, the weighted value of the instantaneous rate, and the weighted value of the historical rate.

[0469] The second target indication (i.e. the first indication information) is used to indicate whether the second network device has fully accepted the scheduling information of step (3) in the model inference process;

[0470] The actual scheduling information of UE2 in the second cell (i.e., the cell where the second network device is located) includes at least one of the following: Serving TRP / beam ID, MCS, precoding, Single User (SU) transmission, and Multi User (MU) transmission.

[0471] Timestamp.

[0472] Wherein, the first measurement quantity includes at least one of the following:

[0473] The measured CQI corresponding to the predicted CQI of UE2;

[0474] The measured SINR corresponding to the predicted SINR of UE2.

[0475] The first KPI includes at least one of the following:

[0476] The accuracy of the predicted CQI for UE2;

[0477] The accuracy of the predicted SINR for UE2;

[0478] UE2's cumulative transmission rate;

[0479] The scheduling fairness of the second network device;

[0480] The average transmission rate of UE2 within the second network device, i.e., the average transmission rate over a period of time, can be characterized by at least one of the following:

[0481] ∑ k log2r k ;

[0482] Where, r k It is UE2 in the second network device k The instantaneous transmission rate at a certain moment;

[0483] It is UE2 in the second network device k The historical transmission rate up to a certain point in time;

[0484] α: Weighted value of instantaneous velocity;

[0485] β: Weighted value of historical rates.

[0486] Step (1b): The first network device receives the ninth target information (i.e., the first information) from the second network device, the ninth target information including information required for model monitoring, including at least one of the following:

[0487] The first quantity measured;

[0488] First KPI;

[0489] The parameters associated with the first KPI (optional);

[0490] The second target indication is used to indicate whether the second network device has fully accepted the scheduling information of step (3) in the model inference process;

[0491] The actual serving TRP / beam ID of UE2 in the second cell (i.e., the cell where the second network device is located);

[0492] Timestamp.

[0493] Step (2): The first network device determines the monitoring results of the AI ​​unit based on the ninth target information.

[0494] Example 2-2: Supervised Learning - Captain-Member Pattern – Model Monitoring Process - ICBM

[0495] Example 2-2 is a model monitoring process that, under the assumptions of Example 0, further assumes that multiple cells use ICBM transmission. Optionally, the model inference of Examples 1-1 to 1-3 can be used as the basic process for model monitoring.

[0496] As shown in Figures 9a and 9b, the operation execution method includes the following process:

[0497] Step (1a) (optional): The first network device (in this example, an information acquisition device) sends the eighth target information (i.e., the third information) to the second network device (in this example, a data source device), the eighth target information being used to request information required for model monitoring by the AI ​​unit.

[0498] Optionally, the information required for model monitoring of the AI ​​unit includes at least one of the following:

[0499] The first quantity measured;

[0500] First KPI;

[0501] The parameters associated with the first KPI include, for example, the duration of observation, the time to time (TTI) of observation, the weighted value of the instantaneous rate, and the weighted value of the historical rate.

[0502] The second target indication is used to indicate whether the second network device has fully accepted the scheduling information of step (3) in the model inference process;

[0503] The actual scheduling information of UE2 in the cell where the second network device is located includes at least one of the following: Serving TRP / beam ID, MCS, precoding, SU transmission, MU transmission;

[0504] The actual scheduling information of UE3 in the cell where the second network device is located includes at least one of the following: serving TRP / beam ID, MCS, precoding, SU transmission, MU transmission; the serving cell of UE3 is the cell where the first network device is located, but the downlink transmission is only sent by the second network device.

[0505] Timestamp.

[0506] Wherein, the first measurement quantity includes at least one of the following:

[0507] The measured CQI corresponding to the predicted CQI of UE2;

[0508] The measured SINR corresponding to the predicted SINR of UE2.

[0509] The first KPI includes at least one of the following:

[0510] The accuracy of the predicted CQI of UE2 (optional, can be sent directly by UE3 to the cell where the first network device is located);

[0511] The accuracy of the predicted SINR for UE2;

[0512] The accuracy of the predicted CQI of UE3 (optional, can be sent directly by UE3 to the cell where the first network device is located);

[0513] Accuracy of predicted SINR for UE3;

[0514] UE2's cumulative transmission rate;

[0515] UE3's cumulative transmission rate;

[0516] The scheduling fairness of the second network device;

[0517] The average transmission rate of UE2 within the second network device, i.e., the average transmission rate over a period of time, can be characterized by at least one of the following: ∑ k log2r k ;

[0518] Where, r k It is UE2 in the second network device k The instantaneous transmission rate at a certain moment;

[0519] It is UE2 in the second network device k The historical transmission rate up to a certain point in time;

[0520] α: Weighted value of instantaneous velocity;

[0521] β: Weighted value of historical rates.

[0522] Step (1b): The first network device receives the ninth target information (i.e., the first information) from the second network device, the ninth target information including information required for model monitoring, including at least one of the following:

[0523] The first quantity measured;

[0524] First KPI;

[0525] The parameters associated with the first KPI include, for example, the duration of observation, the time to time (TTI) of observation, the weighted value of the instantaneous rate, and the weighted value of the historical rate.

[0526] The second target indication is used to indicate whether the second network device has fully accepted the scheduling information of step (3) in the model inference process;

[0527] The actual scheduling information of UE2 in the cell where the second network device is located includes at least one of the following: Serving TRP / beam ID, MCS, precoding, SU transmission, MU transmission;

[0528] The actual scheduling information of UE3 in the cell where the second network device is located includes at least one of the following: Serving TRP / beam ID, MCS, precoding, SU transmission, MU transmission;

[0529] Timestamp.

[0530] Step (2): The first network device determines the monitoring results of the AI ​​unit based on the ninth target information.

[0531] Example 2-3: Supervised Learning - Captain-Member Pattern – Model Monitoring Process - CJT

[0532] Examples 2-3 are model monitoring procedures based on the assumptions of Example 0, further assuming that CJT transmission is used between multiple cells. Optionally, the model inference of Examples 1-1 to 1-3 can be used as the basic process for model monitoring.

[0533] As shown in Figures 10a and 10b, the operation execution method includes the following process:

[0534] Step (1a) (optional): The first network device (in this example, an information acquisition device) sends the eighth target information (i.e., the third information) to the second network device (in this example, a data source device), the eighth target information being used to request information required for model monitoring by the AI ​​unit.

[0535] Optionally, the information required for model monitoring of the AI ​​unit includes at least one of the following:

[0536] The first quantity measured;

[0537] First KPI;

[0538] The parameters associated with the first KPI include, for example, the duration of observation, the time to time (TTI) of observation, the weighted value of the instantaneous rate, and the weighted value of the historical rate.

[0539] The second target indication is used to indicate whether the second network device has fully accepted the scheduling information of step (3) in the model inference process;

[0540] The actual scheduling information of UE2 in the cell where the second network device is located includes at least one of the following: Serving TRP / beam ID, MCS, precoding, SU transmission, MU transmission;

[0541] Timestamp.

[0542] Wherein, the first measurement quantity includes at least one of the following:

[0543] The measured CQI under joint precoding corresponding to the predicted CQI of UE2;

[0544] The measured SINR under joint precoding corresponds to the predicted SINR of UE2.

[0545] The first KPI includes at least one of the following:

[0546] The accuracy of the predicted CQI under joint precoding by UE2 (optional, can be sent directly to cell 1 by UE3);

[0547] The accuracy of the predicted SINR under UE2 joint precoding;

[0548] UE2's cumulative transmission rate;

[0549] The scheduling fairness of the second network device;

[0550] The average transmission rate of UE2 within the second network device, i.e., the average transmission rate over a period of time, can be characterized by at least one of the following: ∑ k log2r k ;

[0551] Where, r k It is UE2 in the second network device k The instantaneous transmission rate at a certain moment;

[0552] It is UE2 in the second network device k The historical transmission rate up to a certain point in time;

[0553] α: Weighted value of instantaneous velocity.

[0554] β: Weighted value of historical rates.

[0555] Step (1b): The first network device receives the ninth target information (i.e., the first information) from the second network device. The ninth target information includes information required for model monitoring of the AI ​​unit, including at least one of the following:

[0556] The first quantity measured;

[0557] First KPI;

[0558] The parameters associated with the first KPI include, for example, the duration of observation, the time to time (TTI) of observation, the weighted value of the instantaneous rate, and the weighted value of the historical rate.

[0559] The second target indication is used to indicate whether the second network device has fully accepted the scheduling information of step (3) in the model inference process;

[0560] The actual scheduling information of UE2 in the cell where the second network device is located includes at least one of the following: Serving TRP / beam ID, MCS, precoding, SU transmission, MU transmission;

[0561] Timestamp.

