WO2023186048A1 - 一种ai服务信息获取方法、装置及系统 - Google Patents

一种ai服务信息获取方法、装置及系统 Download PDF

Info

Publication number
WO2023186048A1
WO2023186048A1 PCT/CN2023/085246 CN2023085246W WO2023186048A1 WO 2023186048 A1 WO2023186048 A1 WO 2023186048A1 CN 2023085246 W CN2023085246 W CN 2023085246W WO 2023186048 A1 WO2023186048 A1 WO 2023186048A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
node
service
identification information
subscription
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Ceased
Application number
PCT/CN2023/085246
Other languages
English (en)
French (fr)
Inventor
乔云飞
王坚
张公正
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP23778413.7A priority Critical patent/EP4492856A4/en
Publication of WO2023186048A1 publication Critical patent/WO2023186048A1/zh
Priority to US18/901,120 priority patent/US20250024276A1/en
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/08—Testing, supervising or monitoring using real traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/5003—Managing SLA; Interaction between SLA and QoS
    • H04L41/5009—Determining service level performance parameters or violations of service level contracts, e.g. violations of agreed response time or mean time between failures [MTBF]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0053—Allocation of signalling, i.e. of overhead other than pilot signals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/10—Scheduling measurement reports ; Arrangements for measurement reports
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W84/00—Network topologies
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W36/00—Hand-off or reselection arrangements
    • H04W36/12—Reselecting a serving backbone network switching or routing node

Definitions

  • This application relates to the field of communication technology, and in particular to a method, device and system for obtaining AI service information.
  • the intelligence of wireless networks is an important evolutionary trend.
  • the fifth generation mobile communication (5G) network has introduced artificial intelligence (AI) capabilities.
  • AI nodes are deployed in networks with AI capabilities.
  • AI nodes can provide AI services for terminal devices, thus Bring a better business experience to users.
  • the terminal device needs to dynamically select an appropriate AI node, so that the selected AI node can provide the terminal device with AI services that meet the needs.
  • the appropriate AI node can be selected based on the relevant information about the AI services provided by each AI node, that is, the AI service information. Therefore, in order to enable terminal devices in the network to dynamically select AI nodes and optimize the overall network performance, a feasible solution is to enable the terminal devices and AI nodes in the network to exchange AI service information, thereby facilitating the exchange of AI nodes. collaborative management. However, there is currently no solution for how to enable terminal devices and AI nodes in the network to exchange AI service information.
  • Embodiments of the present application provide a method, device and system for obtaining AI service information, which are used to solve the problem of how to realize the exchange of AI service information between terminal devices and AI nodes in the network.
  • the communication device (subject) that executes the method may be a network device corresponding to the first AI node, or may be a component of the network device corresponding to the first AI node, such as a processing device, chip, or chip system.
  • the following description takes the execution subject as the first AI node as an example.
  • the method includes: the first AI node sends measurement configuration information to the terminal device; the measurement configuration information is used to indicate measuring the AI service of the AI node to be measured; the first AI node receives from The measurement report of the terminal device; the measurement report includes candidate AI service information; the candidate AI service information is the AI service information that satisfies the first condition among the one or more AI service information obtained by measuring the AI node to be measured according to the measurement configuration information; first The conditions are carried in the measurement configuration information.
  • AI nodes can obtain AI service information that meets certain conditions of other AI nodes through interaction with terminal devices, which facilitates collaborative management among different AI nodes in the AI network.
  • the candidate AI service information includes AI service identification information and performance information.
  • the AI service identification information is used to identify the AI service corresponding to the candidate AI service information
  • the performance information is used to characterize the candidate AI. Performance of the AI service corresponding to the service information.
  • the method also includes: the first AI node, according to the AI service identification information and performance information, respectively determines the candidate AI corresponding to each AI service among the one or more AI services corresponding to the candidate AI service information.
  • the target AI service information in the service information is added to the local AI service list, where the target AI service information is the AI service with the best performance among the candidate AI service information corresponding to each AI service. Excellent candidate AI service information. Based on this solution, from one or more candidate AI service information corresponding to the AI service, the corresponding candidate AI service information with the best performance can be selected and added to the local AI service list.
  • the method further includes: the first AI node adds the target AI service information in the candidate AI service information to the local AI service list, where the target AI service information is the candidate AI In the service information, candidate AI service information is different from the AI service information recorded in the local AI service list. Based on this solution, candidate AI service information that is different from the locally recorded AI service information can be added to the local AI service list to ensure that the same candidate AI service information does not appear in the local AI service list.
  • the candidate AI service information includes AI node identification information, and the AI node identification information is used to identify the AI node to be measured corresponding to the candidate AI service information.
  • the method also includes: a first AI node Determine the AI node corresponding to the one or more AI service information according to the AI node identification information included in the one or more AI service information in the local AI service list; the first AI node obtains the AI corresponding to the one or more AI service information The neural network model of the node; the neural network model is used to implement AI services. Based on this solution, the neural network model can be obtained from the corresponding AI node to implement AI services based on the AI service information recorded in the local AI service list.
  • the candidate AI service information includes model identification information
  • the model identification information is used to identify the neural network model corresponding to the candidate AI service information
  • the first AI node obtains one or more AI services
  • the neural network model of the AI node corresponding to the information includes: the first AI node obtains the neural network model corresponding to the model identification information included in the one or more AI service information from the AI node corresponding to the one or more AI service information. Based on this solution, a way to obtain the neural network model from the AI node is provided.
  • the first AI node obtains the neural network model of the AI node corresponding to one or more AI service information, including: the first AI node obtains the neural network model of the AI node corresponding to one or more AI service information from Obtain the preconfigured neural network model on the AI node. Based on this solution, a way to obtain the neural network model from the AI node is provided.
  • the measurement configuration information includes identification information of the AI node to be measured and identification information of the AI service that needs to be measured.
  • the candidate AI service information includes at least one of the following information: AI service identification information, AI node identification information, model identification information, model parameter information, performance information and timeliness information; where , the AI service identification information is used to identify the AI service corresponding to the candidate AI service information, the AI node identification information is used to identify the AI node to be measured corresponding to the candidate AI service information, and the model identification information is used to identify the neural network model corresponding to the candidate AI service information.
  • the model parameter information is the parameter information of the neural network model corresponding to the candidate AI service information
  • the performance information is used to characterize the performance of the AI service corresponding to the candidate AI service information
  • the aging information is used to characterize the validity time of the candidate AI service information.
  • the first AI node receives the measurement report from the terminal device, including: the first AI node receives the measurement report from the terminal device through the second network device, and the second network device is A network device corresponding to the second AI node among the AI nodes to be measured; wherein the candidate AI service information is the candidate AI service information corresponding to the second AI node.
  • the terminal device can send the corresponding measurement report to the first AI node through the AI node to be measured.
  • the measurement report also includes identification information of the first AI node; the identification information of the first AI node is used by the second network device to send the measurement report to the first AI node.
  • a method for obtaining AI service information is provided.
  • the communication device (main body) that executes the method can be a terminal device or a component of the terminal device, such as a processor, chip, or chip system of the terminal device.
  • the following description takes the execution subject as a terminal device as an example.
  • the method includes: the terminal device receives measurement configuration information from the first AI node; the measurement configuration information is used to indicate measuring the AI service of the AI node to be measured; the terminal device sends a request to the first AI node.
  • AI nodes can obtain AI service information that meets certain conditions of other AI nodes through interaction with terminal devices, which facilitates collaborative management among different AI nodes in the AI network.
  • the measurement configuration information includes identification information of the AI node to be measured and identification information of the AI service that needs to be measured.
  • the candidate AI service information includes at least one of the following information: AI service identification information, AI node identification information, model identification information, model parameter information, performance information and timeliness information; where , the AI service identification information is used to identify the AI service corresponding to the candidate AI service information, the AI node identification information is used to identify the AI node to be measured corresponding to the candidate AI service information, and the model identification information is used to identify the neural network model corresponding to the candidate AI service information.
  • the model parameter information is the parameter information of the neural network model corresponding to the candidate AI service information
  • the performance information is used to characterize the performance of the AI service corresponding to the candidate AI service information
  • the aging information is used to characterize the validity time of the candidate AI service information.
  • the terminal device sends a measurement report to the first AI node, including:
  • the terminal device sends a measurement report to the first AI node through a second network device.
  • the second network device is a network device corresponding to the second AI node among the AI nodes to be measured; wherein the candidate AI service information is corresponding to the second AI node.
  • Candidate AI service information is a network device corresponding to the second AI node among the AI nodes to be measured.
  • the measurement report also includes identification information of the first AI node; the identification information of the first AI node is used by the second network device to send the measurement report to the first AI node.
  • a method for obtaining AI service information is provided.
  • the communication device (subject) that executes the method may be a network device corresponding to the third AI node, or may be a component of the network device corresponding to the third AI node, such as a processing device, chip, or chip system.
  • the following description takes the execution subject as the third AI node as an example.
  • the method includes: the third AI node receives subscription information from the terminal device; the subscription information is used to request to subscribe to an AI service that satisfies the second condition in the subscription area; the third AI The node sends subscription information to one or more fourth AI nodes in the subscription area, so that the fourth AI node sends a subscription result to the terminal device according to the subscription information; the subscription result includes the target AI service information and the AI corresponding to the target AI service information.
  • the service meets the second condition.
  • the terminal device can subscribe to the desired AI service from the AI node and obtain the AI service information corresponding to the AI service that meets the conditions, so that the terminal device can subscribe to the AI service in the AI network. AI services that meet the requirements.
  • the subscription information is identified by the first identification information. Based on this solution, a first identification information for identifying subscription information is defined.
  • the first identification information is generated by the third AI node; or the first identification information is generated by the terminal device. Based on this solution, two methods for generating first identification information are provided. Mode.
  • the method further includes: the third AI node determines that the terminal device is currently located in the first notification area, and sets the first identifier to The information is associated with the second identification information; the second identification information is used to identify the first notification area; the third AI node sends the second identification information associated with the first identification information to one or more fourth AI nodes.
  • the third AI node may send the second identification information associated with the first identification information to the fourth AI node, so that the fourth AI node may The second identification information associated with the identification information indicates that the terminal device that sent the subscription information is currently located in the first notification area.
  • the first notification area is set according to cells and/or beams. Based on this solution, a way to plan notification areas is provided.
  • the subscription information includes at least one of the following information: AI service identification information used to identify the AI service requested to be subscribed, user identification information used to identify the terminal device, and characterization information.
  • the service level information of the second condition is used to characterize the subscription area information, model identification information, model parameter information and subscription validity period information of the subscription area; among which, the model identification information is used to identify the neural network model corresponding to the AI service subscribed by the subscription information request , the model parameter information is the parameter information of the neural network model corresponding to the AI service subscribed by the subscription information request, and the subscription validity period information is used to represent the validity time of the subscription information.
  • the subscription area information is determined by the terminal device; or the subscription area information is determined by the third AI node. Based on this solution, two methods of determining subscription information are provided.
  • the fourth aspect provides an AI service information method.
  • the communication device (main body) that executes the method can be a terminal device or a component of the terminal device, such as a processor, chip, or chip system of the terminal device.
  • the following description takes the execution subject as a terminal device as an example.
  • the method includes: the terminal device sends subscription information to the third AI node; the subscription information is used to request to subscribe to an AI service that meets the second condition in the subscription area; the terminal device receives data from the subscription area The subscription result of one or more fourth AI nodes in the fourth AI node; the subscription result includes target AI service information, and the AI service corresponding to the target AI service information satisfies the second condition.
  • the terminal device can subscribe to the desired AI service from the AI node and obtain the AI service information corresponding to the AI service that meets the conditions, so that the terminal device can subscribe to the AI service in the AI network.
  • AI services that meet the requirements.
  • the subscription information is identified by the first identification information. Based on this solution, a first identification information for identifying subscription information is defined.
  • the first identification information is generated by the third AI node; or the first identification information is generated by the terminal device. Based on this solution, two methods of generating first identification information are provided.
  • the method further includes: the terminal device obtains second identification information associated with the first identification information; the second identification information Used to identify the first notification area where the terminal device is currently located. Based on this solution, the terminal device can obtain the second identification information used to identify the notification area currently located.
  • the method further includes: the terminal device obtains third identification information; the third identification information is used to identify the second notification area where the terminal device is currently located; and the terminal device determines the third identification information is inconsistent with the second identification information, and sends an update request to one or more fourth AI nodes.
  • the update request Used to request to associate the first identification information with the third identification information.
  • the terminal device can initiate an update request to the fourth AI node, so that the fourth AI node can communicate with the first notification area.
  • the second identification information associated with the identification information is updated to the third identification information, and it is determined that the terminal device is currently located in the second notification area.
  • the terminal device obtaining the third identification information includes: the terminal device receiving the third identification information from the third network device; and the third network device is located in the second notification area. Based on this solution, the terminal device can obtain the third identification information through the network device in the second notification area.
  • the first notification area is set according to cells and/or beams. Based on this solution, a way to plan notification areas is provided.
  • the subscription information includes at least one of the following information: AI service identification information used to identify the AI service requested to be subscribed, user identification information used to identify the terminal device,
  • the service level information of the second condition is used to characterize the subscription area information, model identification information, model parameter information and subscription validity period information of the subscription area; among which, the model identification information is used to identify the neural network model corresponding to the AI service subscribed by the subscription information request , the model parameter information is the parameter information of the neural network model corresponding to the AI service subscribed by the subscription information request, and the subscription validity period information is used to represent the validity time of the subscription information.
  • the subscription area information is determined by the terminal device; or the subscription area information is determined by the third AI node. Based on this solution, two methods of determining subscription information are provided.
  • a communication device for implementing the various methods mentioned above.
  • the communication device may be the network device corresponding to the first AI node in the first aspect, or a device including the network device corresponding to the first AI node, or a device included in the network device corresponding to the first AI node, such as a chip;
  • the communication device may be the terminal device in the second aspect, or a device including the terminal device, or a device included in the terminal device; or the communication device may be the third AI node in the third aspect.
  • the corresponding network device, or a device including the network device corresponding to the third AI node, or a device included in the network device corresponding to the third AI node, such as a chip; or, the communication device may be the terminal device in the fourth aspect. , or a device including the above-mentioned terminal equipment, or a device included in the above-mentioned terminal equipment; the communication device includes a module, unit, or means (means) corresponding to the above method, and the module, unit, or means can be implemented by hardware, Software implementation, or corresponding software implementation executed by hardware.
  • the hardware or software includes one or more modules or units corresponding to the above functions.
  • a sixth aspect provides a communication device, including: a processor and a memory; the memory is used to store computer instructions, the processor is used to execute the instructions stored in the memory, and when the processor executes the instructions, the communication
  • the device performs the method described in any of the above aspects.
  • the communication device may be the network device corresponding to the first AI node in the first aspect, or a device including the network device corresponding to the first AI node, or a device included in the network device corresponding to the first AI node, such as a chip;
  • the communication device may be the terminal device in the above second aspect, or a device including the above terminal device, or a device included in the above terminal device, such as a chip; or the communication device may be the third device in the above third aspect.
  • the network equipment corresponding to the three AI nodes, or the device including the network equipment corresponding to the third AI node, or the device included in the network equipment corresponding to the third AI node, such as a chip; or, the communication device can be the fourth aspect mentioned above terminal equipment, or a device including the above-mentioned terminal equipment, or a device included in the above-mentioned terminal equipment, such as a chip.
  • a communication device including: a processor and an interface circuit, the interface circuit being used to communicate with The module communicates outside the communication device; the processor is used to run computer programs or instructions to perform the method described in any of the above aspects.
  • the communication device may be the network device corresponding to the first AI node in the first aspect, or a device including the network device corresponding to the first AI node, or a device included in the network device corresponding to the first AI node, such as a chip;
  • the communication device may be the terminal device in the above second aspect, or a device including the above terminal device, or a device included in the above terminal device, such as a chip; or the communication device may be the third device in the above third aspect.
  • the network equipment corresponding to the three AI nodes, or the device including the network equipment corresponding to the third AI node, or the device included in the network equipment corresponding to the third AI node, such as a chip; or, the communication device can be the fourth aspect mentioned above terminal equipment, or a device including the above-mentioned terminal equipment, or a device included in the above-mentioned terminal equipment, such as a chip.