[0562] Step (2): The first network device determines the monitoring results of the AI ​​unit based on the ninth target information.

[0563] Example 3-1: Supervised Learning - Captain-Member Model - Training Data Collection Process - DPS

[0564] Example 3-1 is the training data collection process under the assumptions of Example 0, further assuming that DPS transmission is used between multiple cells. Some steps of the training data collection process can be found in the model inference process and model monitoring process.

[0565] As shown in Figures 11a and 11b, the operation execution method includes the following process:

[0566] Step (1a) (optional): Same as step (1a) in Example 1-1.

[0567] Step (1b): Same as step (1b) in Example 1-1.

[0568] Step (2): The first network device obtains the third target information. For example, the method for obtaining the third target information can be a high-complexity method such as exhaustive search. The third target information includes scheduling information of the first network device and at least one second network device, and includes at least one of the following:

[0569] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp.

[0570] Step (3a): Same as step (3a) in Example 1-1.

[0571] Step (3b): Same as step (3b) in Example 1-1.

[0572] Step (4): Same as step (4) in Example 1-1.

[0573] Step (5a) (optional): Same as step (1a) in Example 2-1.

[0574] Step (5b): Same as step (1b) in Example 2-1.

[0575] Step (6): The first network device acquires the dataset.

[0576] Example 4-1: Reinforcement Learning - Captain-Member Model - Training Data Collection Process - DPS

[0577] Example 3-1 illustrates the training data collection process under the assumptions of Example 0, where DPS (Digital Per Second) is used for transmission between multiple cells. Example 4-1 illustrates the training data collection process assuming a reinforcement learning approach is adopted. Some steps in Example 4-1 can be referenced from the model inference and model monitoring processes.

[0578] As shown in Figures 12a and 12b, the operation execution method includes the following process:

[0579] Step (1a) (optional): Same as step (1a) in Example 1-1.

[0580] Step (1b): Same as step (1b) in Example 1-1.

[0581] Step (2): The first network device performs inference for the AI ​​unit. Specifically, it obtains the state information of the input AI unit based on the second target information. Through the inference of the AI ​​unit, the third target information is obtained.

[0582] The third objective information can also be referred to as the behavior or action in reinforcement learning. The third objective information can be determined directly based on the inference results of the AI ​​unit, or randomly selected from the set of candidate actions.

[0583] The third target information includes scheduling information for the first network device and at least one second network device, and includes at least one of the following:

[0584] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp.

[0585] For example, the state inputs as AI units include at least one of the following:

[0586] cell ID;

[0587] TRP (e.g., one or more TRPs from a historical service) ID;

[0588] Beam ID;

[0589] Measurement quantities, such as RSRP, RSRQ, SINR, CQI, PMI, MCS and / or HARQ.

[0590] The TRP ID / beam ID can be represented by at least one of the following: TCI ID, SSB index, CSI-RS ID (type A index), and CRI.

[0591] Step (3a): Same as step (3a) in Example 1-1.

[0592] Step (3b): Same as step (3b) in Example 1-1.

[0593] Step (4): Same as step (4) in Example 1-1.

[0594] Step (5a) (optional): Same as step (1a) in Example 2-1.

[0595] Step (5b): Same as step (1b) in Example 2-1.

[0596] Step (6): The first network device determines the environmental feedback or reward based on the ninth target information and obtains the reinforcement learning dataset.

[0597] Specifically, the reinforcement learning dataset (current state, next state, action, reward) is obtained based on the third and ninth objective information.

[0598] Example 2: Peer-to-Peer Mode

[0599] Example 1:

[0600] This example illustrates the functional division of each network device in a peer-to-peer model.

[0601] Example 1 assumes multiple cells, each deploying an intelligent agent. Each agent performs AI unit inference to obtain the selection result of multiple TRPs. This example analyzes the division of different scheduling functions, thereby avoiding the problem that the cell of equipment vendor 1 must obey the scheduling result of the cell of equipment vendor 2.

[0602] Based on whether UEs and TRPs are shared, there are four possible scenarios:

[0603] Case 1: The UE cannot enter or exit, and the TRP is not shared. This case does not belong to multi-cell TRP optimization.

[0604] Scenario 2: UE can enter and exit freely, TRP is not shared, as shown in Figure 13;

[0605] Case 3: UE cannot enter or exit, TRP is shared, as shown in Figure 14;

[0606] Scenario 4: UE can enter and exit freely, TRP is shared, as shown in Figure 15.

[0607] In this context, "UE cannot enter or leave" means that the UE cannot switch serving cells; "UE can enter and leave" means that the UE can switch serving cells.

[0608] Example 2-1:

[0609] This example demonstrates how to implement multi-agent negotiation in scenario 2, where the negotiation is an alternating negotiation.

[0610] Example 1 assumes multiple cells, each deployed with an agent. Each agent performs AI unit inference to obtain the selection result of multiple TRPs. Example 2-1 addresses the interaction problem between multiple peer cells in scenario 2, where UEs can enter and exit freely and TRPs are not shared. The main difference from deploying a single agent is that the model input includes handover information.

[0611] The implementation scenario for Case 2 is shown in Table 2 below.

[0612] Table 2

[0613] As shown in Figure 16, the operation execution method includes the following process:

[0614] Step (1a) (optional): The first network device (in this step, the information acquisition device) sends the first target information (i.e., the third information) to the second network device (in this step, the data source device). The first target information is used to request the information required for the first AI unit to reason.

[0615] Optionally, the information required for the first AI unit inference associated with the first target information includes at least one of the following:

[0616] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under DPS, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0617] UE2 carries at least one measurement result from the second network device, for example, under DPS;

[0618] UE2 carries at least one serving beam and / or TRP of the second network device, for example, under DPS;

[0619] UE2 carries at least one scheduling information from the second network device, for example, under DPS;

[0620] UE2 carries at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, if the first network device cannot obtain HACK feedback under DPS, this option is carried.

[0621] Timestamp;

[0622] The second network device is recommended to switch to the UE ID of UE2 of the first network device.

[0623] Wherein, at least one serving beam and / or TRP of UE2 in the second network device can be indicated by at least one of the following:

[0624] TCI state ID; TCI ID; SSB index; CSI-RS index; type A index; CRI.

[0625] It should be noted that the CRIs mentioned above are local CRIs, for example, CRIs 0-7. The TCIs mentioned above are global TCIs, indicating a wider range of TCIs, for example: TCI IDs 0-127. The network can translate CRIs into TCIs, meaning there is a mapping relationship between CRIs and TCIs. CRIs are not global CRIs, and the UE is aware of this mapping relationship. The SSB or CSI-RS indexes mentioned above are global SSB or CSI-RS indexes.

[0626] Wherein, at least one of the scheduling information of UE2 in the second network device includes at least one of the following:

[0627] The allocated MCS; the precoding matrix; the allocated frequency band resources (e.g., resource block (RB) information).

[0628] At least one scheduling information of UE2 in the second network device can be associated with at least one serving beam and / or TRP of UE2 in the second network device.

[0629] Additionally, the measurement results in the inference-related information of the first AI unit may refer to historical measurement results, which include at least one of the following: RSRP; RSRQ; SINR of at least one serving beam and / or TRP.

[0630] Step (1b) (optional): The first network device (the data source device in this step) receives the tenth target information (i.e. the third information) sent by the second network device (the information acquisition device in this step), the tenth target information being used to request the information required for the second AI unit to reason.

[0631] Optionally, the information required for the first AI unit inference associated with the tenth target information includes at least one of the following:

[0632] At least one measurement result of UE2 in the cell corresponding to the first network device (hereinafter referred to as at least one measurement result of UE2 in the first network device);

[0633] At least one measurement result of UE1 in the cell corresponding to the first network device (hereinafter referred to as at least one measurement result of UE1 in the first network device);

[0634] UE1 has at least one serving beam and / or TRP in the cell corresponding to the first network device (hereinafter referred to as at least one serving beam and / or TRP of UE1 in the first network device). For example, the tenth target information requests this option under DPS.

[0635] At least one scheduling information of UE1 in the cell corresponding to the first network device (hereinafter referred to as at least one scheduling information of UE1 in the first network device), for example, the tenth target information requests this option under DPS;

[0636] At least one scheduling feedback information of UE1 in the cell corresponding to the first network device (hereinafter referred to as at least one scheduling feedback information of UE1 in the first network device), such as HACK feedback. For example, under DPS, the second network device cannot obtain the HACK feedback. At this time, the tenth target information requests this option.

[0637] Timestamp.

[0638] in,

[0639] At least one serving beam and / or TRP of UE1 in the first network device can be indicated by at least one of the following:

[0640] TCI state ID; TCI ID; SSB index; CSI-RS index; type A index; CRI.