  • the interface circuit can be a code/data read-write interface circuit, which is used to receive computer execution instructions (computer execution instructions are stored in the memory, may be read directly from the memory, or may pass through other devices) and transmitted to the A processor, such that the processor executes computer execution instructions to perform the method described in any of the above aspects.
  • the communication device may be a chip or a system on a chip.
  • An eighth aspect provides a communication device, including: a processor; the processor is configured to be coupled to a memory, and after reading instructions in the memory, execute the method as described in any of the above aspects according to the instructions.
  • the communication device may be the network device corresponding to the first AI node in the first aspect, or a device including the network device corresponding to the first AI node, or a device included in the network device corresponding to the first AI node, such as a chip;
  • the communication device may be the terminal device in the above second aspect, or a device including the above terminal device, or a device included in the above terminal device, such as a chip; or the communication device may be the third device in the above third aspect.
  • the network equipment corresponding to the three AI nodes, or the device including the network equipment corresponding to the third AI node, or the device included in the network equipment corresponding to the third AI node, such as a chip; or, the communication device can be the fourth aspect mentioned above terminal equipment, or a device including the above-mentioned terminal equipment, or a device included in the above-mentioned terminal equipment, such as a chip.
  • a computer-readable storage medium Instructions are stored in the computer-readable storage medium, and when run on a communication device, the communication device can perform the method described in any of the above aspects.
  • the communication device may be the network device corresponding to the first AI node in the first aspect, or a device including the network device corresponding to the first AI node, or a device included in the network device corresponding to the first AI node, such as a chip;
  • the communication device may be the terminal device in the above second aspect, or a device including the above terminal device, or a device included in the above terminal device, such as a chip; or the communication device may be the third device in the above third aspect.
  • the network equipment corresponding to the three AI nodes, or the device including the network equipment corresponding to the third AI node, or the device included in the network equipment corresponding to the third AI node, such as a chip; or, the communication device can be the fourth aspect mentioned above terminal equipment, or a device including the above-mentioned terminal equipment, or a device included in the above-mentioned terminal equipment, such as a chip.
  • a computer program product containing instructions which, when run on a communication device, enables the communication device to perform the method described in any of the above aspects.
  • the communication device may be the network device corresponding to the first AI node in the first aspect, or a device including the network device corresponding to the first AI node, or a device included in the network device corresponding to the first AI node, such as a chip;
  • the communication device may be the terminal device in the above second aspect, or a device including the above terminal device, or a device included in the above terminal device.
  • the communication device may be a network device corresponding to the third AI node in the above third aspect, or a device including a network device corresponding to the third AI node, or in a network device corresponding to the third AI node A device included, such as a chip; alternatively, the communication device may be the terminal device in the fourth aspect, or a device including the above terminal device, or a device included in the above terminal device, such as a chip.
  • An eleventh aspect provides a communication device (for example, the communication device may be a chip or a chip system).
  • the communication device includes a processor for implementing the functions involved in any of the above aspects.
  • the communication device further includes a memory, which is used to store necessary program instructions and data.
  • the communication device is a chip system, it may be composed of a chip, or may include a chip and other discrete devices.
  • a communication system which includes network equipment and terminal equipment.
  • the network device can perform the method described in the first aspect, and the terminal device can perform the method described in the second aspect; or, the network device can perform the method described in the third aspect, and the terminal device can perform the method described in the fourth aspect. method described.
  • Figure 1a is a schematic diagram 1 of the architecture of a communication system provided by an embodiment of the present application.
  • Figure 1b is a schematic diagram 2 of the architecture of a communication system provided by an embodiment of the present application.
  • Figure 2 is a schematic structural diagram of a communication device provided by an embodiment of the present application.
  • Figure 3 is an interactive schematic diagram of an AI service information acquisition method provided by an embodiment of the present application.
  • Figure 4 is a schematic diagram of communication between a first AI node and a RAN device provided by an embodiment of the present application;
  • Figure 5 is an interactive schematic diagram of a method for obtaining a neural network model provided by an embodiment of the present application
  • Figure 6 is an interactive schematic diagram of another method for obtaining a neural network model provided by an embodiment of the present application.
  • Figure 7 is an interactive schematic diagram of another method for obtaining AI service information provided by an embodiment of the present application.
  • Figure 8 is a schematic diagram of a tracking area
  • Figure 9 is a schematic diagram of a method for mobility management of terminal devices in an AI network provided by an embodiment of the present application.
  • Figure 10 is a schematic diagram of an exemplary AI service information acquisition method provided by an embodiment of the present application.
  • Figure 11 is a schematic structural diagram of another communication device provided by an embodiment of the present application.
  • Figure 12 is a schematic structural diagram of another communication device provided by an embodiment of the present application.
  • AI nodes are deployed in networks with AI capabilities.
  • the network can provide AI services such as data analysis, speech recognition, or image recognition to terminal devices through the neural network model on the AI node.
  • AI nodes can also support functions such as data collection and training of neural network models.
  • the network can use AI nodes to replace the numerical formula-based methods in the original network functions with AI-based methods, thereby improving the efficiency of network resource usage and improving user business experience.
  • AI nodes can be deployed independently on network equipment dedicated to realizing AI node functions, or they can be deployed as functional modules on existing network equipment, such as on radio access network (RAN) equipment. . It can be understood that when the AI node is used as an independent logical unit (in hardware or software), the AI node can form an AI network that is logically separated from the access network composed of RAN equipment.
  • RAN radio access network
  • a locally deployed AI node can only obtain data within the local observation range of the AI node.
  • the local AI node may not be able to provide optimal AI services for terminal devices.
  • the terminal device needs to dynamically select an appropriate AI node, so that the selected AI node can provide the terminal device with AI services that meet the needs.
  • the appropriate AI node can be selected based on the relevant information about the AI services provided by each AI node, that is, the AI service information.
  • a feasible solution is to enable the terminal devices and AI nodes in the AI network to exchange AI service information, thereby facilitating the AI node collaborative management.
  • AI service information there is currently no solution for how to enable terminal devices and AI nodes in the AI network to exchange AI service information.
  • embodiments of the present application provide a method for obtaining AI service information, which is used to realize the exchange of AI service information between terminal devices and AI nodes.
  • At least one of the following or similar expressions thereof refers to any combination of these items, including any combination of a single item (items) or a plurality of items (items).
  • at least one of a, b or c can mean: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, c Can be single or multiple.
  • words such as “first” and “second” are used to distinguish identical or similar items with basically the same functions and effects.
  • words such as “first” and “second” do not limit the number and execution order, and words such as “first” and “second” do not limit the number and execution order.
  • words such as “exemplary” or “for example” are used to represent examples, illustrations or explanations. Any embodiment or design described as “exemplary” or “such as” in the embodiments of the present application is not to be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as “exemplary” or “such as” is intended to present related concepts in a concrete manner that is easier to understand.
  • the technical solutions of the embodiments of this application can be used in various communication systems.
  • the communication system can be a 3rd generation partnership project (3GPP) communication system, for example, a long term evolution (LTE) system, or 5G system, vehicle to everything (V2X) system, or LTE and 5G hybrid networking system, or device-to-device (D2D) communication system, machine to machine (M2M) Communication systems, Internet of Things (IoT), and other next-generation communication systems.
  • 3GPP 3rd generation partnership project
  • LTE long term evolution
  • V2X vehicle to everything
  • LTE and 5G hybrid networking system or device-to-device (D2D) communication system, machine to machine (M2M) Communication systems, Internet of Things (IoT), and other next-generation communication systems.
  • the communication system can also be a non-3GPP communication system without limitation.
  • eMBB enhanced mobile broadband
  • URLLC ultra-reliable low-latency communication
  • MTC machine type communication
  • mMTC massive machine type communications
  • D2D V2X
  • IoT IoT and other communication scenarios.
  • this application provides a communication system to which this application is applicable.
  • the communication system may include one or more first AI nodes and one or more terminal devices, or the communication system may include one or more third AI nodes and one or more terminal devices.
  • the AI node can be independently deployed on a network device dedicated to realizing the AI node function.
  • the AI node, or the network device used to realize the AI node function is connected to the wireless access network.
  • the devices have a corresponding relationship, and the AI node can communicate with the corresponding wireless access network device, and thus can communicate with the terminal device through the corresponding wireless access network device.
  • the AI node and the radio access network device may have a one-to-one correspondence, or there may be a one-to-many correspondence in which one AI node corresponds to multiple radio access network devices.
  • the communication system 10 provided by the embodiment of the present application can be as shown in Figure 1a.
  • the first AI node 20 or the third AI node 20 communicates through the corresponding wireless Access network equipment communicates with terminal equipment 30.
  • the AI node can be deployed on the radio access network device as a functional module, and the radio access network device deployed by the AI node can also be called the radio access network device corresponding to the AI node.
  • the AI node can communicate with the terminal device through the wireless communication function provided by the corresponding wireless access network device.
  • the communication system 10 provided by the embodiment of the present application can be as shown in Figure 1b.
  • the first AI node 20 or the third AI node 20 is deployed on the radio access network device to communicate with the terminal device 30.
  • the communication system 10 may also include an AI node to be measured.
  • the first AI node 20 sends measurement configuration information to the terminal device 30, and the measurement configuration The information is used to indicate the AI service of the AI node to be measured.
  • the terminal device 30 measures the AI node to be measured according to the measurement configuration information to obtain one or more AI service information, and based on the first condition carried in the measurement configuration information, the one or more AI service information Determine the candidate AI service information that meets the first condition.
  • the first AI node 20 sends the candidate AI service information to the first AI node 20 through a measurement report.
  • the communication system 10 provided by this application may also include a fourth AI node.
  • the terminal device 30 sends subscription information to the third AI node 20, and the subscription information is Yu requests to subscribe to the AI service that meets the second condition in the subscription area.
  • the third AI node 20 sends the subscription information to each AI node in the subscription area, so that one or more fourth AI nodes in the subscription area send the subscription result to the terminal device according to the subscription information.
  • the result includes target AI service information, and the AI service corresponding to the target AI service information satisfies the second condition.
  • the wireless access network device in the embodiment of this application is a device that connects a terminal device to a wireless network. equipment.
  • the radio access network equipment may be a base station (base station), an evolved base station (evolved NodeB, eNodeB), a transmission reception point (TRP), or a next generation base station (next generation NodeB, gNB) in the 5G mobile communication system.
  • CU centralized unit
  • DU distributed unit
  • the embodiments of this application do not limit the specific technology and specific equipment form used by the wireless access network equipment.
  • the terminal device in the embodiment of the present application may be a device used to implement wireless communication functions.
  • the terminal can also be called user equipment (UE), mobile station, mobile terminal, etc.
  • Terminals can be mobile phones, tablets, computers with wireless transceiver functions, virtual reality terminal equipment, augmented reality terminal equipment, wireless terminals in industrial control, wireless terminals in driverless driving, wireless terminals in remote surgery, and smart grids.
  • the embodiments of this application do not limit the specific technology and specific equipment form used by the terminal equipment.
  • the radio access network device and the terminal device in the embodiment of the present application can communicate through licensed spectrum, communicate through unlicensed spectrum, or communicate through licensed spectrum and unlicensed spectrum at the same time.
  • Wireless access network equipment and terminal equipment can communicate through spectrum below 6 gigahertz (GHz) or above 6 GHz, and can also use spectrum below 6 GHz and spectrum above 6 GHz at the same time. communicate.
  • GHz gigahertz
  • the embodiments of the present application do not limit the spectrum resources used between the radio access network device and the terminal device.
  • the network equipment and terminal equipment in the embodiments of this application can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on water; they can also be deployed on aircraft, balloons and artificial satellites in the air.
  • the embodiments of this application do not limit the application scenarios of network devices and terminal devices.
  • the relevant functions of the AI node or terminal device involved in this application can be implemented by one device, or can be implemented by multiple devices together, or can be implemented by one or more functional modules in one device, or can be implemented by one or more
  • the chip may also be a system on chip (SOC) or a chip system.
  • SOC system on chip
  • the chip system may be composed of chips, or may include chips and other discrete devices. This is not specifically limited in the embodiments of the present application.
  • FIG. 2 shows a schematic structural diagram of a communication device 200 provided by an embodiment of the present application.
  • the communication device 200 includes one or more processors 201, communication lines 202, and at least one communication interface (FIG. 2 is only an example of including a communication interface 204 and a processor 201 for illustration).
  • Optional and may also include a memory 203.
  • the processor 201 can be a general central processing unit (CPU), a microprocessor, an application-specific integrated circuit (ASIC), or one or more processors used to control the execution of the program of the present application. integrated circuit.
  • CPU central processing unit
  • ASIC application-specific integrated circuit
  • Communication lines 202 may be used for communication between different components included in communication device 200 .
  • the communication interface 204 can be a transceiver module used to communicate with other devices or communication networks, such as Ethernet, without Wireless access networks (RAN), wireless local area networks (WLAN), etc.
  • the transceiver module may be a device such as a transceiver or a transceiver.
  • the communication interface 204 may also be a transceiver circuit located in the processor 201 to implement signal input and signal output of the processor.
  • the memory 203 may be a device with a storage function. For example, it can be read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (random access memory, RAM) or other types of things that can store information and instructions. Dynamic storage devices can also be electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disk storage, optical disc storage ( Including compressed optical discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or can be used to carry or store desired program code in the form of instructions or data structures and can be stored by a computer. any other medium, but not limited to this.
  • the memory may exist independently and be connected to the processor through the communication line 202 . Memory can also be integrated with the processor.
  • the memory 203 is used to store computer execution instructions for executing the solution of the present application, and is controlled by the processor 201 for execution.
  • the processor 201 is used to execute computer execution instructions stored in the memory 203, thereby implementing the method provided in the embodiment of the present application.
  • the processor 201 may also perform processing-related functions in the methods provided in the following embodiments of the present application, and the communication interface 204 is responsible for communicating with other devices or communication networks.
  • This application implements The example does not specifically limit this.
  • the computer-executed instructions in the embodiments of the present application may also be called application codes, which are not specifically limited in the embodiments of the present application.
  • the processor 201 may include one or more CPUs, such as CPU0 and CPU1 in FIG. 2 .
  • the communication device 200 may include multiple processors, such as the processor 201 and the processor 208 in FIG. 2 .
  • processors can be a single-core processor or a multi-core processor.
  • the processor here may include but is not limited to at least one of the following: central processing unit (CPU), microprocessor, digital signal processor (DSP), microcontroller unit (MCU), or artificial intelligence
  • CPU central processing unit
  • DSP digital signal processor
  • MCU microcontroller unit
  • computing devices such as processors that run software.
  • Each computing device may include one or more cores for executing software instructions to perform calculations or processing.
  • the communication device 200 may also include an output device 205 and an input device 206.
  • Output device 205 communicates with processor 201 and can display information in a variety of ways.
  • the output device 205 may be a liquid crystal display (LCD), a light emitting diode (LED) display device, a cathode ray tube (CRT) display device, or a projector (projector), etc.
  • Input device 206 communicates with processor 201 and may receive user input in a variety of ways.
  • the input device 206 may be a mouse, a keyboard, a touch screen device, a sensing device, or the like.
  • composition structure shown in Figure 2 does not constitute a limitation on the communication device.
  • the communication device may include more or less components than those shown in the figure, or a combination of Certain parts, or different arrangements of parts.
  • the components illustrated may be implemented in hardware, software, or a combination of software and hardware.
  • the AI service information acquisition method includes S301-S302:
  • the first AI node sends measurement configuration information to the terminal device; the measurement configuration information is used to indicate measuring the AI service of the AI node to be measured.
  • the first AI node receives a measurement report from the terminal device; the measurement report includes candidate AI service information; the candidate AI service information is one or more AI service information obtained by measuring the AI node to be measured according to the measurement configuration information and satisfies the first Conditional AI service information; the first condition is carried in the measurement configuration information.
  • the first AI node sends measurement configuration information to the terminal device through the corresponding network device. After receiving the measurement configuration information, the terminal device measures the AI service of the AI node to be measured according to the measurement configuration information.
  • the measurement configuration information includes the first condition, the identification information of the AI node to be measured, and the identification information of the AI service that needs to be measured, which are introduced separately below.
  • the first condition is used to indicate the conditions that the terminal device needs to meet to send (or report) a measurement report to the first AI node based on the measurement results after performing measurements based on the measurement configuration information. Therefore, the first condition can also be called It is the reporting condition for measurement report.
  • the measurement configuration information may include one or more first conditions.
  • the first condition can be a threshold that the performance of the AI service needs to meet. If the measured performance of the AI service meets the threshold, the terminal device can report the measurement result obtained by measuring the AI service to the first AI node. Otherwise, the terminal device can report the measurement result obtained by measuring the AI service to the first AI node. Not reported.
  • the identification information of the AI node to be measured is used by the terminal device to determine the AI node to be measured that needs to be measured, which may also be called measurement list information.
  • the measurement configuration information may include identification information of one or more AI nodes to be measured.
  • the identification information of the AI node to be measured may be the identification number (identity document, ID) of the AI node to be measured, that is, the AI node ID (AI node ID) of the AI node to be measured.
  • the AI nodes to be measured that need to be measured may be AI nodes within a certain area determined by the first AI node based on actual needs or business scenarios.
  • the identification information of the AI service that needs to be measured is used by the terminal device to determine the AI service that needs to be measured. In other words, it is used by the terminal device to determine the measurement target that needs to be measured.
  • the identification information of the AI service that needs to be measured can also be called measurement. target information.
  • the measurement configuration information may include the identification information of one or more AI services that need to be measured.
  • the identification information of the AI service may be the ID of the AI service, that is, the AI service ID (AI service ID).
  • AI service ID AI service ID
  • the AI services that need to be measured may be determined by the first AI node based on actual needs or business scenarios.
  • the first condition in the measurement configuration information, may be in a corresponding relationship with the identification information of the AI service that needs to be measured.
  • different AI services can correspond to different reporting conditions.
  • the first condition corresponding to the AI service with AI service ID 1 can be that the performance is higher than the first threshold
  • the corresponding condition for the AI service with AI service ID 2 The first condition may be performance above a second threshold.
  • different AI services may also correspond to the same reporting conditions, and the measurement configuration information may only include a first condition. This embodiment of the present application does not specifically limit this.
  • the corresponding relationship between the identification information of the AI node and the AI node, and the corresponding relationship between the identification information of the AI service and the AI service are protocols or configurations between the terminal device and the AI node.
  • the terminal device and the AI node can determine the same AI node based on the identification information of the same AI node, or the same AI service based on the identification information of the same AI service.
  • the terminal device After receiving the measurement configuration information, the terminal device can determine the AI node to be measured that needs to be measured, the AI service that needs to be measured, and the conditions that need to be met to report the measurement results based on the information included in the measurement configuration information. After the terminal device measures the AI services that need to be measured on the AI node to be measured according to the instructions of the measurement configuration information, the terminal device can obtain one or more AI service information for each AI node to be measured, that is, the AI service information of the AI node to be measured. Measurement results. Then, the terminal device determines the AI service information that satisfies the first condition as the candidate AI service information among the obtained candidate AI service information.
  • the terminal device can measure the AI node to be measured through the network device (such as a RAN device) corresponding to the AI node to be measured.
  • the network device such as a RAN device
  • the AI service information obtained by the terminal device measuring the node to be measured may include at least one of the following information:
  • AI service identification information AI node identification information, model identification information, model parameter information, performance information and timeliness information.
  • the AI service identification information is used to identify the AI service corresponding to the AI service information.
  • the terminal device can indicate the type of measured AI service through the AI service identification information.
  • the AI node identification information is used to identify the AI node corresponding to the AI service information.
  • the terminal device can indicate the measured AI node through the AI node identification information.
  • the model identification information is used to identify the neural network model corresponding to the AI service information.
  • one or more neural network models can be configured on the AI node to implement AI services, and the terminal device can indicate the type of measured neural network model through the model identification information.
  • the model parameter information is the parameter information of the neural network model corresponding to the AI service information.
  • the model parameter information may include one or more parameter information, such as the size, input format or output format of the neural network model. In the embodiment of the present application, There are no specific restrictions on this.
  • Performance information is used to characterize the performance of the AI service corresponding to the AI service information.
  • the performance information may include one or more dimensional information, such as accuracy, inference accuracy, or model training convergence time. The embodiment of the present application does not specifically limit this.
  • the performance of the AI service is explained below with a specific example. Assume that the AI service is image classification and that the AI service can correctly classify 99 out of 100 images. The accuracy of the AI service is 99%.
  • the performance information can be characterized by values, intervals, or levels specified by the protocol. For example, when the performance information includes accuracy, the corresponding level is A when the accuracy is 90%-95%. If the terminal device measures the AI service If the obtained accuracy is 90%, then the accuracy included in the performance information in the AI service information can be a numerical value of 90%, or level A.
  • the aging information is used to represent the validity time of the AI service information.
  • the AI service information is valid within the valid time and becomes invalid after the valid time.
  • the aging information can be represented by time, such as time 0 (time0, T0) ⁇ time 1 (time1, T1), or it can also be represented by duration, such as 5 minutes (minute, min).
  • AI service information includes AI service identification information, AI node identification information, model parameter information, performance information and timeliness information, where the model parameter information includes size, input format and output format parameter information, and the performance information includes accuracy information
  • model parameter information includes size, input format and output format parameter information
  • performance information includes accuracy information
  • AI service ID is the AI service identification information, and its corresponding service 1 indicates that the AI service information is obtained by measuring AI service: service 1.
  • AI node ID is the AI node identification information, and its corresponding 001 indicates that the AI service information is obtained by measuring AI node 001.
  • Model information is model parameter information, which includes size and size corresponding to 3 and 10, indicating that the neural network model on the AI node 001 has 3 layers, each layer has 10 neurons, input format and the integer vector corresponding to input format Indicates that the input format of the neural network model on the AI node 001 is an integer vector.
  • output format and the integer vector corresponding to the output format indicate that the input format of the neural network model on the AI node 001 is an integer vector.
  • Performance is performance information, which includes accuracy and accuracy corresponding to 99%, indicating that the accuracy of AI service service 1 is 99%.
  • Validity is aging information, and its corresponding T0 ⁇ T1 indicates that the validity time of the AI service information lasts from T0 to T1.
  • the candidate AI service information is the AI service information that satisfies the first condition among the AI service information obtained by the terminal device measuring the node to be measured, therefore if the introduction of the information included above is replaced with For candidate AI service information, if the AI node is replaced by the AI node to be measured, the above introduction to the AI service information can also be applied to the candidate AI service information.
  • the terminal device determines the candidate AI service information, it sends the candidate AI service information to the first AI node through a measurement report, so that the first AI node can obtain AI service information that satisfies the first condition based on the received measurement report.
  • AI nodes can obtain AI service information that meets certain conditions of other AI nodes through interaction with terminal devices, which facilitates collaborative management among different AI nodes in the AI network.
  • the terminal device can send a measurement report to a connected network device, and the network device connected to the terminal device sends a measurement report to the first AI node.
  • the measurement report contains one or more AIs to be measured.
  • Candidate AI service information corresponding to the node Among them, assuming that the network device corresponding to the first AI node is the first network device, if the network device connected to the terminal device is not the first network device, the network device connected to the terminal device can send the measurement report to the first network device through forwarding, Thereby sending the measurement report to the first AI node.
  • the terminal device may send a measurement report corresponding to the AI node to be measured to the AI node to be measured through a connected network device, so that the AI node to be measured sends its respective measurement report to the first AI node.
  • the measurement report corresponding to the AI node to be measured includes candidate AI service information corresponding to the AI node to be measured, but does not include candidate AI service information corresponding to other AI nodes to be measured.
  • the terminal device may include The measurement report of the candidate AI service information corresponding to the second AI node is sent to the network device connected to the terminal device, so that the network device connected to the terminal device sends the measurement report corresponding to the second AI node to the second network device.
  • the network device connected to the terminal device may send the measurement report to the second network device in a forwarding manner.
  • the terminal device sends the measurement report corresponding to the AI node to be measured to the AI node to be measured through the connected network device, so that the AI node to be measured sends its respective measurement report to the first AI node.
  • the measurement report also includes the identification information of the first AI node.
  • the network device corresponding to the AI node to be measured (for example, the second network device) can send the measurement report to the first AI node based on the identification information of the first AI node.
  • the terminal device sends the measurement report corresponding to the AI node to be measured to the AI node to be measured through the connected network device, so that the AI node to be measured sends its respective measurement report to the first AI node, and the forwarding of the measurement report
  • This can be accomplished by interfaces between network devices deployed independently by AI nodes, or by interfaces between network devices (such as RAN devices) that are connected to each other and can provide access network functions for terminal devices.
  • the first AI node can determine the target AI service information based on the candidate AI service information, thereby updating the local AI service list of the first AI node.
  • the following is an introduction to how the first AI node updates the local AI service list based on candidate AI service information.
  • the first AI node can determine the candidate AI service information with the best performance of the AI service as the target AI service information among one or more candidate AI service information corresponding to the same type of AI service. And add the target AI service information to the local AI service list.
  • the type of AI service can be determined by the AI service identification information included in the candidate AI service information, and the AI service types corresponding to the same AI service identification information are the same.
  • the performance of the AI service can be determined through the performance information included in the candidate AI service information.
  • the first AI node can, according to the AI service identification information and performance information included in the candidate AI service information, respectively combine one or more AI services corresponding to the candidate AI service information into the candidate AI service information corresponding to each AI service.
  • the target AI service information is added to the local AI service list, where the target AI service information is the candidate AI service information with the best performance among the candidate AI service information corresponding to each AI service.
  • the first AI node may determine the candidate AI service information with the best performance according to preset rules.
  • the preset rule may be that if the performance information is represented by a numerical value, the candidate AI service information with the largest performance information value is the candidate AI service information with the best performance.
  • the first AI node can determine the target AI service information regularly. For example, every 10 minutes, the first AI node determines the target AI service among all candidate AI service information received within these 10 minutes. information. Alternatively, the first AI node can determine the target AI service information when the number of received candidate AI service information meets a preset threshold. For example, the first AI node can determine the target AI service information every time it receives 10 pieces of candidate AI service information. The target AI service information is determined in the service information.
  • the first AI node may determine the candidate AI service information that is different from the AI service information recorded in the local AI service list as the target AI service information, and set the target AI service information as the target AI service information. The information is added to the local AI service list. In other words, if the first AI node determines that the candidate AI service information has been recorded in the local AI service list, the first AI node ends the process and does not execute subsequent steps. step, otherwise the first AI node adds, or updates, the candidate AI service information to the local AI service list.
  • the candidate AI service information that is different from the AI service information recorded in the local AI service list refers to the candidate AI service information that is of a different type from the AI service information recorded in the local AI service list.
  • the type of AI service information can be characterized by AI service identification information.
  • the first AI node can determine whether the type of the candidate AI service information is consistent with the AI recorded in the local AI service list based on the AI service identification information included in the AI service information.
  • the types of service information are different, which determines whether to add candidate AI service information to the local AI service list.
  • the two implementation methods of updating the local AI service list based on candidate AI service information introduced above can be applied independently or in combination. It can be understood that when these two implementation methods are combined and applied, the AI service corresponding to the AI service information retained in the local AI service list is the AI service with the best performance among the same type of AI services.
  • the first AI node can be independently deployed on a network device dedicated to realizing the function of the AI node.
  • the network device deployed by the first AI node corresponds to the RAN device, and the RAN device can communicate with the terminal device.
  • a possible schematic diagram in which the first AI node instructs the terminal device to measure the AI node to be measured through the RAN device and updates the local AI service list according to the measurement report reported by the terminal device is as shown in Figure 4.
  • the network device deployed by the first AI node has two modules: model management and model measurement, and maintains a local AI service list.
  • the model measurement module sends measurement configuration information to the terminal device through the radio resource control (RRC) connection with the RAN device, instructs the terminal device to measure the AI node to be measured, and receives the measurement results reported by the terminal device. (Candidate AI service information).
  • the model management module can update the local AI service list through the measurement results collected by the model measurement module.
  • the specific implementation of the model measurement module instructing the terminal device to measure the AI node to be measured and receiving the measurement results reported by the terminal device can refer to the introduction to S301 above.
  • the specific implementation of the model management module updating the local AI service list can refer to the above description.
  • the first AI node after the first AI node updates the local AI service list based on the candidate AI service information, it can also obtain it from the corresponding AI node based on one or more AI service information recorded in the local AI service list.
  • Neural network model to implement AI services to implement AI services.
  • the candidate AI service information includes AI node identification information. It can be understood that, therefore, the AI service information recorded in the local AI service list also includes AI node identification information.
  • the first AI node can be based on one of the local AI service lists. or the AI node identification information included in multiple AI service information, determine the AI node corresponding to one or more AI service information (hereinafter referred to as the target AI node), and obtain the neural network model from the target AI node.
  • the first AI node can regularly obtain the corresponding neural network model from the corresponding target AI node. For example, every 10 minutes, the first AI node can obtain the corresponding neural network model based on the local AI service list within these 10 minutes.
  • the added AI service information determines the target AI node and obtains the corresponding neural network model from the target AI node.
  • the first AI node can obtain the corresponding neural network model from the corresponding target AI node when the number of newly added AI service information in the local AI service list meets the preset threshold. For example, each new AI service information in the local AI service list Add 10 pieces of AI service information, and the first AI node determines the target AI node based on the 10 newly added pieces of AI service information and obtains the neural network model from the target AI node.
  • the first AI node can obtain the target from the target through the following two different implementation methods: Obtain the neural network model at the AI node.
  • the candidate AI service information includes model identification information. It is understandable that, therefore, the AI service information recorded in the local AI service list also includes model identification information. After the first AI node determines the target AI node corresponding to the AI service information in the local AI service list, it can obtain the corresponding model identification information from the target AI node based on the model identification information included in the AI service information in the local AI service list. neural network model.
  • the first AI node can obtain the neural network model preconfigured by the target AI node from the target AI node.
  • one or more neural network models can be pre-configured on the target AI node.
  • the first AI node can request the neural network model from the target AI node by sending a request message to the target AI node, and the target AI node can send a response message to the first AI node. , sending the neural network model to the first AI node.
  • the embodiments of the present application provide that the first AI node sends measurement configuration information to the terminal device, receives the measurement report from the terminal device, and obtains the neural network from the target AI node according to the measurement report.
  • the model scheme is introduced.
  • Example 1 Assume that one of the AI nodes to be measured is a second AI node.
  • the AI node is independently deployed on a network device dedicated to realizing the AI node function.
  • the first AI node corresponds to RAN device 1
  • the second AI node The node corresponds to the RAN device 2, and the first AI node and the second AI node can communicate through respectively deployed network devices.
  • a possible solution for the first AI node to obtain the neural network model from the second AI node includes S501-S504:
  • S501 The first AI node sends measurement configuration information to the terminal device through the RAN device 1.
  • S501 please refer to the introduction of S301 above.
  • S502 The terminal device measures the second AI node through the RAN device 2, and obtains the measurement result (AI service information corresponding to the second AI node).
  • AI service information corresponding to the second AI node.
  • S503 The terminal device determines that the AI service information corresponding to the second AI node meets the first condition, carries the AI service information corresponding to the second AI node in the measurement report, and sends it to the first AI node through RAN device 1.
  • the specific implementation of S503 please refer to the introduction of S302 above.
  • S504 The terminal device adds the AI service information corresponding to the second AI node to the local AI service list, and obtains the corresponding neural network model from the second AI node.
  • the terminal device adds the AI service information corresponding to the second AI node to the local AI service list, and obtains the corresponding neural network model from the second AI node.
  • Example 2 Assume that one of the AI nodes to be measured is a second AI node.
  • the AI node is independently deployed on a network device dedicated to realizing the AI node function.
  • the first AI node corresponds to RAN device 1
  • the second AI node The node corresponds to the RAN device 2, and the first AI node and the second AI node can communicate through respectively deployed network devices.
  • a possible solution for the first AI node to obtain the neural network model from the second AI node includes S601-S604:
  • S601 The first AI node sends measurement configuration information to the terminal device through the RAN device 1.
  • S601 please refer to the introduction to S301 above.