[0641] It should be noted that the CRIs mentioned above are local CRIs, for example, CRIs 0-7. The TCIs mentioned above are global TCIs, indicating a wider range of TCIs, for example: TCI IDs 0-127. The network can translate CRIs into TCIs, meaning there is a mapping relationship between CRIs and TCIs. CRIs are not global CRIs, and the UE is aware of this mapping relationship. The SSB or CSI-RS indexes mentioned above are global SSB or CSI-RS indexes.

[0642] Wherein, at least one of the scheduling information of UE1 in the first network device includes at least one of the following:

[0643] Allocated MCS; precoding matrix (e.g., independent precoding matrix, joint precoding matrix); allocated frequency band resources (e.g., RB information).

[0644] At least one scheduling information of UE1 in the first network device can be associated with at least one serving beam and / or TRP of UE1 in the first network device.

[0645] The measurement results in the inference-related information of the second AI unit may refer to historical measurement results, which include at least one of the following: RSRP; RSRQ; SINR of at least one serving beam and / or TRP.

[0646] Step (2): The first network device (in this step, the information acquisition device) receives the second target information (i.e., the first information) from the second network device (in this step, the data source device). The second target information includes: relevant information required for the inference of the first AI unit. A detailed description of the relevant information required for the inference of the first AI unit can be found in step (1a) of this example, and will not be repeated here.

[0647] It should be noted that steps (1b) and (2) can be implemented through a single signaling instruction or through multiple separate signaling instructions. The signaling instructions used for implementation can be related to processes such as switching, load balancing, and resource management, or they can be new signaling instructions. This example does not limit the name of the signaling instructions.

[0648] Step (3): The first network device performs inference using the first AI unit to obtain third target information. The third target information includes at least one of the following:

[0649] The selected UE, for example, the third target information under DPS carries the identifier of UE1, which is used to indicate UE1 that will become UE2;

[0650] UE switched to the second network device;

[0651] UE switching from the second network device to the first network device;

[0652] The selection of TRP within the first network device;

[0653] Precoding matrix, for example, the precoding matrix of the first target UE;

[0654] The assigned MCS, for example, the assigned MCS of the first target UE;

[0655] Predicted CQI, for example, the predicted CQI of the first target UE;

[0656] Timestamp.

[0657] The first target UE includes at least one of the following: UE1, UE2, a UE that will change from UE1 to the second UE, and a UE that will change from UE2 to UE1.

[0658] For example, the model input of the first AI unit includes at least one of the following: cell ID; TRP ID / beam ID (TRP and / or BEAM of historical service); measurement quantity (such as RSRP, RSRQ, SINR, CQI, PMI, MCS and / or HARQ).

[0659] The TRP ID / beam ID can be represented by TCI ID, SSB index, CSI-RS ID (type A index), or CRI.

[0660] Step (4a) (optional): The first network device (in this step, the information acquisition device) sends fourth target information to UE1. The fourth target information includes a handover command, which is determined based on the third target information.

[0661] For example, if the selected UE includes UE1, and the selected TRP belongs to the second network device, then the handover target cell of the fourth target information is the corresponding second network device.

[0662] Step (4b): The first network device (in this step, the information acquisition device) sends the fifth target information (i.e., the second information) to the second network device (in this step, the information source device). The fifth target information includes at least one of the following:

[0663] Switch command; sixth target information; first target indication.

[0664] The first target indication is used to indicate the fifth target information and / or switching command, which is obtained based on AI prediction.

[0665] The sixth target information is scheduling information related to the second network device, including at least one of the following:

[0666] The selected UE, such as UE2 which can be received from the second network device, and / or, the recommended UE1 which can be received by the second network device;

[0667] The selected TRP, for example, recommends that the second network device receive the TRP of UE1 in the second network device, and in case 1, this information can be optionally carried;

[0668] The precoding matrix, for example, is the precoding matrix of UE1 recommended by the second network device, which may optionally carry this information in case 1;

[0669] The assigned MCS, for example, recommends that the second network device receive UE1 in the second network device's MCS, and in case 1, this information may be carried.

[0670] The predicted CQI, for example, is the predicted CQI of UE1 received by the second network device. In case 1, this information can be optionally carried.

[0671] Timestamp.

[0672] Step (5): The first network device (the data source device in this step) sends the eleventh target information (the first information in this step) to the second network device (the information acquisition device in this step). The eleventh target information is the data acquisition feedback of the second AI unit, including at least one of the following:

[0673] At least one measurement result of UE2 in the first network device;

[0674] At least one measurement result of UE1 in the first network device, for example, if the second network device cannot directly obtain at least one measurement result of UE1 in the first network device under DPS, this option is carried;

[0675] UE1 carries at least one serving beam and / or TRP of the first network device, for example, under DPS;

[0676] UE1 carries at least one scheduling message from the first network device, for example, under DPS;

[0677] UE1 carries at least one scheduling feedback message (e.g., HACK feedback) from the first network device. For example, if the second network device cannot obtain the HACK feedback under DPS, then this option is carried.

[0678] Timestamp;

[0679] The fourth target indication is used to indicate whether the first network device has fully accepted the scheduling information previously recommended by the second network device. It should be noted that this option is not carried before the second AI unit has made inference.

[0680] The instruction for UE1 to switch from the second network device to the first network device should be noted that this option is not carried before the second AI unit has started inference.

[0681] Step (6): The second network device performs inference in the second AI unit to obtain the twelfth target information, which includes at least one of the following:

[0682] The selected UE, for example, includes UE2, and / or UE2 that will become UE1;

[0683] Switch to UE2, the first network device;

[0684] Switch to UE1 on the second network device;

[0685] The selection of TRP within the second network device;

[0686] Precoding matrix, for example, the precoding matrix of the second target UE;

[0687] The assigned MCS, for example, the assigned MCS of the second target UE;

[0688] Predicted CQI, for example, the predicted CQI of the second target UE;

[0689] Timestamp.

[0690] The second target UE includes at least one of the following: UE1, UE2, a UE that will change from UE1 to the second UE, and a UE that will change from UE2 to UE1.

[0691] Step (7a) (optional): The second network device (in this step, the information acquisition device) sends the seventh target information to UE2. The seventh target information includes a handover command, which is determined based on the twelfth target information.

[0692] For example, if the selected UE includes UE2, and the selected TRP belongs to the first network device, then the handover target cell of the thirteenth target information is the corresponding first network device.

[0693] Step (7b): The second network device (in this step, the information acquisition device) sends the fourteenth target information (in this step, the second information) to the first network device (in this step, the data source device), wherein the fourteenth target information includes at least one of the following:

[0694] Switching command; Thirteenth target information; Third target indication.

[0695] The third target indication is used to indicate the thirteenth target information and / or switching commands, which are obtained based on AI prediction.

[0696] Step (8): The first network device (in this step, the information acquisition device) receives the second target information (in this step, the first information) from the second network device (in this step, the data source device). The second target information includes: inference-related information of the first AI unit.

[0697] The inference-related information of the first AI unit includes at least one of the following:

[0698] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under DPS, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0699] UE2 carries at least one measurement result from the second network device, for example, under DPS;

[0700] UE2 carries at least one serving beam and / or TRP of the second network device, for example, under DPS;

[0701] UE2 carries at least one scheduling information from the second network device, for example, under DPS;

[0702] UE2 carries at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, if the first network device cannot obtain HACK feedback under DPS, this option is carried.

[0703] The second target indication is used to indicate whether the second network device has fully accepted the scheduling recommendation information previously provided by the first network device.

[0704] The instruction for UE2 to switch from the first network device to the second network device should be noted that this option is not carried before the first AI unit has started inference.

[0705] Timestamp.

[0706] It should be noted that steps (7b) and (8) can be implemented by a single signaling instruction, or by multiple signaling instructions respectively.

[0707] Steps 2, 3, 4a, 4b, 5, 6, 7a, 7b, and 8 can be a continuously looping process until the first network device and / or the second network device terminates the process. The process can be terminated by the data source device sending a message to the information acquisition device, or by the information acquisition device sending a message to the data source device.

[0708] Example 2-2:

[0709] This example demonstrates how to implement multi-agent negotiation in scenario 3, where the negotiation is an alternating negotiation.

[0710] Example 2-2 addresses the issue of how multiple peer cells interact in Case 3, where the UE is not allowed to enter or leave and the TRP can be shared.

[0711] The implementation scenarios for one example of Case 3 are shown in Table 3 below.

[0712] Table 3

[0713] As shown in Figure 17, the operation execution method includes the following process:

[0714] Step (1a) (optional): The first network device (in this step, the information acquisition device) sends the first target information (i.e., the third information) to the second network device (in this step, the data source device). The first target information is used to request the relevant information required for the inference of the first AI unit.

[0715] Optionally, the relevant information required for the inference of the first AI unit associated with the first target information includes at least one of the following:

[0716] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under ICBM or CJT, the first network device can know at least one measurement result of UE1 in the second network device, but only one cell can know it. In this case, this option can be omitted.

[0717] At least one measurement result of UE2 in the second network device. For example, under ICBM, the first network device can know at least one measurement result of UE2 in the second network device, but only one cell can know it. In this case, this option can be omitted. Under CJT, the first network device cannot know at least one measurement result of UE2 in the second network device. In this case, this option can be included.

[0718] UE2 has at least one serving beam and / or TRP in the second network device;

[0719] At least one scheduling information of UE2 in the second network device;

[0720] UE2 has at least one scheduling feedback message (e.g., HACK feedback) from the second network device. Under ICBM, the first network device can be aware of the HACK feedback, but only one cell can be aware of it. In this case, the option can be omitted. Under CJT, the first network device cannot be aware of the HACK feedback. In this case, the option can be included.

[0721] Timestamp.

[0722] The first target information includes the types of services that can be provided to UE2.

[0723] The types of services that can be provided to UE2 include at least one of the following:

[0724] Individual uplink or downlink transmissions are performed by the TRP of the first network device (e.g., this option is carried under DPS);

[0725] The TRP of the first network device performs joint uplink or downlink transmission (e.g., carrying this option under ICBM);

[0726] The device can be switched from a second network device to a first network device (e.g., this option is carried under DPS).

[0727] Wherein, at least one of the scheduling information of UE2 in the second network device includes at least one of the following:

[0728] Assigned MCS;

[0729] Precoding matrices, such as joint precoding matrices, independent precoding matrices, single-user precoding matrices, or multi-user precoding matrices;

[0730] Allocated frequency band resources (e.g., RB information).

[0731] At least one scheduling information of UE2 in the second network device can be associated with at least one serving beam and / or TRP of UE2 in the second network device.

[0732] Step (1b) (optional): The first network device (the data source device in this step) receives the tenth target information (the third information in this step) sent by the second network device (the information acquisition device in this step). The tenth target information is used to request the relevant information required for the second AI unit to reason.

[0733] Optionally, the relevant information required for the second AI unit inference associated with the first target information includes at least one of the following:

[0734] At least one measurement result of UE2 in the first network device;

[0735] At least one measurement result of UE1 in the first network device;

[0736] UE1 carries at least one serving beam and / or TRP of the first network device, for example, under DPS;

[0737] UE1 carries at least one scheduling message from the first network device, for example, under DPS;

[0738] UE1 carries at least one scheduling feedback message (e.g., HACK feedback) from the first network device. For example, under DPS, the first network device cannot obtain the HACK feedback and can carry this option.

[0739] Timestamp.

[0740] The tenth target information includes the types of services that can be provided to UE1, including at least one of the following:

[0741] Individual uplink or downlink transmissions are performed by the TRP of the second network device (e.g., this option is carried under ICBM);

[0742] Joint uplink or downlink transmission is performed by the TRP of the second network device (e.g., this option is carried under CJT);

[0743] The device can be switched from a second network device to a first network device (e.g., this option is carried under DPS).

[0744] in,

[0745] At least one serving beam and / or TRP of UE1 in the first network device can be indicated by at least one of the following:

[0746] TCI state ID; TCI ID; SSB index; CSI-RS index; type A index; CRI.

[0747] It should be noted that the CRIs mentioned above are local CRIs, for example, CRIs 0-7. The TCIs mentioned above are global TCIs, indicating a wider range of TCIs, for example: TCI IDs 0-127. The network can translate CRIs into TCIs, meaning there is a mapping relationship between CRIs and TCIs. CRIs are not global CRIs, and the UE is aware of this mapping relationship. The SSB or CSI-RS indexes mentioned above are global SSB or CSI-RS indexes.

[0748] At least one of the scheduling information of UE1 in the first network device includes:

[0749] Assigned MCS;

[0750] Precoding matrices, such as joint precoding matrices, independent precoding matrices, single-user precoding matrices, or multi-user precoding matrices;

[0751] Allocated frequency band resources (e.g., RB information).

[0752] At least one scheduling information of UE1 in the first network device can be associated with at least one serving beam and / or TRP of UE1 in the first network device.

[0753] The measurement results in the inference-related information of the second AI unit may refer to historical measurement results, which include at least one of the following: RSRP; RSRQ; SINR of at least one serving beam and / or TRP.

[0754] Step (2): The first network device (in this step, the information acquisition device) receives the second target information (i.e., the first information) from the second network device (in this step, the data source device). The second target information includes: relevant information required for the inference of the first AI unit.

[0755] The relevant information required for the inference of the first AI unit includes at least one of the following:

[0756] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under DPS, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0757] UE2 carries at least one measurement result from the second network device, for example, under DPS;

[0758] UE2 carries at least one serving beam and / or TRP of the second network device, for example, under DPS;

[0759] UE2 carries at least one scheduling information from the second network device, for example, under DPS;

[0760] UE2 carries at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, if the first network device cannot obtain HACK feedback under DPS, this option is carried.

[0761] Timestamp;

[0762] The second network device recommends that the UE ID of UE2 be transmitted separately uplink or downlink by the TRP of the first network device, for example, by carrying this option under ICBM;

[0763] The second network device recommends that the UE ID of UE2 be jointly transmitted uplink or downlink by the TRP of the first network device, for example, by carrying this option under CJT.

[0764] Step (3): The first network device (in this step, the information acquisition device) performs inference on the first AI unit to obtain third target information, wherein the third target information includes at least one of the following:

[0765] Selected UEs, such as UE1, and / or UE2;

[0766] UE switched to the second network device;

[0767] UE switching from the second network device to the first network device;

[0768] The selection of TRP within the first network device;

[0769] The selection of TRP within the second network device;

[0770] Precoding matrices, for example, the precoding matrix of UE1, and / or the precoding matrix of UE2;

[0771] The assigned MCS, for example, the assigned MCS of UE1, and / or the assigned MCS of UE2;

[0772] Predicted CQI, for example, the predicted CQI of UE1, and / or the predicted CQI of UE2;

[0773] Timestamp.

[0774] For example, the model input of the first AI unit includes at least one of the following: cell ID; TRP ID / beam ID (such as TRP and / or BEAM of historical services); measurement quantity (such as RSRP, RSRQ, SINR, CQI, PMI, MCS and / or HARQ).

[0775] The TRP ID / beam ID can be represented by TCI ID, SSB index, CSI-RS ID (type A index), or CRI.

[0776] Step (4b): The first network device (in this step, the information acquisition device) sends the fifth target information (i.e., the second information) to the second network device (in this step, the data source device), wherein the fifth target information includes at least one of the following:

[0777] Switch command; sixth target information; first target indication.

[0778] The first target indication is used to indicate the fifth target information and / or switching command, which is obtained based on AI prediction.

[0779] The sixth target information is scheduling information related to the second network device, including at least one of the following:

[0780] Selected UE; Selected TRP; Precoding matrix; Assigned MCS; Predicted CQI; Timestamp; UE ID of UE2 that can provide independent precoding (example, this option is carried in case 3); UE ID of UE2 that can provide joint precoding (example, this option is carried in case 3); UE ID of UE1 that is recommended for independent precoding by a second network device (example, this option is carried in case 3); UE ID of UE1 that is recommended for joint precoding by a second network device (example, this option is carried in case 3).

[0781] Step (5): The first network device (the data source device in this step) sends the eleventh target information (the first information in this step) to the second network device (the information acquisition device in this step). The eleventh target information is the data acquisition feedback of the second AI unit, including at least one of the following:

[0782] At least one measurement result of UE2 in the first network device;

[0783] At least one measurement result of UE1 in the first network device, for example, if the second network device cannot directly obtain at least one measurement result of UE1 in the first network device under DPS, this option is carried;

[0784] UE1 carries at least one serving beam and / or TRP of the first network device, for example, under DPS;

[0785] UE1 carries at least one scheduling message from the first network device, for example, under DPS;

[0786] UE1 carries at least one scheduling feedback message (e.g., HACK feedback) from the first network device. For example, if the second network device cannot obtain the HACK feedback under DPS, then this option is carried.

[0787] Timestamp;

[0788] The fourth target indication is used to indicate whether the first network device has fully accepted the scheduling information previously recommended by the second network device. It should be noted that this option is not carried before the second AI unit has made inference.

[0789] The UE ID of UE2 is provided independently precoded by the first network device, and this option is carried as an example in case 3;

[0790] The UE ID of UE2 is provided by the first network device in joint precoding, and this option is carried as an example in case 3.