  • S602 The terminal device measures the second AI node and obtains the measurement result (AI service information corresponding to the second AI node).
  • AI service information corresponding to the second AI node.
  • S603 The terminal device determines that the AI service information corresponding to the second AI node meets the first condition, carries the AI service information corresponding to the second AI node in the measurement report, and sends it to the first AI node through RAN device 2.
  • the specific implementation of S603 please refer to the introduction of S302 above.
  • the first AI node adds the AI service information corresponding to the second AI node to the local AI service list, and obtains the corresponding neural network model from the second AI node.
  • the specific implementation of S604 please refer to the above introduction of the first AI node obtaining the neural network model from the target AI node.
  • the AI service information acquisition method includes S701-S703:
  • the terminal device sends subscription information to the third AI node, and correspondingly, the third AI node receives the subscription information from the terminal device.
  • the subscription information is used to request to subscribe to the AI service that meets the second condition in the subscription area.
  • the third AI node sends subscription information to one or more fourth AI nodes in the subscription area.
  • One or more fourth AI nodes in the subscription area send subscription results to the terminal device.
  • the terminal device receives the subscription results from one or more fourth AI nodes.
  • the subscription result includes target AI service information, and the AI service corresponding to the target AI service information satisfies the second condition.
  • the terminal device requests to subscribe to the AI service that the terminal device expects to obtain by sending subscription information to the third AI node, that is, requesting to subscribe to the AI service that meets the second condition.
  • the subscription information includes at least one of the following information:
  • AI service identification information user identification information, service level information, subscription area information, model identification information, model parameter information and subscription validity period information.
  • the AI service identification information in the subscription information is used to identify the AI service that the terminal device requests to subscribe to.
  • the terminal device can indicate the type of AI service it expects to obtain through the AI service identification information in the subscription information.
  • the AI service that the terminal device requests to subscribe to may be determined by the terminal device based on actual needs or business scenarios.
  • the subscription information may include one or more AI service identification information.
  • the identification information of the AI service may be the ID of the AI service, that is, the AI service ID (AI service ID).
  • the user identification information in the subscription information is used to identify the terminal device that sends the subscription information.
  • the user identification information may also be called the subscription initiator identification.
  • the user identification information can reuse existing information used to identify the terminal device.
  • the user identification information can be a temporary mobile subscriber identity (TMSI).
  • TMSI temporary mobile subscriber identity
  • the user identification information may be newly defined identification information used to identify the terminal device in the AI network.
  • the service level information in the subscription information is used to characterize the second condition, in other words, it is used to characterize the conditions that need to be met for the AI service that the terminal device expects to obtain.
  • the subscription information may include one or more service level information.
  • the service level information may be a threshold that the performance of the AI service requested to be subscribed to needs to meet. If the performance of the AI service of the fourth AI node meets the threshold, the fourth AI node may feed back the subscription result corresponding to the subscription information. To the terminal device, otherwise the subscription result will not be fed back.
  • different AI services can correspond to different second conditions.
  • the service level information can be corresponding to the identification information of the AI service.
  • the service level information corresponding to the AI service with an AI service ID of 1 may indicate that the performance is higher than the first threshold
  • the service level information corresponding to the AI service with the AI service ID of 2 may indicate that the performance is higher than the second threshold.
  • different AI services may also correspond to the same second condition, and the subscription information may only include the same service level information. This embodiment of the present application does not specifically limit this.
  • the subscription area information in the subscription information is used to characterize the subscription area.
  • the subscription area information may include the identification information of the subscribed AI node, and the third AI node sends the subscription information to the corresponding AI node based on the identification information of the AI node in the subscription information.
  • each AI node corresponding to the identification information of each AI node in the subscription information can also be referred to as each AI node in the subscription area.
  • the subscription information may include identification information of a pre-planned subscription area, and the third AI node sends subscription information to all AI nodes in the corresponding subscription area based on the identification information of the subscription area.
  • the subscription area information may be determined by the terminal device or the third AI node according to actual needs or business scenarios.
  • the terminal device can send subscription information including the subscription area information to the third AI node, and the third AI node sends the received subscription information to the AI node in the subscription area.
  • the third AI node determines the subscription area information, after receiving the subscription information from the terminal device that does not include the subscription area information, the third AI node determines the subscription area information and carries it in the subscription information, and then adds the subscription area information carrying the subscription area information. Subscription information is sent to AI nodes in the subscription area.
  • the model identification information in the subscription information is used to identify the neural network model corresponding to the AI service that the subscription information requests to subscribe to.
  • the terminal device may indicate the type of neural network model used to implement the AI service requested to be subscribed through the model identification information.
  • the identification information of the AI service in the subscription information may be in a corresponding relationship with the model identification information, and the identification information of the same AI service may correspond to one or more model identification information.
  • the model parameter information in the subscription information is the parameter information of the neural network model corresponding to the AI service subscribed by the subscription information request, and is used to indicate the parameter information of the neural network model expected by the terminal device.
  • the model parameter information in the subscription information may include one or more parameter information, such as the size, input format or output format of the neural network model.
  • the embodiment of the present application does not specifically limit this. For example, assuming that the terminal device requests to subscribe to the AI service and the terminal device expects the neural network model used to implement the AI service to have 100 neurons, the size information included in the subscription information may be 100 in the model parameter information.
  • the model identification information in the subscription information may be in a corresponding relationship with the model parameter information.
  • the parameter information expected by the terminal device may also be different.
  • the model parameter information includes size information
  • the terminal device expects the neural network model 1 used to implement AI services to have 100 neurons
  • the terminal device expects the neural network model 2 used to implement AI services to have 200 neurons
  • the subscription information , the size information corresponding to model identification information 1 may be 100
  • the size information corresponding to model identification information 2 may be 200.
  • different neural network models model identification information
  • the subscription validity period information in the subscription information is used to represent the validity time of the subscription information.
  • the subscription information is valid within the validity time and becomes invalid after the validity time.
  • the subscription validity period information can be represented by time, such as time 0 (time0, T0) ⁇ time 1 (time1, T1), or it can also be represented by duration, such as 5 minutes.
  • the subscription information may be identified by the first identification information.
  • the first identification information may be newly defined identification information specifically used to identify subscription information.
  • a possible form of the first identification information may be a subscription identification number (subscribe-ID, S-ID).
  • the first identification information may be generated by a third AI node.
  • the third AI node may generate the first identification information after receiving the subscription information from the terminal device, and The first identification information is carried in a response message and sent to the terminal device, where the response message is used to respond to the subscription information sent by the terminal device.
  • the first identification information may be generated by the terminal device.
  • the terminal device sends the subscription information identified by the first identification information to the third AI node.
  • the third AI node sends the subscription information to each AI node in the subscription area.
  • the AI node that receives the subscription information can determine whether it can provide the AI service requested by the terminal device based on the subscription information. If it can provide the AI service requested by the terminal device, it will monitor whether the corresponding AI service it provides meets the second requirement. condition.
  • an AI node that can provide an AI service that satisfies the second condition and is requested to be subscribed is called a fourth AI node.
  • the third AI node may send subscription information to one or more fourth AI nodes within the subscription area.
  • the third AI node may be an AI node within the subscription area.
  • the third AI node may also monitor whether it can provide the corresponding AI service that satisfies the second condition based on the subscription information.
  • the third node is also a fourth AI node.
  • the third AI node may not be an AI node within the subscription area, and the third AI node does not need to monitor whether it can provide corresponding AI services that meet the second condition.
  • the fourth AI node in the subscription area determines that the AI service provided by itself can satisfy the second condition, it carries the AI service information corresponding to the AI service that satisfies the second condition in the subscription result and sends it to the terminal device, so that the terminal The device obtains AI service information corresponding to the desired AI service.
  • the AI service information corresponding to the AI service that meets the second condition may also be called target AI service information.
  • the terminal device can subscribe to the desired AI service from the AI node and obtain the AI service information corresponding to the AI service that meets the conditions, so that the terminal device can subscribe to the AI service in the AI network.
  • AI services that meet the requirements.
  • the AI service information corresponding to the AI service that satisfies the second condition refer to the above introduction to the AI service information in S301.
  • the fourth AI node sends the subscription result to the terminal device in the following two ways:
  • Implementation method 1 The fourth AI node sends the subscription result to the third AI node, and the third AI node sends the subscription result to the terminal device.
  • Implementation method two The fourth AI node directly sends the subscription result to the terminal device. It should be noted that "direct” in this implementation means that there is no need to go through the third AI node, and it does not mean forwarding without the need for an intermediate device.
  • the way in which the third AI node or the fourth AI node sends the subscription result to the terminal device is related to the status of the terminal device, and will be introduced below.
  • the third AI node or the fourth AI node can send the subscription result to the terminal device through the RAN device currently connected to the terminal device.
  • the RAN device currently accessed by the terminal device may send the subscription result through a terminal-specific parameter signal (UE-specific) message.
  • UE-specific terminal-specific parameter signal
  • the third AI node or the fourth AI node sending the subscription result to the RAN device currently accessed by the terminal device please refer to the above introduction to the specific implementation of the terminal device sending the measurement report to the first AI node.
  • the fourth AI node sends the subscription result to the terminal device in the same manner as the method of moving the terminal device provided by the embodiments of the present application that will be introduced below. It is related to the sexual management plan and will not be introduced here.
  • the AI service acquisition method provided by the embodiment of this application may also include S704.a or S704.b:
  • the terminal device After the terminal device obtains the subscription result corresponding to the subscription information for the first time, it notifies the currently resident AI node that the terminal device has obtained the subscription result, and the currently resident AI node notifies other AI nodes in the subscription area that it can be terminated. Subscribe to tasks.
  • the AI node where the terminal device currently resides refers to the AI node corresponding to the network device (such as RAN device) currently connected to the terminal device. Terminating the subscription task means that the AI node can stop monitoring whether it can provide AI services that meet the subscription information requirements (satisfying the second condition).
  • S704.b In the subscription area, the fourth AI node that is the first in time to determine that it can provide the AI service required by the subscription information will notify other AI nodes in the subscription area while sending the subscription result to the terminal device. Subscription tasks can be terminated.
  • the S701-S703 is introduced above.
  • the S701-S703 solution allows terminal devices to subscribe to AI services that meet the requirements in the AI network.
  • the current AI network still has the following problems.
  • TA tracking area
  • Multiple cells are defined as a tracking area for mobility management on the wireless access network side.
  • the terminal device has been moving within a tracking area, there is no need to inform the wireless access network.
  • the terminal device in the idle state is paged, paging is performed in all cells in the tracking area where the terminal device is located.
  • the wireless access network and the AI network are logically separated, and there is currently no mechanism to manage the mobility of terminal devices on the AI network side. Therefore, embodiments of the present application also provide a solution that can perform mobility management on terminal devices on the AI network side. This solution will be described below.
  • the method for obtaining AI service information also includes S702.1-S702.3:
  • the third AI node determines that the terminal device is currently located in the first notification area, and associates the first identification information with the second identification information; the second identification information is used to identify the first notification area.
  • the third AI node sends the second identification information associated with the first identification information to one or more fourth AI nodes.
  • the terminal device obtains the second identification information associated with the first identification information.
  • the third AI node may determine that the terminal device is currently located in the first notification area according to the location information of the terminal device, thereby determining the second identification information used to identify the first notification area, and establishing the relationship between the first identification information and the third notification area.
  • the correlation relationship between the two identification information is to obtain the second identification information associated with the first identification information.
  • the notification area (notification-area, N-Area) may be a newly defined concept used to track terminal devices in the AI network and perform mobility management. Specifically, how to manage the mobility of terminal devices based on the notification area is introduced below.
  • the notification area (such as the first notification area or the second notification area introduced below) Notification area) is set or planned based on cells and/or beams.
  • the notification area can be planned in units of cells, and three adjacent cells constitute one notification area.
  • the notification area can be planned in units of beams, and three beams constitute one notification area.
  • the identification information used to identify the notification area may also be newly defined information.
  • one possible form of identification information used to identify the notification area may be a notification area identification number (notification-area-ID, N-ID).
  • the third AI node sends the second identification information and the association between the first identification information and the second identification information to each AI node in the subscription area. It can be understood that one or more fourth AI nodes in the subscription area can receive the second identification information, as well as the association between the first identification information and the second identification information, and based on the first identification information associated with the first identification information. The second identification information determines that the terminal device is currently located in the first notification area.
  • the third AI node may simultaneously send the subscription information identified by the first identification information, the second identification information, and the association between the first identification information and the second identification information to one or more fourth AI nodes. .
  • the third AI node may first send the subscription information identified by the first identification information to one or more fourth AI nodes, and then send the second identification information and the association between the first identification information and the second identification information,
  • the embodiments of this application are not specifically limited.
  • the terminal device can obtain the second identification information associated with the first identification information in the following manner:
  • the third AI node sends the second identification information associated with the first identification information to the terminal device.
  • the third AI node may also send the association relationship between the first identification information and the second identification information to the terminal device.
  • the third AI node may send the second identification information to the network device in the first notification area, thereby sending the second identification information to the terminal device.
  • the third AI node may carry the second identification information in a response message sent to the terminal device, where the response message is used to respond to the subscription information sent by the terminal device.
  • Implementation manner two the terminal device receives the second identification information sent from the network device in the first notification area.
  • network devices in the notification area such as RAN devices, can broadcast identification information used to identify the notification area, so that after the terminal device moves to the notification area, it can receive identification information used to identify the notification area. Identification information for this notification area. Therefore, in this implementation, when the terminal device is located in the first notification area, it can receive the second identification information broadcast by the RAN device in the first notification area.
  • this embodiment of the present application does not limit the timing of S702.3 to be after S702.1 and S702.2.
  • the terminal device can obtain the second identification information after moving to the first notification area, and it is not related to the third AI node.
  • the first identification information is not associated with the second identification information.
  • the following describes how to manage the mobility of terminal devices in the AI network through the notification area.
  • the AI service information acquisition method also includes S702.4-S702.5:
  • the terminal device obtains the third identification information, and the third identification information is used to identify the second notification area where the terminal device is currently located.
  • the terminal device may receive the third identification information from the third network device, where the third network device The device is located in the second notification area.
  • the RAN device in the second notification area broadcasts third identification information used to identify the second notification area. After the terminal device moves to the second notification area, it can receive the third identification information broadcast by the RAN device.
  • the terminal device determines that the third identification information is inconsistent with the second identification information, and sends an update request to one or more fourth AI nodes, where the update request is used to request that the first identification information be associated with the third identification information.
  • the terminal device determines that the obtained third identification information is inconsistent with the previously obtained second identification information, it means that the terminal device has moved out of the first notification area where it was previously located and moved to the second notification area. Therefore, the terminal device needs to send an update request to each AI node (including one or more fourth AI nodes) in the subscription area, requesting that the first identification information be associated with the third identification information.
  • the AI node in the subscription area updates the association between the first identification information and the second identification information according to the update request to the association between the first identification information and the third identification information. After the update, the first identification information associated with third identification information. Therefore, the AI node in the subscription area can learn that the terminal device has moved to the second notification area according to the third identification information associated with the first identification information.
  • the terminal device sends an S-ID to the third AI node (not shown in Figure 9) as
  • the subscription information of S1 includes notification area 1, notification area 2 and notification area 3.
  • the corresponding N-IDs of notification area 1, notification area 2 and notification area 3 are 001, 002 and 003 respectively.
  • Notification area 1, notification area 2 and notification area 3 correspond to one AI node 2.
  • the third AI node determines that the terminal device is currently located in notification area 1 (first notification area), associates S1 with the identification information 001 of notification area 1, and associates the associated information (including the association relationship ) and the subscription information are forwarded to AI node 2.
  • the terminal device When the terminal device is located in the first notification area, it receives N-ID: 001 broadcast by the RAN device (not shown in Figure 9) in notification area 1. Afterwards, the terminal device moves from notification area 1 to notification area 2. After the terminal device moves to notification area 2, it receives N-ID: 002 broadcast by the RAN device (not shown in Figure 9) in notification area 2.
  • the terminal device determines that 002 is inconsistent with 001, it sends an update request to the third AI section and AI node 2, requesting to associate S1 with 002. After receiving the update request, the third AI node and AI node 2 update 001 associated with S1 to 002. After the terminal device moves from notification area 2 to notification area 3, and determines that 003 and 002 are inconsistent based on the N-ID: 003 broadcast by the RAN device in notification area 3, it will send an update to the third AI node and AI node 2 again. request, so that the third AI node and AI node 2 update 002 associated with S1 to 003.