[0791] Step (6): The second network device (in this step, the information acquisition device) performs inference in the second AI unit to obtain the twelfth target information, which includes at least one of the following:

[0792] The selected UE, for example, includes UE2, and / or UE2 that will become UE1;

[0793] Switch to UE2, the first network device;

[0794] Switch to UE1 on the second network device;

[0795] The selection of TRP within the second network device;

[0796] Precoding matrix, for example, the precoding matrix of the second target UE;

[0797] The assigned MCS, for example, the assigned MCS of the second target UE;

[0798] Predicted CQI, for example, the predicted CQI of the second target UE;

[0799] Timestamp.

[0800] The second target UE includes at least one of the following: UE1, UE2, a UE that will change from UE1 to the second UE, and a UE that will change from UE2 to UE1.

[0801] Step (7b): The second network device (in this step, the information acquisition device) sends the fourteenth target information (in this step, the second information) to the first network device (in this step, the data source device), wherein the fourteenth target information includes at least one of the following:

[0802] Switching command; Thirteenth target information; Third target indication.

[0803] The third target indication is used to indicate the thirteenth target information and / or switching commands, which are obtained based on AI prediction.

[0804] Step (8): The first network device (in this step, the information acquisition device) receives the second target information (in this step, the first information) from the second network device (in this step, the data source device). The second target information includes: relevant information required for the inference of the first AI unit.

[0805] The relevant information required for the inference of the first AI unit includes at least one of the following:

[0806] At least one measurement result of UE1 in the second network device, wherein the serving cell of UE1 is the cell where the first network device is located. For example, under DPS, the first network device can obtain at least one measurement result of UE1 in the second network device. In this case, this option may not be included.

[0807] UE2 carries at least one measurement result from the second network device, for example, under DPS;

[0808] UE2 carries at least one serving beam and / or TRP of the second network device, for example, under DPS;

[0809] UE2 carries at least one scheduling information from the second network device, for example, under DPS;

[0810] UE2 carries at least one scheduling feedback message (e.g., HACK feedback) from the second network device. For example, if the first network device cannot obtain HACK feedback under DPS, this option is carried.

[0811] The second target indication is used to indicate whether the second network device has fully accepted the scheduling recommendation information previously provided by the first network device.

[0812] The UE ID of UE1 is provided by a second network device with independent precoding, for example, in case 3, this option is carried;

[0813] The UE ID of UE1 is jointly precoded by the second network device, for example, this option is carried in case 3;

[0814] Timestamp.

[0815] Steps 2, 3, 4b, 5, 6, 7b, and 8 can be a continuously looping process until the first network device and / or the second network device terminates the process. The process can be terminated by the data source device sending a message to the information acquisition device, or by the information acquisition device sending a message to the data source device.

[0816] Example 2-3:

[0817] This example demonstrates how to implement multi-agent negotiation in scenario 4, where the negotiation is an alternating negotiation.

[0818] Example 2-3 addresses the issue of how multiple peer cells interact in Case 4, where UEs can enter and exit and TRPs can be shared.

[0819] The implementation scenarios for one example of Case 4 are shown in Table 4 below.

[0820] Table 4

[0821] As shown in Figure 18, the operation execution method includes the following process:

[0822] Step (1a) (optional): The first network device (in this step, the information acquisition device) sends the first target information (i.e., the third information) to the second network device (in this step, the data source device). The first target information is used to request the relevant information required for the inference of the first AI unit.

[0823] Step (1b) (optional): The first network device (the data source device in this step) receives the tenth target information (i.e. the third information) sent by the second network device (the information acquisition device in this step), the tenth target information being used to request the relevant information required for the second AI unit to infer.

[0824] Step (2): The first network device (in this step, the information acquisition device) receives the second target information (i.e., the first information) from the second network device (in this step, the data source device). The second target information includes: relevant information required for the inference of the first AI unit.

[0825] Step (3): The first network device (in this step, the information acquisition device) performs inference of the first AI unit to obtain the third target information.

[0826] Step (4a) (optional): The first network device (in this step, the information acquisition device) sends fourth target information to UE1, the fourth target information including a handover command.

[0827] Step (4b): The first network device (in this step, the information acquisition device) sends the fifth target information (i.e., the second information) to the second network device (in this step, the data source device).

[0828] Step (5): The first network device (the data source device in this step) sends the sixth target information (i.e. the first information) to the second network device (the information acquisition device in this step). The sixth target information is the data acquisition feedback of the second AI unit.

[0829] Step (6): The second network device (in this step, the information acquisition device) performs reasoning in the second AI unit to obtain the twelfth target information.

[0830] Step (7a) (optional): The second network device (in this step, the information acquisition device) sends the seventh target information to UE2, the seventh target information including a handover command.

[0831] Step (7b): The second network device (in this step, the information acquisition device) sends the thirteenth target information (i.e., the second information) to the first network device (in this step, the data source device).

[0832] Step (8): The first network device (in this step, the information acquisition device) receives the second target information (i.e., the first information) from the second network device (in this step, the data source device). The second target information includes: relevant information required for the inference of the first AI unit.

[0833] It should be noted that the information in each step of Example 2-3 is a collection of information with the same name in Example 2-1 and Example 2-2 of Example 2. To avoid duplication, it will not be repeated here.

[0834] This application provides a multi-agent-based multi-TRP transmission method that can reduce the complexity of multi-TRP transmission and improve the transmission rate by leveraging AI.

[0835] Referring to Figure 19, which is a flowchart of an operation execution method provided in an embodiment of this application, the operation execution method includes the following steps:

[0836] Step 201: The information source device sends first information to the information acquisition device, wherein the first information is the transmission-related information of the information source device;

[0837] The first information is used to perform the first operation of the target AI unit;

[0838] The target AI unit is associated with the information acquisition device;

[0839] The first operation includes at least one of the following:

[0840] Model inference, model monitoring, and training data collection.

[0841] Optionally, when the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following:

[0842] The first terminal has at least one measurement result in the cell corresponding to the information source device;

[0843] The second terminal has at least one measurement result in the cell corresponding to the information source device;

[0844] The second terminal is in at least one serving beam of the cell corresponding to the information source device;

[0845] The second terminal is in at least one TRP in the cell corresponding to the information source device;

[0846] The second terminal has at least one scheduling information in the cell corresponding to the information source device;

[0847] The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device.

[0848] The user throughput rate of the information source device;

[0849] First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device;

[0850] The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device;

[0851] The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device;

[0852] The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device;

[0853] TRP of candidate information source devices;

[0854] Beams of candidate information source devices;

[0855] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0856] Optionally, the beam and / or TRP in the model inference-related information are indicated by at least one of the following:

[0857] TCI status identifier; TCI identifier; SSB index; CSI-RS index; CRI.

[0858] Optionally, if the first operation includes model inference, the method further includes:

[0859] The information source device receives at least one of the second information and the second indication information sent by the information acquisition device. The second indication information is used to indicate that the second information is obtained based on AI unit prediction. The second information is scheduling information related to the transmission of the information source device. The second information is obtained based on the model inference result.

[0860] Optionally, the second information includes the model inference result, which includes at least one of the following:

[0861] The identifier of the target terminal; the TRP corresponding to the target terminal; the beam corresponding to the target terminal; the precoding matrix corresponding to the target terminal; the modulation and coding information corresponding to the target terminal; the predicted channel quality information corresponding to the target terminal; and the first predicted measurement.

[0862] The target terminal includes at least one of the following:

[0863] It is recommended to switch to the first terminal in the cell corresponding to the information source device; it is recommended to switch to the second terminal in the cell corresponding to the information acquisition device; it is recommended that the second terminal be scheduled by the information source device; it is recommended that the first terminal be scheduled by the information acquisition device.

[0864] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0865] Optionally, the second information further includes at least one of the following:

[0866] The information acquisition device can provide an independently pre-encoded identifier for the second terminal;

[0867] The information acquisition device can provide a jointly pre-coded identifier for the second terminal;

[0868] The identifier of the first terminal is recommended to be independently pre-encoded by the information source device;

[0869] The identifier of the first terminal is recommended to be jointly pre-coded by the information source device.

[0870] Optionally, when the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following:

[0871] The first measured quantity; the first KPI; the parameters associated with the first KPI; the actual scheduling information of the second terminal in the cell corresponding to the information source device;

[0872] The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device.

[0873] The serving cell of the second terminal is the cell corresponding to the information source device.

[0874] Optionally, the method further includes:

[0875] The information source device receives third information sent by the information acquisition device;

[0876] Alternatively, the information source device receives the third information sent by the information acquisition device and sends confirmation information of the third information to the information source device;

[0877] The third information is used to request the first information;

[0878] The confirmation information is used to instruct the information source device to agree to the request corresponding to the third information.

[0879] Optionally, the third information includes at least one of the following: the service type that the information acquisition device can provide to the second terminal, and the service type that the information acquisition device requests from the information source device;

[0880] Wherein, the serving cell of the second terminal is the cell corresponding to the information source device;

[0881] The service type includes at least one of the following: DPS; ICBM; CJT; NCJT.

[0882] It should be noted that this embodiment is an implementation of the information source device corresponding to the embodiment shown in Figure 3. For the specific implementation, please refer to the relevant description of the embodiment shown in Figure 3. To avoid repetition, this embodiment will not be described again.

[0883] The operation execution method provided in this application can be executed by an operation execution device. This application uses an operation execution device executing the operation execution method as an example to illustrate the operation execution device provided in this application.

[0884] This application provides an operation execution device. As an example, the operation execution device may be a communication device or a component within a communication device, such as a chip. The communication device may be a terminal, a network-side device, or a server, etc. Exemplarily, the terminal may include, but is not limited to, the type of terminal 11 listed above, and the network-side device may include, but is not limited to, the type of network-side device 12 listed above. This application does not impose specific limitations.

[0885] The operation execution device includes a receiving module, a transmitting module, and a processing module. These modules can be implemented in software or hardware. When implemented in hardware, the processing module can be implemented by a processor. For example, the processor can include general-purpose processors, special-purpose processors, etc., such as central processing units (CPUs), microprocessors, digital signal processors (DSPs), artificial intelligence (AI) processors, graphics processing units (GPUs), application-specific integrated circuits (ASICs), network processors (NPs), field-programmable gate arrays (FPGAs), or other programmable logic devices, gate circuits, transistors, discrete hardware components, etc. The receiving and transmitting modules can be implemented by a communication interface, which can include one or more of the following: transceivers, pins, circuits, buses, radio frequency units, etc.

[0886] Specifically, referring to Figure 20, when the operation execution device is an information acquisition device or a component of an information acquisition device, the operation execution device 300 includes:

[0887] The receiving module 301 is used to receive first information sent by the information source device, wherein the first information is transmission-related information of the information source device;

[0888] Processing module 302 is used to perform a first operation on the target AI unit based on the first information;

[0889] The target AI unit is associated with the information acquisition device;

[0890] The first operation includes at least one of the following:

[0891] Model inference, model monitoring, and training data collection.

[0892] Optionally, when the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following:

[0893] The first terminal has at least one measurement result in the cell corresponding to the information source device;

[0894] The second terminal has at least one measurement result in the cell corresponding to the information source device;

[0895] The second terminal is in at least one serving beam of the cell corresponding to the information source device;

[0896] The second terminal is in at least one TRP in the cell corresponding to the information source device;

[0897] The second terminal has at least one scheduling information in the cell corresponding to the information source device;

[0898] The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device.

[0899] The user throughput rate of the information source device;

[0900] First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device;

[0901] The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device;

[0902] The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device;

[0903] The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device;

[0904] TRP of candidate information source devices;

[0905] Beams of candidate information source devices;

[0906] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0907] Optionally, the beam and / or TRP in the model inference-related information are indicated by at least one of the following:

[0908] Transmission Configuration Indicator (TCI) Status Identifier; TCI Identifier; Synchronization Signal Block (SSB) Index; Channel State Information Reference Signal (CSI-RS) Index; CSI-RS Resource Indicator (CRI).

[0909] Optionally, when the first operation includes model inference, the apparatus further includes:

[0910] The sending module is used to send at least one of second information and second indication information to the information source device. The second indication information is used to indicate that the second information is obtained based on AI unit prediction. The second information is scheduling information related to the transmission of the information source device. The second information is obtained based on model inference results.

[0911] Optionally, the second information includes the model inference result, which includes at least one of the following:

[0912] The identifier of the target terminal; the TRP corresponding to the target terminal; the beam corresponding to the target terminal; the precoding matrix corresponding to the target terminal; the modulation and coding information corresponding to the target terminal; the predicted channel quality information corresponding to the target terminal; and the first predicted measurement.

[0913] The target terminal includes at least one of the following:

[0914] It is recommended to switch to the first terminal in the cell corresponding to the information source device; it is recommended to switch to the second terminal in the cell corresponding to the information acquisition device; it is recommended that the second terminal be scheduled by the information source device; it is recommended that the first terminal be scheduled by the information acquisition device.

[0915] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0916] Optionally, the second information further includes at least one of the following:

[0917] The information acquisition device can provide an independently pre-encoded identifier for the second terminal;

[0918] The information acquisition device can provide a jointly pre-coded identifier for the second terminal;

[0919] The identifier of the first terminal is recommended to be independently pre-encoded by the information source device;

[0920] The identifier of the first terminal is recommended to be jointly pre-coded by the information source device.

[0921] Optionally, when the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following:

[0922] The first measurement quantity; the first key performance indicator (KPI); the parameters associated with the first KPI; and the scheduling information of the second terminal in the cell corresponding to the information source device.

[0923] The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device.

[0924] The serving cell of the second terminal is the cell corresponding to the information source device.

[0925] Optionally, when the first operation includes training data collection, the first information includes at least one of model inference-related information and model monitoring-related information;

[0926] The information acquisition device performs a first operation on the target AI unit based on the first information, including any one of the following:

[0927] The information acquisition device generates a dataset based on the model inference-related information and the model monitoring-related information;

[0928] The information acquisition device performs model inference of the target AI unit based on the model inference-related information, and determines the environmental feedback or reward value of the target AI unit based on the model monitoring-related information; the information acquisition device generates a dataset based on the model inference result of the target AI unit and the environmental feedback or reward value.

[0929] Optionally, the sending module is further configured to: send third information to the information source device;

[0930] Alternatively, the sending module is further configured to: send third information to the information source device, and the receiving module is further configured to: receive confirmation information of the third information sent by the information source device;

[0931] The third information is used to request the first information;

[0932] The confirmation information is used to instruct the information source device to agree to the request corresponding to the third information.

[0933] Optionally, the third information includes at least one of the following: the service type that the information acquisition device can provide to the second terminal, and the service type that the information acquisition device requests from the information source device;

[0934] Wherein, the serving cell of the second terminal is the cell corresponding to the information source device;

[0935] The service type includes at least one of the following: DPS; ICBM; CJT; NCJT.

[0936] Referring to Figure 21, when the operation execution device is an information source device or a component of an information source device, the operation execution device 400 includes:

[0937] Sending module 401 is used to send first information to information acquisition device, wherein the first information is transmission-related information of the information source device;

[0938] The first information is used to perform the first operation of the target AI unit;

[0939] The target AI unit is associated with the information acquisition device;

[0940] The first operation includes at least one of the following:

[0941] Model inference, model monitoring, and training data collection.

[0942] Optionally, when the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following:

[0943] The first terminal has at least one measurement result in the cell corresponding to the information source device;

[0944] The second terminal has at least one measurement result in the cell corresponding to the information source device;

[0945] The second terminal is in at least one serving beam of the cell corresponding to the information source device;

[0946] The second terminal is in at least one TRP in the cell corresponding to the information source device;

[0947] The second terminal has at least one scheduling information in the cell corresponding to the information source device;

[0948] The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device.

[0949] The user throughput rate of the information source device;

[0950] First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device;

[0951] The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device;

[0952] The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device;

[0953] The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device;

[0954] TRP of candidate information source devices;

[0955] Beams of candidate information source devices;

[0956] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0957] Optionally, the beam and / or TRP in the model inference-related information are indicated by at least one of the following:

[0958] TCI status identifier; TCI identifier; SSB index; CSI-RS index; CRI.

[0959] Optionally, when the first operation includes model inference, the apparatus further includes:

[0960] The receiving module is configured to receive at least one of the second information and the second indication information sent by the information acquisition device, wherein the second indication information is used to indicate that the second information is obtained based on prediction by the AI ​​unit, the second information is scheduling information related to the transmission of the information source device, and the second information is obtained based on the model inference result.

[0961] Optionally, the second information includes the model inference result, which includes at least one of the following:

[0962] The identifier of the target terminal; the TRP corresponding to the target terminal; the beam corresponding to the target terminal; the precoding matrix corresponding to the target terminal; the modulation and coding information corresponding to the target terminal; the predicted channel quality information corresponding to the target terminal; and the first predicted measurement.

[0963] The target terminal includes at least one of the following:

[0964] It is recommended to switch to the first terminal in the cell corresponding to the information source device; it is recommended to switch to the second terminal in the cell corresponding to the information acquisition device; it is recommended that the second terminal be scheduled by the information source device; it is recommended that the first terminal be scheduled by the information acquisition device.