  • the fourth AI node may The notification area in which the terminal device is currently located is determined based on the identification information associated with the first identification information and used to identify the notification area, so that the subscription result is sent to the terminal device.
  • the fourth AI node may determine the notification area where the terminal device is currently located based on the identification information associated with the first identification information for identifying the notification area, thereby instructing the terminal All RAN devices in the notification area where the device is currently located page the terminal device through paging based on the user identification information or the first identification information, so that the terminal device reestablishes an air interface with a certain RAN device in the notification area where the device is currently located. link, so that the fourth AI node can send the subscription result to the terminal device through the RAN device that establishes an air interface link with the terminal device.
  • the fourth AI node can determine the notification area where the terminal device is currently located based on the identification information associated with the first identification information for identifying the notification area, so that the subscription can be more quickly
  • the result is sent to the RAN device currently accessed by the terminal device.
  • An exemplary AI service information acquisition method provided by embodiments of the present application is introduced below with reference to examples and figures. As shown in Figure 10, the method includes S1001-S1007.
  • the terminal device sends subscription information to AI node 1 (the third AI node) through the connected RAN device.
  • the subscription information may also be called a subscription request (subscribe request).
  • the subscription information request is to subscribe to AI node 2 to AI node N for AI service 1 that meets the second condition, where N is a positive integer greater than or equal to 2.
  • N is a positive integer greater than or equal to 2.
  • AI node 1 forwards the subscription information to AI node 2 to AI node N.
  • AI node 1 allocates S-ID: S1 (first identification information) to the subscription information, and associates S1 with 001 according to N-ID: 001 (second identification information) of notification area 1 where the terminal device is currently located.
  • AI node 1 sends the information 001, the association between S1 and 001, and the subscription information identified by S1 to AI nodes 2 to AI node N.
  • AI nodes 2 to AI nodes N monitor whether the AI services provided by themselves meet the second condition.
  • AI node 1 sends the response message of the subscription information to the terminal device. This response message may also be called a subscribe response.
  • AI node 1 carries the 001 information associated with S1 in the response message and sends it to the terminal device.
  • the terminal device receives a broadcast message from the connected RAN device.
  • the terminal device moves from notification area 1 to notification area 2 and is connected to the RAN device in notification area 2.
  • the RAN device in notification area 2 broadcasts the N-ID of notification area 2: 002 (third identification information). After the terminal device is connected to the RAN device in notification area 2, it receives 002 broadcast by the RAN device.
  • the terminal device sends an update request to AI node 1 and AI node 2 to AI node N.
  • This update request may also be called a notification area update (N-Area update) request.
  • the terminal device determines that 002 is inconsistent with 001 currently associated with S1, and then sends an update request to AI node 1 and AI node 2 to AI node N, requesting that S1 be associated with 002.
  • AI node 1 and AI node 2 to AI node N update the information associated with S1.
  • One or more AI nodes among AI nodes 2 to N determine that the AI service provided by itself can meet the second condition.
  • AI node 1 and AI node 2 to AI node N update 001 associated with S1 to 002. And during this period, one or more AI nodes among AI nodes 2 to AI nodes N determine that the AI services they provide can meet the second condition, that is, there are one or more fourth AI nodes among AI nodes 2 to AI nodes N. .
  • One or more fourth AI nodes among AI nodes 2 to AI nodes N determine that the terminal device is currently located in the notification area 2 based on 002 associated with S1, and send the subscription result to the RAN device in the notification area 2, thereby passing the notification
  • the RAN device in area 002 sends the subscription result to the terminal device.
  • S1007 please refer to the introduction of S703 above.
  • the methods and/or steps implemented by the first AI node/third AI node can also be implemented by the network device corresponding to the first AI node/third AI node, or by the network device that can be used to
  • the first AI node/third AI node is implemented by components (such as chips or circuits) of the network device corresponding to the first AI node/third AI node, and the methods and/or steps implemented by the terminal device can also be implemented by components that can be used in the terminal device.
  • embodiments of the present application also provide a communication device, which is used to implement the above various methods.
  • the communication device may be the first AI node/third AI node in the above method embodiment, or a network device corresponding to the first AI node/third AI node in the above method embodiment, or include the above first AI node/
  • the device of the network device corresponding to the third AI node may be a component that can be used in the network device corresponding to the first AI node/the third AI node; or the communication device may be the terminal device in the above method embodiment, or include the above A device for terminal equipment, or a component usable in terminal equipment.
  • the communication device includes corresponding hardware structures and/or software modules for performing each function.
  • the present application can be implemented in the form of hardware or a combination of hardware and computer software. Whether a function is performed by hardware or computer software driving the hardware depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each specific application, but such implementations should not be considered beyond the scope of this application.
  • Embodiments of the present application can divide the communication device into functional modules according to the above method embodiments.
  • each functional module can be divided corresponding to each function, or two or more functions can be integrated into one processing module.
  • the above integrated modules can be implemented in the form of hardware or software function modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. In actual implementation, there may be other division methods.
  • FIG. 11 shows a schematic structural diagram of a communication device 110.
  • the communication device 110 includes a sending module 1101 and a receiving module 1102.
  • the sending module 1101 and the receiving module 1102 may also be called transceiving modules.
  • the sending module 1101 and the receiving module 1102 may also be called a transceiver unit.
  • the transceiver unit is used to implement sending and/or receiving functions, and may be, for example, a transceiver circuit, a transceiver, a transceiver, or a communication interface.
  • the sending module 1101 can be used to perform the sending steps performed by the first AI node/third AI node in the above method embodiment
  • the receiving module 1102 can be used to perform the steps performed by the first AI node in the above method embodiment. /The steps of receiving the class performed by the third AI node.
  • the sending module 1101 is used to send measurement configuration information to the terminal device; the measurement configuration information is used to indicate measuring the AI service of the AI node to be measured; the receiving module 1102 is used to receive a measurement report from the terminal device; the measurement report includes candidate AI Service information; candidate AI service information is AI service information that satisfies the first condition among one or more AI service information obtained by measuring the AI node to be measured according to the measurement configuration information; the first condition is carried in the measurement configuration information.
  • the communication device 110 may also include a processing module 1103.
  • the processing module 1103 may be used to perform other steps in addition to receiving and sending steps performed by the first AI node/third AI node in the above method embodiment. step.
  • the communication device 110 is presented in the form of dividing various functional modules in an integrated manner.
  • a “module” here may refer to a specific ASIC, circuit, processor and memory that executes one or more software or firmware programs, integrated logic circuits, and/or other devices that may provide the above functions.
  • the communication device 110 may take the form of the communication device 200 shown in FIG. 2 .
  • the processor 201 in the communication device 200 shown in FIG. 2 can cause the communication device 200 to execute the AI service information acquisition method in the above method embodiment by calling the computer execution instructions stored in the memory 203.
  • the functions/implementation processes of the sending module 1101, the receiving module 1102 and the processing module 1103 in Figure 11 can be implemented by the processor 201 in the communication device 200 shown in Figure 2 calling the computer execution instructions stored in the memory 203.
  • the function/implementation process of the processing module 1103 in Figure 11 can be realized by the processor 201 in the communication device 200 shown in Figure 2 calling the computer execution instructions stored in the memory 203.
  • the sending module 1101 and the receiving module in Figure 11 The function/implementation process of module 1102 can be implemented through the communication interface 204 in the communication device 200 shown in FIG. 2 .
  • the communication device 110 provided in this embodiment can execute the above-mentioned AI service information acquisition method, the technical effects it can obtain can be referred to the above-mentioned method embodiments, which will not be described again here.
  • FIG. 12 shows a schematic structural diagram of a communication device 120.
  • the communication device 120 includes a sending module 1201 and a receiving module 1202.
  • the sending module 1201 and the receiving module 1202 can also be collectively called a transceiving module.
  • the sending module 1201 and the receiving module 1202 can also be called a transceiver unit.
  • the transceiver unit is used to implement sending and/or receiving functions, and can be, for example, a transceiver circuit, a transceiver, a transceiver, or a communication interface.
  • the sending module 1201 can be used to perform the sending steps performed by the terminal device in the above method embodiment
  • the receiving module 1202 can be used to perform the receiving step performed by the terminal device in the above method embodiment.
  • the receiving module 1202 is used to receive measurement configuration information from the first AI node; the measurement configuration information is used to indicate measuring the AI service of the AI node to be measured; the sending module 1201 is used to send a measurement report to the first AI node; measurement The report includes candidate AI service information; the candidate AI service information is the AI service information that satisfies the first condition among the one or more AI service information obtained by measuring the AI node to be measured according to the measurement configuration information; the first condition is carried in the measurement configuration information .
  • the sending module 1201 is used to send subscription information to the third AI node; the subscription information is used to request to subscribe to an AI service that meets the second condition in the subscription area; the receiving module 1202 is used to receive one or more messages from the subscription area.
  • the communication device 120 may also include a processing module 1203.
  • the processing module 1203 may be used to perform other steps other than the receiving and sending steps performed by the terminal device in the above method embodiment.
  • the communication device 120 is presented in the form of dividing various functional modules in an integrated manner.
  • a “module” here may refer to a specific ASIC, circuit, processor and memory that executes one or more software or firmware programs, integrated logic circuits, and/or other devices that may provide the above functions.
  • the communication device 120 may take the form of the communication device 200 shown in FIG. 2 .
  • the processor 201 in the communication device 200 shown in FIG. 2 can cause the communication device 200 to execute the AI service information acquisition method in the above method embodiment by calling the computer execution instructions stored in the memory 203.
  • the functions/implementation processes of the sending module 1201, the receiving module 1202 and the processing module 1203 in Figure 12 can be implemented by the processor 201 in the communication device 200 shown in Figure 2 calling the computer execution instructions stored in the memory 203.
  • the function/implementation process of the processing module 1203 in Figure 12 can be realized by the processor 201 in the communication device 200 shown in Figure 2 calling the computer execution instructions stored in the memory 203.
  • the sending module 1201 and the receiving module in Figure 12 The function/implementation process of module 1202 can be implemented through the communication interface 204 in the communication device 200 shown in FIG. 2 .
  • the communication device 120 provided in this embodiment can execute the above-mentioned AI service information acquisition method, the technical effects it can obtain can be referred to the above-mentioned method embodiments, which will not be described again here.
  • the embodiment of the present application also provides a communication device (for example, the communication device may be a chip or a chip system).
  • the communication device includes a processor and is used to implement the method in any of the above method embodiments.
  • the communication device further includes a memory.
  • the memory is used to store necessary program instructions and data.
  • the processor can call the program code stored in the memory to instruct the communication device to execute the method in any of the above method embodiments.
  • the memory may not be in the communication device.
  • the communication device further includes an interface circuit, which is a code/data reading and writing interface circuit.
  • the interface circuit is used to receive computer execution instructions (computer execution instructions are stored in the memory, and may be directly from memory, or possibly through other devices) and transferred to the processor.
  • the communication device is a chip system, it may be composed of a chip or may include a chip and other discrete devices, which is not specifically limited in the embodiments of the present application.
  • the embodiment of the present application also provides a communication device.
  • the communication device may include a processor and an interface circuit.
  • the interface circuit may be used to communicate with other modules other than the communication device.
  • the processor may be used to execute Computer programs or instructions to enable the communication device to implement the method in any of the above method embodiments.
  • the communication device may be a chip or a chip system.
  • the above embodiments it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof.
  • a software program it may be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions.
  • computer program instructions When computer program instructions are loaded and executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
  • the computer instructions may be stored in or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transferred from a website, computer, server, or data center Transmission to another website, computer, server or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) means.
  • the computer-readable storage medium can be any available medium that can be accessed by a computer or include one or more data storage devices such as servers and data centers that can be integrated with the medium.
  • the available media It may be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium (for example, a solid state disk (SSD)), etc.
  • the computer may include the aforementioned device.