[0965] Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device.

[0966] Optionally, the second information further includes at least one of the following:

[0967] The information acquisition device can provide an independently pre-encoded identifier for the second terminal;

[0968] The information acquisition device can provide a jointly pre-coded identifier for the second terminal;

[0969] The identifier of the first terminal is recommended to be independently pre-encoded by the information source device;

[0970] The identifier of the first terminal is recommended to be jointly pre-coded by the information source device.

[0971] Optionally, when the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following:

[0972] The first measured quantity; the first KPI; the parameters associated with the first KPI; the actual scheduling information of the second terminal in the cell corresponding to the information source device;

[0973] The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device.

[0974] The serving cell of the second terminal is the cell corresponding to the information source device.

[0975] Optionally, the receiving module is further configured to: receive third information sent by the information acquisition device;

[0976] Alternatively, the receiving module is further configured to: receive third information sent by the information acquisition device, and the sending module is further configured to: send confirmation information of the third information to the information source device;

[0977] The third information is used to request the first information;

[0978] The confirmation information is used to instruct the information source device to agree to the request corresponding to the third information.

[0979] Optionally, the third information includes at least one of the following: the service type that the information acquisition device can provide to the second terminal, and the service type that the information acquisition device requests from the information source device;

[0980] Wherein, the serving cell of the second terminal is the cell corresponding to the information source device;

[0981] The service type includes at least one of the following: DPS; ICBM; CJT; NCJT.

[0982] The operation execution device provided in this application embodiment can implement the various processes implemented in the method embodiments of FIG3 and FIG19 and achieve the same technical effect. To avoid repetition, it will not be described again here.

[0983] As shown in Figure 22, this application embodiment also provides a communication device 500, including a processor 501 and a memory 502. The memory 502 stores a program or instructions that can run on the processor 501. For example, when the communication device 500 is a terminal, when the program or instructions are executed by the processor 501, they implement the various steps of the above-described operation execution method embodiment for the information acquisition device side, and achieve the same technical effect. When the communication device 500 is a network-side device, when the program or instructions are executed by the processor 501, they implement the various steps of the above-described operation execution method embodiment for the information source device side, and achieve the same technical effect. To avoid repetition, this will not be described again here.

[0984] This application also provides a network-side device, including a processor and a communication interface. The communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the steps of the method embodiments shown in FIG3 or FIG19. This network-side device embodiment corresponds to the above-described network-side device method embodiment. All implementation processes and methods of the above-described method embodiments can be applied to this network-side device embodiment and can achieve the same technical effect.

[0985] Specifically, this application embodiment also provides a network-side device, which may be the operation execution device shown in FIG20 or FIG21. As shown in FIG23, the network-side device 600 includes: an antenna 601, a radio frequency device 602, a baseband device 603, a processor 604, and a memory 605. The antenna 601 is connected to the radio frequency device 602. In the uplink direction, the radio frequency device 602 receives information through the antenna 601 and sends the received information to the baseband device 603 for processing. In the downlink direction, the baseband device 603 processes the information to be transmitted and sends it to the radio frequency device 602. The radio frequency device 602 processes the received information and transmits it through the antenna 601.

[0986] The method executed by the network-side device in the above embodiments can be implemented in the baseband device 603, which includes a baseband processor.

[0987] The baseband device 603 may include at least one baseband board, on which multiple chips are disposed, as shown in FIG23. One of the chips is, for example, a baseband processor, which is connected to the memory 605 via a bus interface to call the program or instructions in the memory 605 to execute the network-side device operation shown in the above method embodiment.

[0988] The network-side device may also include a network interface 606, such as a Common Public Radio Interface (CPRI).

[0989] In the case where the network-side device is an information acquisition device, the radio frequency device 602 is used to: receive first information sent by the information source device, wherein the first information is the transmission-related information of the information source device;

[0990] Processor 604 is used to perform a first operation on the target AI unit based on the first information;

[0991] The target AI unit is associated with the information acquisition device;

[0992] The first operation includes at least one of the following:

[0993] Model inference, model monitoring, and training data collection.

[0994] When the network-side device is an information source device, the radio frequency device 602 is used for:

[0995] Send first information to the information acquisition device, wherein the first information is transmission-related information of the information source device;

[0996] The first information is used to perform the first operation of the target AI unit;

[0997] The target AI unit is associated with the information acquisition device;

[0998] The first operation includes at least one of the following:

[0999] Model inference, model monitoring, and training data collection.

[1000] In addition, the network-side device 600 of this application embodiment also includes: a program or instructions stored in a memory 605 and executable on a processor 604. The processor 604 calls the program or instructions in the memory 605 to execute the methods executed by the modules shown in FIG20 or FIG21 and achieve the same technical effect. To avoid repetition, it will not be described in detail here.

[1001] Specifically, this application also provides a network-side device. As shown in FIG24, the network-side device 700 includes a processor 701, a network interface 702, and a memory 703. The network-side device may be the operation execution device shown in FIG20 or FIG21. The network interface 702 is, for example, a common public radio interface (CPRI).

[1002] In the case where the network-side device is an information acquisition device, the network interface 702 is used to: receive first information sent by the information source device, wherein the first information is the transmission-related information of the information source device;

[1003] Processor 701 is used to perform a first operation on the target AI unit based on the first information;

[1004] The target AI unit is associated with the information acquisition device;

[1005] The first operation includes at least one of the following:

[1006] Model inference, model monitoring, and training data collection.

[1007] When the network-side device is an information source device, network interface 702 is used for:

[1008] Send first information to the information acquisition device, wherein the first information is transmission-related information of the information source device;

[1009] The first information is used to perform the first operation of the target AI unit;

[1010] The target AI unit is associated with the information acquisition device;

[1011] The first operation includes at least one of the following:

[1012] Model inference, model monitoring, and training data collection.

[1013] In addition, the network-side device 700 of this application embodiment also includes: a program or instructions stored in a memory 703 and executable on a processor 701. The processor 701 calls the program or instructions in the memory 703 to execute the methods executed by the modules shown in FIG20 or FIG21 and achieve the same technical effect. To avoid repetition, it will not be described in detail here.

[1014] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described operation execution method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[1015] The processor mentioned above is either the processor in the terminal described in the above embodiments or the processor in the network-side device. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk. In some examples, the readable storage medium may be a non-transient readable storage medium.

[1016] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described operation execution method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[1017] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[1018] This application also provides a computer program / program product, which is stored in a storage medium and executed by at least one processor to implement the various processes of the above-described operation execution method embodiments, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[1019] This application embodiment also provides a wireless communication system, including: an information source device and an information acquisition device. The information source device can be used to execute the steps of the operation execution method applied to the information source device as described above, and the information acquisition device can be used to execute the steps of the operation execution method applied to the information acquisition device as described above.

[1020] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[1021] From the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of computer software products plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware. The computer software product is stored in a storage medium (such as ROM, RAM, magnetic disk, optical disk, etc.), and the computer software product includes several instructions to cause the terminal or network-side device to execute the methods described in the various embodiments of this application.

[1022] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other implementations under the guidance of this application without departing from the spirit and scope of the claims. All of these implementations are within the protection scope of this application.