Landscapes

  • Engineering & Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Databases & Information Systems (AREA)
  • Evolutionary Computation (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本申请涉及通信技术领域，提供一种AI服务信息获取方法、装置及系统，可以使网络中的终端设备和AI节点可以交互AI服务信息。该方法中，第一AI节点向终端设备发送测量配置信息；测量配置信息用于指示测量待测量AI节点的AI服务；第一AI节点接收来自终端设备的测量报告；测量报告包括候选AI服务信息；候选AI服务信息为根据测量配置信息对待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；第一条件携带在测量配置信息中。

Description

一种AI服务信息获取方法、装置及系统
本申请要求于2022年04月01日提交国家知识产权局、申请号为202210349458.2、发明名称为“一种AI服务信息获取方法、装置及系统”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及通信技术领域，尤其涉及一种AI服务信息获取方法、装置及系统。
背景技术
无线网络的智能化是一个重要的演进趋势。目前，第五代移动通信(5th generation mobile communication，5G)网络引入了人工智能(artificial intelligence，AI)能力，具备AI能力的网络中部署有AI节点，AI节点可以为终端设备提供AI服务，从而为用户带来更好的业务体验。
当网络中存在多个可选择的AI节点时，为了满足不同业务的需求，终端设备需要动态选择合适的AI节点，从而使选择的AI节点可以为终端设备提供满足需求的AI服务。目前，可以根据各个AI节点提供的AI服务的相关信息，即AI服务信息，选择合适的AI节点。因此，若要使网络中的终端设备可以动态选择AI节点，实现网络整体性能的优化，一种可行的方案是使网络中的终端设备和AI节点间可以交互AI服务信息，从而便于AI节点间的协同管理。但是，如何实现网络中的终端设备和AI节点可以交互AI服务信息，目前还没有解决方案。
发明内容
本申请实施例提供一种AI服务信息获取方法、装置及系统，用于解决如何实现网络中的终端设备和AI节点间交互AI服务信息的问题。
为达到上述目的，本申请的实施例采用如下技术方案：
第一方面，提供了一种AI服务信息获取方法，执行该方法的通信装置(主体)可以为第一AI节点对应的网络设备，也可以为第一AI节点对应的网络设备的部件，例如处理器、芯片、或芯片系统。下面以执行主体为第一AI节点为例描述，该方法包括：第一AI节点向终端设备发送测量配置信息；测量配置信息用于指示测量待测量AI节点的AI服务；第一AI节点接收来自终端设备的测量报告；测量报告包括候选AI服务信息；候选AI服务信息为根据测量配置信息对待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；第一条件携带在测量配置信息中。基于本申请实施例的技术方案，AI节点可以通过与终端设备的交互，获取其他AI节点的满足一定条件的AI服务信息，便于AI网络中不同AI节点间的协同管理。
结合上述第一方面，在一种可能的设计中，候选AI服务信息包括AI服务标识信息和性能信息，AI服务标识信息用于标识候选AI服务信息对应的AI服务，性能信息用于表征候选AI服务信息对应的AI服务的性能，方法还包括：第一AI节点根据AI服务标识信息和性能信息，分别将候选AI服务信息对应的一个或多个AI服务中，每个AI服务对应的候选AI服务信息中的目标AI服务信息添加至本地AI服务列表中，其中，目标AI服务信息为每个AI服务对应的候选AI服务信息中，AI服务的性能最 优的候选AI服务信息。基于本方案，可以将AI服务对应的一个或多个候选AI服务信息中，选择对应的性能最优的候选AI服务信息添加至本地AI服务列表中。
结合上述第一方面，在一种可能的设计中，方法还包括：第一AI节点将候选AI服务信息中的目标AI服务信息添加至本地AI服务列表中，其中，目标AI服务信息为候选AI服务信息中，与本地AI服务列表中记录的AI服务信息不同的候选AI服务信息。基于本方案，可以将与本地记录的AI服务信息不同的候选AI服务信息添加至本地AI服务列表中，保证本地AI服务列表中不会出现相同的候选AI服务信息。
结合上述第一方面，在一种可能的设计中，候选AI服务信息包括AI节点标识信息，AI节点标识信息用于标识候选AI服务信息对应的待测量AI节点，方法还包括：第一AI节点根据本地AI服务列表中的一个或多个AI服务信息包括的AI节点标识信息，确定与一个或多个AI服务信息对应的AI节点；第一AI节点获取一个或多个AI服务信息对应的AI节点的神经网络模型；神经网络模型用于实现AI服务。基于本方案，可以根据本地AI服务列表中记录的AI服务信息，从对应的AI节点处获取神经网络模型来实现AI服务。
结合上述第一方面，在一种可能的设计中，候选AI服务信息包括模型标识信息，模型标识信息用于标识候选AI服务信息对应的神经网络模型，第一AI节点获取一个或多个AI服务信息对应的AI节点的神经网络模型，包括：第一AI节点从一个或多个AI服务信息对应的AI节点上获取与一个或多个AI服务信息包括的模型标识信息对应的神经网络模型。基于本方案，提供了一种从AI节点处获取神经网络模型的方式。
结合上述第一方面，在一种可能的设计中，第一AI节点获取一个或多个AI服务信息对应的AI节点的神经网络模型，包括：第一AI节点从一个或多个AI服务信息对应的AI节点上获取预配置的神经网络模型。基于本方案，提供了一种从AI节点处获取神经网络模型的方式。
结合上述第一方面，在一种可能的设计中，测量配置信息包括待测量AI节点的标识信息和需要测量的AI服务的标识信息。
结合上述第一方面，在一种可能的设计中，候选AI服务信息包括以下至少一项信息：AI服务标识信息、AI节点标识信息、模型标识信息、模型参数信息、性能信息和时效信息；其中，AI服务标识信息用于标识候选AI服务信息对应的AI服务，AI节点标识信息用于标识候选AI服务信息对应的待测量AI节点，模型标识信息用于标识候选AI服务信息对应的神经网络模型，模型参数信息为候选AI服务信息对应的神经网络模型的参数信息，性能信息用于表征候选AI服务信息对应的AI服务的性能，时效信息用于表征候选AI服务信息的有效时间。
结合上述第一方面，在一种可能的设计中，第一AI节点接收来自终端设备的测量报告，包括：第一AI节点通过第二网络设备接收来自终端设备的测量报告，第二网络设备为与待测量AI节点中的第二AI节点对应的网络设备；其中，候选AI服务信息为与第二AI节点对应的候选AI服务信息。基于本方案，终端设备可以通过待测量AI节点，发送对应的测量报告给第一AI节点。
结合上述第一方面，在一种可能的设计中，测量报告还包括第一AI节点的标识信息；第一AI节点的标识信息用于第二网络设备将测量报告发送给第一AI节点。
第二方面，提供了一种AI服务信息获取方法，执行该方法的通信装置(主体)可以为终端设备，也可以为终端设备的部件，例如终端设备的处理器、芯片、或芯片系统。下面以执行主体为终端设备为例描述，该方法包括：终端设备接收来自第一AI节点的测量配置信息；测量配置信息用于指示测量待测量AI节点的AI服务；终端设备向第一AI节点发送测量报告；测量报告包括候选AI服务信息；候选AI服务信息为根据测量配置信息对待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；第一条件携带在测量配置信息中。基于本申请实施例的技术方案，AI节点可以通过与终端设备的交互，获取其他AI节点的满足一定条件的AI服务信息，便于AI网络中不同AI节点间的协同管理。
结合上述第二方面，在一种可能的设计中，测量配置信息包括待测量AI节点的标识信息和需要测量的AI服务的标识信息。
结合上述第二方面，在一种可能的设计中，候选AI服务信息包括以下至少一项信息：AI服务标识信息、AI节点标识信息、模型标识信息、模型参数信息、性能信息和时效信息；其中，AI服务标识信息用于标识候选AI服务信息对应的AI服务，AI节点标识信息用于标识候选AI服务信息对应的待测量AI节点，模型标识信息用于标识候选AI服务信息对应的神经网络模型，模型参数信息为候选AI服务信息对应的神经网络模型的参数信息，性能信息用于表征候选AI服务信息对应的AI服务的性能，时效信息用于表征候选AI服务信息的有效时间。
结合上述第二方面，在一种可能的设计中，终端设备向第一AI节点发送测量报告，包括：
终端设备通过第二网络设备向第一AI节点发送测量报告，第二网络设备为与待测量AI节点中的第二AI节点对应的网络设备；其中，候选AI服务信息为与第二AI节点对应的候选AI服务信息。
结合上述第二方面，在一种可能的设计中，测量报告还包括第一AI节点的标识信息；第一AI节点的标识信息用于第二网络设备将测量报告发送给第一AI节点。
第三方面，提供了一种AI服务信息获取方法，执行该方法的通信装置(主体)可以为第三AI节点对应的网络设备，也可以为第三AI节点对应的网络设备的部件，例如处理器、芯片、或芯片系统。下面以执行主体为第三AI节点为例描述，该方法包括：第三AI节点接收来自终端设备的订阅信息；订阅信息用于请求在订阅区域内订阅满足第二条件的AI服务；第三AI节点向订阅区域内的一个或多个第四AI节点发送订阅信息，以使得第四AI节点根据订阅信息，向终端设备发送订阅结果；订阅结果包括目标AI服务信息，目标AI服务信息对应的AI服务满足第二条件。基于本申请实施例提供的AI服务信息获取方法，终端设备可以向AI节点订阅期望的AI服务，并获取到满足条件的AI服务对应的AI服务信息，从而使终端设备可以在AI网络中订阅到满足要求的AI服务。
结合上述第三方面，在一种可能的设计中，订阅信息由第一标识信息标识。基于本方案，定义了一种用于标识订阅信息的第一标识信息。
结合上述第三方面，在一种可能的设计中，第一标识信息由第三AI节点生成；或者，第一标识信息由终端设备生成。基于本方案，提供了两种第一标识信息的生成方 式。
结合上述第三方面，在一种可能的设计中，在第三AI节点接收来自终端设备的订阅信息之后，方法还包括：第三AI节点确定终端设备当前位于第一通知区域，将第一标识信息与第二标识信息相关联；第二标识信息用于标识第一通知区域；第三AI节点将与第一标识信息相关联的第二标识信息发送给一个或多个第四AI节点。基于本方案，第三AI节点在确定终端设备位于第一通知区域后，可以将与第一标识信息关联的第二标识信息发送给第四AI节点，以使第四AI节点可以根据与第一标识信息关联的第二标识信息，获知发送了订阅信息的终端设备当前位于第一通知区域。
结合上述第三方面，在一种可能的设计中，第一通知区域是根据小区和/或波束设置的。基于本方案，提供了一种规划通知区域的方式。
结合上述第三方面，在一种可能的设计中，订阅信息包括以下至少一项信息：用于标识请求订阅的AI服务的AI服务标识信息，用于标识终端设备的用户标识信息，用于表征第二条件的服务等级信息，用于表征订阅区域的订阅区域信息，模型标识信息，模型参数信息和订阅有效期信息；其中，模型标识信息用于标识订阅信息请求订阅的AI服务对应的神经网络模型，模型参数信息为订阅信息请求订阅的AI服务对应的神经网络模型的参数信息，订阅有效期信息用于表征订阅信息的有效时间。
结合上述第三方面，在一种可能的设计中，订阅区域信息由终端设备确定；或者，订阅区域信息由所述第三AI节点确定。基于本方案，提供了两种确定订阅信息的方式。
第四方面，提供了一种AI服务信息方法，执行该方法的通信装置(主体)可以为终端设备，也可以为终端设备的部件，例如终端设备的处理器、芯片、或芯片系统。下面以执行主体为终端设备为例描述，该方法包括：终端设备向第三AI节点发送订阅信息；订阅信息用于请求在订阅区域内订阅满足第二条件的AI服务；终端设备接收来自订阅区域内的一个或多个第四AI节点的订阅结果；订阅结果包括目标AI服务信息，目标AI服务信息对应的AI服务满足第二条件。基于本申请实施例提供的AI服务信息获取方法，终端设备可以向AI节点订阅期望的AI服务，并获取到满足条件的AI服务对应的AI服务信息，从而使终端设备可以在AI网络中订阅到满足要求的AI服务。
结合上述第四方面，在一种可能的设计中，订阅信息由第一标识信息标识。基于本方案，定义了一种用于标识订阅信息的第一标识信息。
结合上述第四方面，在一种可能的设计中，第一标识信息由第三AI节点生成；或者，第一标识信息由终端设备生成。基于本方案，提供了两种第一标识信息的生成方式。
结合上述第四方面，在一种可能的设计中，在终端设备向第三AI节点发送订阅信息之后，方法还包括：终端设备获取与第一标识信息关联的第二标识信息；第二标识信息用于标识终端设备当前位于的第一通知区域。基于本方案，终端设备可以获取到用于标识当前位于的通知区域的第二标识信息。
结合上述第四方面，在一种可能的设计中，方法还包括：终端设备获取第三标识信息；第三标识信息用于标识终端设备当前位于的第二通知区域；终端设备确定第三标识信息与第二标识信息不一致，向一个或多个第四AI节点发送更新请求，更新请求 用于请求将第一标识信息与第三标识信息相关联。基于本方案，终端设备在确定当前位于的第二通知区域与发送订阅信息时位于的第一通知区域不一致后，可以向第四AI节点发起更新请求，以使第四AI节点可以将与第一标识信息关联的第二标识信息更新为第三标识信息，确定终端设备当前位于第二通知区域。
结合上述第四方面，在一种可能的设计中，终端设备获取第三标识信息包括：终端设备接收来自第三网络设备的第三标识信息；第三网络设备位于第二通知区域中。基于本方案，终端设备可以通过第二通知区域内的网络设备获取第三标识信息。
结合上述第四方面，在一种可能的设计中，第一通知区域是根据小区和/或波束设置的。基于本方案，提供了一种规划通知区域的方式。
结合上述第四方面，在一种可能的设计中，订阅信息包括以下至少一项信息：用于标识请求订阅的AI服务的AI服务标识信息，用于标识终端设备的用户标识信息，用于表征第二条件的服务等级信息，用于表征订阅区域的订阅区域信息，模型标识信息，模型参数信息和订阅有效期信息；其中，模型标识信息用于标识订阅信息请求订阅的AI服务对应的神经网络模型，模型参数信息为订阅信息请求订阅的AI服务对应的神经网络模型的参数信息，订阅有效期信息用于表征订阅信息的有效时间。
结合上述第四方面，在一种可能的设计中，订阅区域信息由终端设备确定；或者，订阅区域信息由第三AI节点确定。基于本方案，提供了两种确定订阅信息的方式。
第五方面，提供了一种通信装置用于实现上述各种方法。该通信装置可以为上述第一方面中的第一AI节点对应的网络设备，或者包含第一AI节点对应的网络设备的装置，或者第一AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第二方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置；或者，该通信装置可以为上述第三方面中的第三AI节点对应的网络设备，或者包含第三AI节点对应的网络设备的装置，或者第三AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第四方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置；所述通信装置包括实现上述方法相应的模块、单元、或手段(means)，该模块、单元、或means可以通过硬件实现，软件实现，或者通过硬件执行相应的软件实现。该硬件或软件包括一个或多个与上述功能相对应的模块或单元。
第六方面，提供了一种通信装置，包括：处理器和存储器；该存储器用于存储计算机指令，该处理器用于执行该存储器存储的指令，当该处理器执行该指令时，以使该通信装置执行上述任一方面所述的方法。该通信装置可以为上述第一方面中的第一AI节点对应的网络设备，或者包含第一AI节点对应的网络设备的装置，或者第一AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第二方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片；或者，该通信装置可以为上述第三方面中的第三AI节点对应的网络设备，或者包含第三AI节点对应的网络设备的装置，或者第三AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第四方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片。
第七方面，提供了一种通信装置，包括：处理器和接口电路，该接口电路用于与 该通信装置之外的模块通信；该处理器用于运行计算机程序或指令以执行上述任一方面所述的方法。该通信装置可以为上述第一方面中的第一AI节点对应的网络设备，或者包含第一AI节点对应的网络设备的装置，或者第一AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第二方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片；或者，该通信装置可以为上述第三方面中的第三AI节点对应的网络设备，或者包含第三AI节点对应的网络设备的装置，或者第三AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第四方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片。
或者，该接口电路可以为代码/数据读写接口电路，该接口电路用于接收计算机执行指令(计算机执行指令存储在存储器中，可能直接从存储器读取，或可能经过其他器件)并传输至该处理器，以使该处理器运行计算机执行指令以执行上述任一方面所述的方法。
在一些可能的设计中，该通信装置可以为芯片或芯片系统。
第八方面，提供了一种通信装置，包括：处理器；所述处理器用于与存储器耦合，并读取存储器中的指令之后，根据所述指令执行如上述任一方面所述的方法。该通信装置可以为上述第一方面中的第一AI节点对应的网络设备，或者包含第一AI节点对应的网络设备的装置，或者第一AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第二方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片；或者，该通信装置可以为上述第三方面中的第三AI节点对应的网络设备，或者包含第三AI节点对应的网络设备的装置，或者第三AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第四方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片。
第九方面，提供了一种计算机可读存储介质，该计算机可读存储介质中存储有指令，当其在通信装置上运行时，使得通信装置可以执行上述任一方面所述的方法。该通信装置可以为上述第一方面中的第一AI节点对应的网络设备，或者包含第一AI节点对应的网络设备的装置，或者第一AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第二方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片；或者，该通信装置可以为上述第三方面中的第三AI节点对应的网络设备，或者包含第三AI节点对应的网络设备的装置，或者第三AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第四方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片。
第十方面，提供了一种包含指令的计算机程序产品，当其在通信装置上运行时，使得通信装置可以执行上述任一方面所述的方法。该通信装置可以为上述第一方面中的第一AI节点对应的网络设备，或者包含第一AI节点对应的网络设备的装置，或者第一AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第二方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的 装置，比如芯片；或者，该通信装置可以为上述第三方面中的第三AI节点对应的网络设备，或者包含第三AI节点对应的网络设备的装置，或者第三AI节点对应的网络设备中包含的装置，比如芯片；或者，该通信装置可以为上述第四方面中的终端设备，或者包含上述终端设备的装置，或者上述终端设备中包含的装置，比如芯片。
第十一方面，提供了一种通信装置(例如，该通信装置可以是芯片或芯片系统)，该通信装置包括处理器，用于实现上述任一方面中所涉及的功能。在一种可能的设计中，该通信装置还包括存储器，该存储器，用于保存必要的程序指令和数据。该通信装置是芯片系统时，可以由芯片构成，也可以包含芯片和其他分立器件。
其中，第五方面至第十一方面中任一种设计方式所带来的技术效果可参见上述第一方面至第四方面中不同设计方式所带来的技术效果，此处不再赘述。
第十二方面，提供一种通信系统，该通信系统包括网络设备和终端设备。网络设备可以执行上述第一方面所述的方法，终端设备可以执行上述第二方面所述的方法；或者，网络设备可以执行上述第三方面所述的方法，终端设备可以执行上述第四方面所述的方法。
附图说明
图1a为本申请实施例提供的一种通信系统的架构示意图一；
图1b为本申请实施例提供的一种通信系统的架构示意图二；
图2为本申请实施例提供的一种通信装置的结构示意图；
图3为本申请实施例提供的一种AI服务信息获取方法的交互示意图；
图4为本申请实施例提供的一种第一AI节点与RAN设备通信的示意图；
图5为本申请实施例提供的一种获取神经网络模型的方法的交互示意图；
图6为本申请实施例提供的另一种获取神经网络模型的方法的交互示意图；
图7为本申请实施例提供的另一种AI服务信息获取方法的交互示意图；
图8为一种跟踪区示意图；
图9为本申请实施例提供的一种在AI网络中对终端设备进行移动性管理的方法的示意图；
图10为本申请实施例提供的一种示例性的AI服务信息获取方法的示意图；
图11为本申请实施例提供的另一种通信装置的结构示意图；
图12为本申请实施例提供的又一种通信装置的结构示意图。
具体实施方式
为了便于理解本申请实施例的技术方案，首先给出本申请涉及的相关技术的简要介绍如下。
无线网络的智能化是一个重要的演进趋势。目前，5G网络引入了AI能力，来为终端设备提供AI服务(AI service)。具体地，具备AI能力的网络中部署有AI节点(AI node)，网络可以通过AI节点上的神经网络模型，为终端设备提供例如数据分析、语音识别或者图像识别等AI服务。可选的，AI节点除了为终端设备提供AI服务，还可以支持收集数据、训练神经网络模型等功能。换言之，网络可以通过AI节点，使用基于AI的方法替换原有网络功能中基于数值公式的方法，从而提高网络资源使用效率，提升用户业务体验。
目前，AI节点可以独立部署在专用于实现AI节点功能的网络设备上，也可以作为功能模块部署在现有的网络设备上，例如可以部署在无线接入网(radio access network，RAN)设备上。可以理解的是，当AI节点作为(硬件或软件上)独立的逻辑单元时，AI节点可以构成逻辑上与RAN设备构成的接入网分离的AI网络。
在AI网络中，本地部署的AI节点只能获得该AI节点的本地观察范围内的数据，根据业务的不同，本地AI节点未必可以为终端设备提供最优的AI服务。可见，为了满足不同业务的需求，当AI网络中存在多个可选择的AI节点时，终端设备需要动态选择合适的AI节点，从而使选择的AI节点可以为终端设备提供满足需求的AI服务。目前，可以根据各个AI节点提供的AI服务的相关信息，即AI服务信息，选择合适的AI节点。因此，若要使网络中的终端设备可以动态选择AI节点，实现网络整体性能的优化，一种可行的方案是使AI网络中的终端设备和AI节点间可以交互AI服务信息，从而便于AI节点间的协同管理。但是，如何实现AI网络中的终端设备和AI节点可以交互AI服务信息，目前还没有解决方案。基于此，本申请实施例提供一种AI服务信息获取方法，用于实现终端设备和AI节点间交互AI服务信息。
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述。其中，在本申请的描述中，除非另有说明，“/”表示前后关联的对象是一种“或”的关系。例如，A/B可以表示A或B。本申请中的“和/或”仅仅是一种描述关联对象的关联关系，表示可以存在三种关系。例如，A和/或B，可以表示：单独存在A，同时存在A和B，单独存在B这三种情况。其中A，B可以为单数或者复数。并且，在本申请的描述中，除非另有说明，“多个”是指两个以上(包括两个)。“以下至少一项(个)”或其类似表达，是指的这些项中的任意组合，包括单项(个)或复数项(个)的任意组合。例如，a，b或c中的至少一项(个)，可以表示：a，b，c，a和b，a和c，b和c，或者a和b和c，其中a，b，c可以是单个，也可以是多个。另外，为了便于清楚描述本申请实施例的技术方案，在本申请的实施例中，采用了“第一”、“第二”等字样对功能和作用基本相同的相同项或相似项进行区分。本领域技术人员可以理解“第一”、“第二”等字样并不对数量和执行次序进行限定，并且“第一”、“第二”等字样也并不限定一定不同。同时，在本申请实施例中，“示例性的”或者“例如”等词用于表示作例子、例证或说明。本申请实施例中被描述为“示例性的”或者“例如”的任何实施例或设计方案不应被解释为比其它实施例或设计方案更优选或更具优势。确切而言，使用“示例性的”或者“例如”等词旨在以具体方式呈现相关概念，便于理解。
本申请实施例的技术方案可用于各种通信系统，该通信系统可以为第三代合作伙伴计划(3rd generation partnership project，3GPP)通信系统，例如，长期演进(long term evolution，LTE)系统，或者5G系统、车联网(vehicle to everything，V2X)系统，或者LTE和5G混合组网的系统，或者设备到设备(device-to-device，D2D)通信系统、机器到机器(machine to machine，M2M)通信系统、物联网(internet of things，IoT)，以及其他下一代通信系统。该通信系统也可以为非3GPP通信系统，不予限制。
本申请实施例的技术方案可以应用于各种通信场景，例如可以应用于以下通信场景中的一种或多种：增强移动宽带(enhanced mobile broadband，eMBB)、超可靠低时延通信(ultra-reliable low latency communication，URLLC)、机器类型通信(machine  type communication，MTC)、大规模机器类型通信(massive machine type communications，mMTC)、D2D、V2X、和IoT等通信场景。
其中，上述适用本申请的通信系统和通信场景仅是举例说明，适用本申请的通信系统和通信场景不限于此，本领域普通技术人员可知，随着网络架构的演变和新业务场景的出现，本申请实施例提供的技术方案对于类似的技术问题，同样适用。在此统一说明，以下不再赘述。
作为一种可能的实现，本申请提供一种本申请适用的通信系统。该通信系统可以包括一个或多个第一AI节点，以及一个或多个终端设备，或者，该通信系统可以包括一个或多个第三AI节点，以及一个或多个终端设备。
可选的，本申请实施例中，AI节点可以独立部署在专用于实现AI节点功能的网络设备上，该场景中，AI节点，或者说用于实现AI节点功能的网络设备与无线接入网设备呈对应关系，AI节点可以与对应的无线接入网设备通信，从而可以通过对应的无线接入网设备与终端设备通信。需要说明的是，AI节点与无线接入网设备可以为一一对应关系，也可以为一个AI节点对应多个无线接入网设备的一对多的对应关系。示例性的，假设AI节点与无线接入网设备为一一对应关系，本申请实施例提供的通信系统10可以如图1a所示，第一AI节点20或者第三AI节点20通过对应的无线接入网设备，与终端设备30通信。
可选的，本申请实施例中，AI节点可以作为功能模块部署在无线接入网设备上，AI节点部署的无线接入网设备也可以称为AI节点对应的无线接入网设备。AI节点可以通过对应的无线接入网设备提供的无线通信功能，与终端设备通信。示例性的，该场景下，本申请实施例提供的通信系统10可以如图1b所示，第一AI节点20或者第三AI节点20部署在无线接入网设备上，与终端设备30通信。
可选的，本申请提供的通信系统10包括第一AI节点20和终端设备30时，通信系统10还可以包括待测量AI节点。以图1a或图1b所示的第一AI节点20与任一终端设备30进行交互为例，在一种可能的实现方式中，第一AI节点20向终端设备30发送测量配置信息，测量配置信息用于指示测量待测量AI节点的AI服务。终端设备30接收到测量配置信息后，根据测量配置信息对待测量AI节点进行测量，得到一个或多个AI服务信息，并根据测量配置信息中携带的第一条件，在一个或多个AI服务信息中确定满足第一条件的候选AI服务信息。第一AI节点20确定候选AI服务信息后，通过测量报告将候选AI服务信息发送给第一AI节点20。该方案的具体实现以及技术效果将在后续方法实施例中详细描述，在此不予赘述。
可选的，本申请提供的通信系统10包括第三AI节点20和终端设备30时，本申请提供的通信系统10还可以包括第四AI节点。以图1a或图1b所示的第三AI节点20与任一终端设备30进行交互为例，在一种可能的实现方式中，终端设备30向第三AI节点20发送订阅信息，订阅信息用于请求在订阅区域内订阅满足第二条件的AI服务。第三AI节点20接收到订阅信息后，向订阅区域内的每个AI节点发送订阅信息，以使得订阅区域内的一个或多个第四AI节点根据订阅信息，向终端设备发送订阅结果，订阅结果包括目标AI服务信息，目标AI服务信息对应的AI服务满足第二条件。
可选的，本申请实施例中的无线接入网设备是一种将终端设备接入到无线网络的 设备。无线接入网设备可以是基站(base station)、演进型基站(evolved NodeB，eNodeB)、发送接收点(transmission reception point，TRP)、5G移动通信系统中的下一代基站(next generation NodeB，gNB)、未来移动通信系统中的基站或无线保真(wireless-fidelity，Wi-Fi)系统中的接入节点等；也可以是完成基站部分功能的模块或单元。例如，集中式单元(central unit，CU)，分布式单元(distributed unit，DU)等。本申请的实施例对无线接入网设备所采用的具体技术和具体设备形态不做限定。
可选的，本申请实施例中的终端设备，可以是用于实现无线通信功能的设备。例如终端或者可用于终端中的芯片等。终端也可以称为用户设备(user equipment，UE)、移动台、移动终端等。终端可以是手机、平板电脑、带无线收发功能的电脑、虚拟现实终端设备、增强现实终端设备、工业控制中的无线终端、无人驾驶中的无线终端、远程手术中的无线终端、智能电网中的无线终端、运输安全中的无线终端、智慧城市中的无线终端、智慧家庭中的无线终端等等。本申请的实施例对终端设备所采用的具体技术和具体设备形态不做限定。
可选的，本申请实施例中的无线接入网设备和终端设备之间可以通过授权频谱进行通信，也可以通过免授权频谱进行通信，也可以同时通过授权频谱和免授权频谱进行通信。无线接入网设备和终端设备之间可以通过6千兆赫兹(gigahertz，GHz)以下的频谱进行通信，也可以通过6GHz以上的频谱进行通信，还可以同时使用6GHz以下的频谱和6GHz以上的频谱进行通信。本申请的实施例对无线接入网设备和终端设备之间所使用的频谱资源不做限定。
可选的，本申请实施例中的网络设备和终端设备可以部署在陆地上，包括室内或室外、手持或车载；也可以部署在水面上；还可以部署在空中的飞机、气球和人造卫星上。本申请的实施例对网络设备和终端设备的应用场景不做限定。
本申请涉及的AI节点或终端设备的相关功能可以由一个设备实现，也可以由多个设备共同实现，还可以是由一个设备内的一个或多个功能模块实现，或者可以为一个或多个芯片，也可以为片上系统(system on chip，SOC)或芯片系统，芯片系统可以由芯片构成，也可以包括芯片和其他分立器件，本申请实施例对此不作具体限定。
可以理解的是，上述功能既可以是硬件设备中的网络元件，也可以是在专用硬件上运行的软件功能，或者是硬件与软件的结合，或者是平台(例如，云平台)上实例化的虚拟化功能。
例如，本申请涉及的AI节点或终端设备的相关功能可以通过图2中的通信装置200来实现。图2所示为本申请实施例提供的通信装置200的结构示意图。该通信装置200包括一个或多个处理器201，通信线路202，以及至少一个通信接口(图2中仅是示例性的以包括通信接口204，以及一个处理器201为例进行说明)，可选的，还可以包括存储器203。
处理器201可以是一个通用中央处理器(central processing unit，CPU)，微处理器，特定应用集成电路(application-specific integrated circuit，ASIC)，或一个或多个用于控制本申请方案程序执行的集成电路。
通信线路202可以用于通信装置200包括的不同组件之间的通信。
通信接口204，可以是收发模块用于与其他设备或通信网络通信，如以太网，无 线接入网(wireless access networks，RAN)，无线局域网(wireless local area networks，WLAN)等。例如，所述收发模块可以是收发器、收发机一类的装置。可选的，所述通信接口204也可以是位于处理器201内的收发电路，用以实现处理器的信号输入和信号输出。
存储器203可以是具有存储功能的装置。例如可以是只读存储器(read-only memory，ROM)或可存储静态信息和指令的其他类型的静态存储设备，随机存取存储器(random access memory，RAM)或者可存储信息和指令的其他类型的动态存储设备，也可以是电可擦可编程只读存储器(electrically erasable programmable read-only memory，EEPROM)、只读光盘(compact disc read-only memory，CD-ROM)或其他光盘存储、光碟存储(包括压缩光碟、激光碟、光碟、数字通用光碟、蓝光光碟等)、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质，但不限于此。存储器可以是独立存在，通过通信线路202与处理器相连接。存储器也可以和处理器集成在一起。
其中，存储器203用于存储执行本申请方案的计算机执行指令，并由处理器201来控制执行。处理器201用于执行存储器203中存储的计算机执行指令，从而实现本申请实施例中提供的方法。
或者，可选的，本申请实施例中，也可以是处理器201执行本申请下述实施例提供的方法中的处理相关的功能，通信接口204负责与其他设备或通信网络通信，本申请实施例对此不作具体限定。
可选的，本申请实施例中的计算机执行指令也可以称之为应用程序代码，本申请实施例对此不作具体限定。
在具体实现中，作为一种实施例，处理器201可以包括一个或多个CPU，例如图2中的CPU0和CPU1。
在具体实现中，作为一种实施例，通信装置200可以包括多个处理器，例如图2中的处理器201和处理器208。这些处理器中的每一个可以是一个单核(single-core)处理器，也可以是一个多核(multi-core)处理器。这里的处理器可以包括但不限于以下至少一种：中央处理单元(central processing unit，CPU)、微处理器、数字信号处理器(DSP)、微控制器(microcontroller unit，MCU)、或人工智能处理器等各类运行软件的计算设备，每种计算设备可包括一个或多个用于执行软件指令以进行运算或处理的核。
在具体实现中，作为一种实施例，通信装置200还可以包括输出设备205和输入设备206。输出设备205和处理器201通信，可以以多种方式来显示信息。例如，输出设备205可以是液晶显示器(liquid crystal display，LCD)，发光二极管(light emitting diode，LED)显示设备，阴极射线管(cathode ray tube，CRT)显示设备，或投影仪(projector)等。输入设备206和处理器201通信，可以以多种方式接收用户的输入。例如，输入设备206可以是鼠标、键盘、触摸屏设备或传感设备等。
需要说明的是，图2中示出的组成结构并不构成对该通信装置的限定，除图2所示部件之外，该通信装置可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置。图示的部件可以以硬件，软件或软件和硬件的组合实现。
下面将结合图1a至图2，以图1a或图1b所示的第一AI节点20与任一终端设备30进行交互为例，对本申请实施例提供的一种AI服务信息获取方法进行展开说明。如图3所示，该AI服务信息获取方法包括S301-S302：