Claims

An operation execution method, comprising: The information acquisition device receives first information sent by the information source device, wherein the first information is transmission-related information of the information source device; The information acquisition device performs a first operation on the target artificial intelligence (AI) unit based on the first information; The target AI unit is associated with the information acquisition device; The first operation includes at least one of the following: Model inference, model monitoring, and training data collection. The method according to claim 1, wherein, When the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following: The first terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal is in at least one serving beam of the cell corresponding to the information source device; The second terminal is in at least one TRP in the cell corresponding to the information source device; The second terminal has at least one scheduling information in the cell corresponding to the information source device; The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device. The user throughput rate of the information source device; First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device; The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device; The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device; The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device; TRP of candidate information source devices; Beams of candidate information source devices; Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device. The method according to claim 2, wherein, The beam and / or TRP in the model inference-related information are indicated by at least one of the following: Transmission Configuration Indicator (TCI) Status Identifier; TCI Identifier; Synchronization Signal Block (SSB) Index; Channel State Information Reference Signal (CSI-RS) Index; CSI-RS Resource Indicator (CRI). The method according to any one of claims 1-3, wherein, If the first operation includes model inference, the method further includes: The information acquisition device sends at least one of second information and second indication information to the information source device. The second indication information is used to indicate that the second information is obtained based on AI unit prediction. The second information is scheduling information related to the transmission of the information source device. The second information is obtained based on model inference results. The method according to claim 4, wherein, The second information includes the model inference result, which includes at least one of the following: The identifier of the target terminal; the TRP corresponding to the target terminal; the beam corresponding to the target terminal; the precoding matrix corresponding to the target terminal; the modulation and coding information corresponding to the target terminal; the predicted channel quality information corresponding to the target terminal; and the first predicted measurement. The target terminal includes at least one of the following: It is recommended to switch to the first terminal in the cell corresponding to the information source device; it is recommended to switch to the second terminal in the cell corresponding to the information acquisition device; it is recommended that the information source device perform scheduling on the second terminal. The first terminal recommended for scheduling by the information acquisition device; Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device. The method according to claim 5, wherein, The second information also includes at least one of the following: The information acquisition device can provide an independently pre-encoded identifier for the second terminal; The information acquisition device can provide a jointly pre-coded identifier for the second terminal; The identifier of the first terminal is recommended to be independently pre-encoded by the information source device; The identifier of the first terminal is recommended to be jointly pre-coded by the information source device. The method according to any one of claims 1-6, wherein, When the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following: The first measurement quantity; the first key performance indicator (KPI); the parameters associated with the first KPI; and the scheduling information of the second terminal in the cell corresponding to the information source device. The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device. The serving cell of the second terminal is the cell corresponding to the information source device. The method according to any one of claims 1-7, wherein, When the first operation includes training data collection, the first information includes at least one of model inference-related information and model monitoring-related information; The information acquisition device performs a first operation on the target AI unit based on the first information, including any one of the following: The information acquisition device generates a dataset based on the model inference-related information and the model monitoring-related information; The information acquisition device performs model inference of the target AI unit based on the model inference-related information, and determines the environmental feedback or reward value of the target AI unit based on the model monitoring-related information; the information acquisition device generates a dataset based on the model inference result of the target AI unit and the environmental feedback or reward value. The method according to any one of claims 1-8, wherein, The method further includes: The information acquisition device sends third information to the information source device; Alternatively, the information acquisition device may send third information to the information source device and receive confirmation information of the third information sent by the information source device; The third information is used to request the first information; The confirmation information is used to instruct the information source device to agree to the request corresponding to the third information. The method according to claim 9, wherein, The third information includes at least one of the following: the service type that the information acquisition device can provide to the second terminal, and the service type that the information acquisition device requests from the information source device; Wherein, the serving cell of the second terminal is the cell corresponding to the information source device; The service type includes at least one of the following: DPS; ICBM; CJT; NCJT. An operation execution method, comprising: The information source device sends first information to the information acquisition device, wherein the first information is transmission-related information of the information source device. The first information is used to perform the first operation of the target AI unit; The target AI unit is associated with the information acquisition device; The first operation includes at least one of the following: Model inference, model monitoring, and training data collection. The method according to claim 11, wherein, When the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following: The first terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal is in at least one serving beam of the cell corresponding to the information source device; The second terminal is in at least one TRP in the cell corresponding to the information source device; The second terminal has at least one scheduling information in the cell corresponding to the information source device; The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device. The user throughput rate of the information source device; First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device; The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device; The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device; The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device; TRP of candidate information source devices; Beams of candidate information source devices; Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device. The method according to claim 12, wherein, The beam and / or TRP in the model inference-related information are indicated by at least one of the following: TCI status identifier; TCI identifier; SSB index; CSI-RS index; CRI. The method according to any one of claims 11-13, wherein, If the first operation includes model inference, the method further includes: The information source device receives at least one of the second information and the second indication information sent by the information acquisition device. The second indication information is used to indicate that the second information is obtained based on AI unit prediction. The second information is scheduling information related to the transmission of the information source device. The second information is obtained based on the model inference result. The method according to claim 14, wherein, The second information includes the model inference result, which includes at least one of the following: The identifier of the target terminal; the TRP corresponding to the target terminal; the beam corresponding to the target terminal; the precoding matrix corresponding to the target terminal; the modulation and coding information corresponding to the target terminal; the predicted channel quality information corresponding to the target terminal; and the first predicted measurement. The target terminal includes at least one of the following: It is recommended to switch to the first terminal in the cell corresponding to the information source device; it is recommended to switch to the second terminal in the cell corresponding to the information acquisition device; it is recommended that the information source device perform scheduling on the second terminal. The first terminal recommended for scheduling by the information acquisition device; Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device. The method according to claim 15, wherein, The second information also includes at least one of the following: The information acquisition device can provide an independently pre-encoded identifier for the second terminal; The information acquisition device can provide a jointly pre-coded identifier for the second terminal; The identifier of the first terminal is recommended to be independently pre-encoded by the information source device; The identifier of the first terminal is recommended to be jointly pre-coded by the information source device. The method according to any one of claims 11-16, wherein, When the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following: The first measured quantity; the first KPI; the parameters associated with the first KPI; the actual scheduling information of the second terminal in the cell corresponding to the information source device; The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device. The serving cell of the second terminal is the cell corresponding to the information source device. The method according to any one of claims 11-17, wherein, The method further includes: The information source device receives third information sent by the information acquisition device; Alternatively, the information source device receives the third information sent by the information acquisition device and sends confirmation information of the third information to the information source device; The third information is used to request the first information; The confirmation information is used to instruct the information source device to agree to the request corresponding to the third information. The method according to claim 18, wherein, The third information includes at least one of the following: the service type that the information acquisition device can provide to the second terminal, and the service type that the information acquisition device requests from the information source device; Wherein, the serving cell of the second terminal is the cell corresponding to the information source device; The service type includes at least one of the following: DPS; ICBM; CJT; NCJT. An operation execution device, comprising: The receiving module is used to receive first information sent by the information source device, wherein the first information is transmission-related information of the information source device; The processing module is used to perform a first operation on the target AI unit based on the first information; The target AI unit is associated with the information acquisition device; The first operation includes at least one of the following: Model inference, model monitoring, and training data collection. The apparatus according to claim 20, wherein, When the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following: The first terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal is in at least one serving beam of the cell corresponding to the information source device; The second terminal is in at least one TRP in the cell corresponding to the information source device; The second terminal has at least one scheduling information in the cell corresponding to the information source device; The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device. The user throughput rate of the information source device; First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device; The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device; The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device; The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device; TRP of candidate information source devices; Beams of candidate information source devices; Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device. The apparatus according to claim 20 or 21, wherein, When the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following: The first measurement quantity; the first key performance indicator (KPI); the parameters associated with the first KPI; and the scheduling information of the second terminal in the cell corresponding to the information source device. The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device. The serving cell of the second terminal is the cell corresponding to the information source device. An operation execution device, comprising: The sending module is used to send first information to the information acquisition device, wherein the first information is transmission-related information of the information source device; The first information is used to perform the first operation of the target AI unit; The target AI unit is associated with the information acquisition device; The first operation includes at least one of the following: Model inference, model monitoring, and training data collection. The apparatus according to claim 23, wherein, When the first operation includes model inference, the first information includes model inference-related information, which includes at least one of the following: The first terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal has at least one measurement result in the cell corresponding to the information source device; The second terminal is in at least one serving beam of the cell corresponding to the information source device; The second terminal is in at least one TRP in the cell corresponding to the information source device; The second terminal has at least one scheduling information in the cell corresponding to the information source device; The second terminal receives at least one scheduling feedback message from the cell corresponding to the information source device. The user throughput rate of the information source device; First indication information, the first indication information is used to indicate whether the information source device accepts the scheduling information of the information acquisition device; The information source device is recommended to switch to the identifier of the second terminal in the cell corresponding to the information acquisition device; The information source device recommends the identification of the second terminal that is independently transmitted uplink or downlink by the TRP of the information acquisition device; The information source device recommends that the identification of the second terminal be jointly transmitted uplink or downlink by the TRP of the information acquisition device; TRP of candidate information source devices; Beams of candidate information source devices; Wherein, the serving cell of the first terminal is the cell corresponding to the information acquisition device, and the serving cell of the second terminal is the cell corresponding to the information source device. The apparatus according to claim 23 or 24, wherein, When the first operation includes model monitoring, the first information includes model monitoring-related information, which includes at least one of the following: The first measurement quantity; the first key performance indicator (KPI); the parameters associated with the first KPI; and the scheduling information of the second terminal in the cell corresponding to the information source device. The first KPI includes at least one of the following: the prediction accuracy of the first measurement; the transmission rate information of the second terminal; and the scheduling fairness of the information source device. The serving cell of the second terminal is the cell corresponding to the information source device. A network-side device includes a processor and a memory, the memory storing a program or instructions executable on the processor, the program or instructions, when executed by the processor, implementing the steps of the operation execution method as claimed in any one of claims 1 to 10, or implementing the steps of the operation execution method as claimed in any one of claims 11 to 19. A readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the operation execution method as claimed in any one of claims 1 to 10, or implement the steps of the operation execution method as claimed in any one of claims 11 to 19.

Images