S301：第一AI节点向终端设备发送测量配置信息；测量配置信息用于指示测量待测量AI节点的AI服务。
S302：第一AI节点接收来自终端设备的测量报告；测量报告包括候选AI服务信息；候选AI服务信息为根据测量配置信息对待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；第一条件携带在测量配置信息中。
对于S301，第一AI节点通过对应的网络设备，向终端设备发送测量配置信息，终端设备接收到测量配置信息后，根据测量配置信息测量待测量AI节点的AI服务。
本申请实施例中，测量配置信息包括第一条件、待测量AI节点的标识信息和需要测量的AI服务的标识信息，以下分别进行介绍。
测量配置信息中，第一条件用于指示终端设备根据测量配置信息进行测量后，根据测量结果向第一AI节点发送(或者说上报)测量报告所需满足的条件，因此第一条件也可以称为测量报告的上报条件。可选的，根据实际需求，测量配置信息可以包括一个或多个第一条件。示例性的，第一条件可以为AI服务的性能需要满足的阈值，若测量的AI服务的性能满足该阈值，则终端设备可以将测量该AI服务得到的测量结果上报给第一AI节点，否则不上报。
测量配置信息中，待测量AI节点的标识信息用于终端设备确定需要测量的待测量AI节点，其也可以称为测量列表信息。可选的，根据需要测量的待测量AI节点，测量配置信息可以包括一个或多个待测量AI节点的标识信息。示例性的，待测量AI节点的标识信息可以为待测量AI节点的标识号(identity document，ID)，即待测量AI节点的AI节点ID(AI node ID)。可选的，本申请实施例中，需要测量的待测量AI节点可以是第一AI节点根据实际需要或者业务场景确定的，一定区域范围内的AI节点。
测量配置信息中，需要测量的AI服务的标识信息用于终端设备确定需要测量的AI服务，换言之，用于终端设备确定需要测量的测量目标，需要测量的AI服务的标识信息也可以称为测量目标信息。可选的，根据需要测量的AI服务，测量配置信息可以包括一个或多个需要测量的AI服务的标识信息。示例性的，AI服务的标识信息可以为AI服务的ID，即AI服务ID(AI service ID)。可选的，本申请实施例中，需要测量的AI服务可以是第一AI节点根据实际需要或者业务场景确定的。
可选的，本申请实施例中，测量配置信息中，第一条件可以与需要测量的AI服务的标识信息呈对应关系。可以理解的是，不同的AI服务可以对应不同的上报条件，例如，AI service ID为1的AI服务对应的第一条件可以为性能高于第一阈值，AI service ID为2的AI服务对应的第一条件可以为性能高于第二阈值。或者，不同的AI服务也可以对应相同的上报条件，测量配置信息中可以仅包括一个第一条件，本申请实施例对此不作具体限制。
需要说明的是，本申请实施例中，AI节点的标识信息与AI节点间的对应关系，以及AI服务的标识信息与AI服务间的对应关系是终端设备和AI节点间协议或配置 好的，换言之，终端设备和AI节点可以根据相同的AI节点的标识信息确定同一AI节点，也可以根据相同的AI服务的标识信息确定同一AI服务。
终端设备接收到测量配置信息后，可以根据测量配置信息包括的信息，确定需要测量的待测量AI节点，需要测量的AI服务，以及上报测量结果所需满足的条件。终端设备根据测量配置信息的指示，对待测量AI节点的需要测量的AI服务进行测量后，对于每个待测量AI节点，终端设备可以得到一个或多个AI服务信息，即该待测量AI节点的测量结果。然后，终端设备在得到的候选AI服务信息中，将满足第一条件的AI服务信息确定为候选AI服务信息。
本申请实施例中，终端设备可以通过待测量AI节点对应的网络设备(例如RAN设备)，对待测量AI节点进行测量。
以下对AI服务信息进行具体介绍。本申请实施例中，终端设备测量待测量节点得到的AI服务信息可以包括以下至少一项信息：
AI服务标识信息、AI节点标识信息、模型标识信息、模型参数信息、性能信息和时效信息。
其中，AI服务标识信息用于标识该AI服务信息对应的AI服务，换言之，终端设备可以通过AI服务标识信息指示测量的AI服务的类型。
AI节点标识信息用于标识该AI服务信息对应的AI节点，换言之，终端设备可以通过AI节点标识信息指示测量的AI节点。
模型标识信息用于标识该AI服务信息对应的神经网络模型。本申请实施例中，AI节点上可以配置一种或多种神经网络模型来实现AI服务，终端设备可以通过模型标识信息指示测量的神经网络模型的种类。
模型参数信息为该AI服务信息对应的神经网络模型的参数信息。本申请实施例中，模型参数信息可以包括一项或多项参数信息，例如包括神经网络模型的规格(size)、输入格式(input format)或者输出格式(output format)等信息，本申请实施例对此不作具体限制。
性能信息用于表征该AI服务信息对应的AI服务的性能。本申请实施例中，性能信息可以包括一项或多项维度信息，例如包括精确度(accuracy)、推理精度或者模型训练收敛时间等信息，本申请实施例对此不作具体限制。以下结合一个具体示例对AI服务的性能进行解释，假设AI服务为图像分类，通过该AI服务可以在100张图片中正确分类99张，则该AI服务的精确度为99％。可选的，性能信息可以通过数值、区间或者协议规定的等级等信息表征，例如性能信息包括精确度时，规定精确度为90％-95％时对应的等级为A，若终端设备测量AI服务得到的精确度为90％，则AI服务信息中，性能信息包括的精确度可以为数值90％，或者为等级A。
时效信息用于表征该AI服务信息的有效时间，在有效时间内该AI服务信息有效，超过有效时间则失效。可选的，时效信息可以通过时刻表征，例如时刻0(time0，T0)～时刻1(time1，T1)，或者，也可以通过时长表征，例如5分钟(minute，min)。
示例性的，假设AI服务信息包括AI服务标识信息、AI节点标识信息、模型参数信息、性能信息和时效信息，其中模型参数信息包括size、input format以及output format参数信息，性能信息包括accuracy信息，则AI服务信息一种可能的形式可以如下表1 所示：
表1
上表1中，AI service ID为AI服务标识信息，其对应的service 1表示该AI服务信息是测量AI服务：service 1得到的。AI node ID为AI节点标识信息，其对应的001表示该AI服务信息是测量AI节点001得到的。Model information为模型参数信息，其包括的size以及size对应的3，10表示该AI节点001上的神经网络模型有3层，每层有10个神经元，input format以及input format对应的整数型向量表示该AI节点001上的神经网络模型的输入格式为整数型向量，output format以及output format对应的整数型向量表示该AI节点001上的神经网络模型的输入格式为整数型向量。performance为性能信息，其包括的accuracy以及accuracy对应的99％表示AI服务service 1的精确度为99％。validity为时效信息，其对应的T0～T1表示该AI服务信息的有效时间从T0持续到T1。
可以理解的是，因为候选AI服务信息为终端设备测量待测量节点得到的AI服务信息中，满足第一条件的AI服务信息，因此若将上文包括的信息的介绍中，AI服务信息替换为候选AI服务信息，AI节点替换为待测量AI节点，则上文对AI服务信息的介绍同样可以适用于候选AI服务信息。
对于S302，终端设备确定候选AI服务信息后，通过测量报告将候选AI服务信息发送给第一AI节点，从而使第一AI节点可以根据接收的测量报告获取到满足第一条件的AI服务信息。
基于本申请实施例的技术方案，AI节点可以通过与终端设备的交互，获取其他AI节点的满足一定条件的AI服务信息，便于AI网络中不同AI节点间的协同管理。
可选的，本申请实施例中，终端设备可以通过向相连的网络设备发送测量报告，由终端设备相连的网络设备向第一AI节点发送测量报告，该测量报告包含一个或多个待测量AI节点对应的候选AI服务信息。其中，假设第一AI节点对应的网络设备为第一网络设备，若终端设备相连的网络设备并非第一网络设备，终端设备相连的网络设备可以通过转发的方式向第一网络设备发送测量报告，从而向第一AI节点发送测量报告。
或者，终端设备可以通过相连的网络设备，向待测量AI节点发送待测量AI节点对应的测量报告，以使待测量AI节点向第一AI节点发送各自的测量报告。其中，待测量AI节点对应的测量报告中包括待测量AI节点对应的候选AI服务信息，不包括其他待测量AI节点对应的候选AI服务信息。假设待测量AI节点中某一待测量节点为第二AI节点，第二AI节点对应的网络设备为第二网络设备，终端设备可以将包括 第二AI节点对应的候选AI服务信息的测量报告发送给终端设备相连的网络设备，以使终端设备相连的网络设备向第二网络设备发送第二AI节点对应的测量报告。其中，终端设备相连的网络设备可以通过转发的方式向第二网络设备发送测量报告。
可选的，终端设备通过相连的网络设备向待测量AI节点发送待测量AI节点对应的测量报告，以使待测量AI节点向第一AI节点发送各自的测量报告的方式中，终端设备发送的测量报告还包括第一AI节点的标识信息，待测量AI节点对应的网络设备(例如第二网络设备)可以根据第一AI节点的标识信息，将测量报告发送给第一AI节点。
可选的，终端设备通过相连的网络设备向待测量AI节点发送待测量AI节点对应的测量报告，以使待测量AI节点向第一AI节点发送各自的测量报告的方式中，测量报告的转发可以由AI节点独立部署的网络设备间的接口完成，也可以由彼此相连的，可以为终端设备提供接入网络功能的网络设备(例如RAN设备)间的接口完成。
以上介绍了S301-S302的技术方案。可选的，在S302之后，第一AI节点可以根据候选AI服务信息，确定目标AI服务信息，从而更新第一AI节点的本地AI服务列表。以下对第一AI节点如何根据候选AI服务信息更新本地AI服务列表进行展开介绍。
一种可能的实现方式中，第一AI节点可以在同类型的AI服务对应的一个或多个候选AI服务信息中，将AI服务的性能最优的候选AI服务信息确定为目标AI服务信息，并将目标AI服务信息添加至本地AI服务列表中。其中，AI服务的类型可以通过候选AI服务信息中包括的AI服务标识信息确定，同一AI服务标识信息对应的AI服务的类型相同。AI服务的性能可以通过候选AI服务信息中包括的性能信息确定。
换言之，第一AI节点可以根据候选AI服务信息中包括的AI服务标识信息和性能信息，分别将候选AI服务信息对应的一个或多个AI服务中，每个AI服务对应的候选AI服务信息中的目标AI服务信息添加至本地AI服务列表中，其中，目标AI服务信息为每个AI服务对应的候选AI服务信息中，AI服务的性能最优的候选AI服务信息。
可选的，第一AI节点在根据候选AI服务信息中的性能信息确定性能最优的候选AI服务信息时，可以根据预设的规则来确定性能最优的候选AI服务信息。例如，预设规则可以为若性能信息以数值表征，性能信息的数值最大的候选AI服务信息是性能最优的候选AI服务信息。
可选的，该实现方式中，第一AI节点可以定时确定目标AI服务信息，例如每隔10分钟，第一AI节点就在这10分钟内接收到的所有候选AI服务信息中确定目标AI服务信息。或者，第一AI节点可以在接收的候选AI服务信息的数量满足预设阈值时确定目标AI服务信息，例如，第一AI节点可以每接收到10条候选AI服务信息就在这10条候选AI服务信息中确定目标AI服务信息。
另一种可能的实现方式中，第一AI节点可以将候选AI服务信息中，与本地AI服务列表中记录的AI服务信息不同的候选AI服务信息确定为目标AI服务信息，并将目标AI服务信息添加至本地AI服务列表中。换言之，若第一AI节点确定候选AI服务信息已经记录在本地AI服务列表中，则第一AI节点结束流程，不执行之后的步 骤，否则第一AI节点将候选AI服务信息添加，或者说更新至本地AI服务列表。
该实现方式中，与本地AI服务列表中记录的AI服务信息不同的候选AI服务信息，指与本地AI服务列表中记录的AI服务信息的类型不同的候选AI服务信息。可选的，AI服务信息的类型可以通过AI服务标识信息表征，第一AI节点可以根据AI服务信息包括的AI服务标识信息，确定候选AI服务信息的类型是否与本地AI服务列表中记录的AI服务信息的类型不同，从而决定是否将候选AI服务信息添加至本地AI服务列表中。
可选的，本申请实施例中，上文介绍的两种根据候选AI服务信息更新本地AI服务列表的实现方式可以独立应用，也可以结合应用。可以理解的是，当这两种实现方式结合应用时，本地AI服务列表中保留的AI服务信息对应的AI服务是同类型的AI服务中，性能最优的AI服务。
示例性的，假设第一AI节点可以独立部署在专用于实现AI节点功能的网络设备上，第一AI节点部署的网络设备与RAN设备对应，该RAN设备可以与终端设备通信。第一AI节点通过RAN设备指示终端设备对待测量AI节点进行测量，并根据终端设备上报的测量报告更新本地AI服务列表的一种可能的示意图可以如图4所示。如图4所示，第一AI节点部署的网络设备具有模型管理和模型测量两个模块，并维护一张本地AI服务列表。其中，模型测量模块通过与RAN设备的无线资源控制(radio resource control，RRC)是指连接，向终端设备下发测量配置信息，指示终端设备测量待测量AI节点，并接收终端设备上报的测量结果(候选AI服务信息)。模型管理模块可以通过模型测量模块收集的测量结果，更新本地AI服务列表。其中，模型测量模块指示终端设备测量待测量AI节点并接收终端设备上报的测量结果的具体实现可以参考上文对S301的介绍，模型管理模块更新本地AI服务列表的具体实现可以参考上文对第一AI节点根据候选AI服务信息更新本地AI服务列表的介绍。
进一步地，本申请实施例中，第一AI节点根据候选AI服务信息更新本地AI服务列表后，还可以根据本地AI服务列表中记录的一个或多个AI服务信息，从对应的AI节点处获取神经网络模型来实现AI服务。
具体地，候选AI服务信息中包括AI节点标识信息，可以理解的是，因此本地AI服务列表中记录的AI服务信息也包括AI节点标识信息，第一AI节点可以根据本地AI服务列表中的一个或多个AI服务信息包括的AI节点标识信息，确定与一个或多个AI服务信息对应的AI节点(以下简称为目标AI节点)，并从目标AI节点处获取神经网络模型。
可选的，本申请实施例中，第一AI节点可以定时从对应的目标AI节点处获取对应的神经网络模型，例如每隔10分钟，第一AI节点就根据这10分钟内本地AI服务列表添加的AI服务信息，确定目标AI节点并从目标AI节点处获取对应的神经网络模型。或者，第一AI节点可以在本地AI服务列表中新添加的AI服务信息的数量满足预设阈值时，从对应的目标AI节点处获取对应的神经网络模型，例如，本地AI服务列表中每新增10条AI服务信息，第一AI节点就根据这新增的10条AI服务信息确定目标AI节点并从目标AI节点处获取神经网络模型。
可选的，本申请实施例中，第一AI节点可以通过以下两种不同的实现方式从目标 AI节点处获取神经网络模型。
一种可能的实现方式中，候选AI服务信息包括模型标识信息，可以理解的是，因此本地AI服务列表中记录的AI服务信息也包括模型标识信息。第一AI节点确定与本地AI服务列表中的AI服务信息对应的目标AI节点后，可以根据本地AI服务列表中的AI服务信息包括的模型标识信息，从目标AI节点上获取与模型标识信息对应的神经网络模型。
另一种可能的实现方式中，第一AI节点可以从目标AI节点处获取目标AI节点预配置的神经网络模型。该实现方式中，目标AI节点上可以预配置有一个或多个神经网络模型。
可选的，本申请实施例中，第一AI节点可以通过向目标AI节点发送请求消息的方式，向目标AI节点请求神经网络模型，目标AI节点可以通过向第一AI节点发送响应消息的方式，将神经网络模型发送给第一AI节点。
以下结合两个示例和对应的附图，对本申请实施例提供的，第一AI节点向终端设备发送测量配置信息，接收来自终端设备的测量报告，并根据测量报告从目标AI节点处获取神经网络模型的方案进行介绍。
示例一，假设待测量AI节点中某一待测量AI节点为第二AI节点，AI节点独立部署在专用于实现AI节点功能的网络设备上，第一AI节点与RAN设备1对应，第二AI节点与RAN设备2对应，第一AI节点与第二AI节点之间可以通过分别部署的网络设备通信。如图5所示，第一AI节点从第二AI节点处获取神经网络模型的一种可能的方案包括S501-S504：
S501：第一AI节点通过RAN设备1向终端设备发送测量配置信息。S501的具体实现可以参考上文对S301的介绍。
S502：终端设备通过RAN设备2对第二AI节点进行测量，获取测量结果(第二AI节点对应的AI服务信息)。S502的具体实现可以参考上文对S301的介绍。
S503：终端设备确定第二AI节点对应的AI服务信息满足第一条件，将第二AI节点对应的AI服务信息携带在测量报告中，通过RAN设备1发送给第一AI节点。S503的具体实现可以参考上文对S302的介绍。
S504：终端设备将第二AI节点对应的AI服务信息添加至本地AI服务列表，并从第二AI节点处获取对应的神经网络模型。S504的具体实现可以参考上文对第一AI节点的从目标AI节点处获取神经网络模型的介绍。
示例二，假设待测量AI节点中某一待测量AI节点为第二AI节点，AI节点独立部署在专用于实现AI节点功能的网络设备上，第一AI节点与RAN设备1对应，第二AI节点与RAN设备2对应，第一AI节点与第二AI节点之间可以通过分别部署的网络设备通信。如图6所示，第一AI节点从第二AI节点处获取神经网络模型的一种可能的方案包括S601-S604：
S601：第一AI节点通过RAN设备1向终端设备发送测量配置信息。S601的具体实现可以参考上文对S301的介绍。
S602：终端设备对第二AI节点进行测量，获取测量结果(第二AI节点对应的AI服务信息)。S602的具体实现可以参考上文对S301的介绍。
S603：终端设备确定第二AI节点对应的AI服务信息满足第一条件，将第二AI节点对应的AI服务信息携带在测量报告中，通过RAN设备2发送给第一AI节点。S603的具体实现可以参考上文对S302的介绍。
S604：第一AI节点将第二AI节点对应的AI服务信息添加至本地AI服务列表，并从第二AI节点处获取对应的神经网络模型。S604的具体实现可以参考上文对第一AI节点的从目标AI节点处获取神经网络模型的介绍。
下面将结合图1a至图2，以图1a或图1b所示的第三AI节点20与任一终端设备30进行交互为例，对本申请实施例提供的另一AI服务信息获取方法进行展开说明。如图7所示，该AI服务信息获取方法包括S701-S703：
S701：终端设备向第三AI节点发送订阅信息，相对应的，第三AI节点接收来自终端设备的订阅信息。其中，订阅信息用于请求在订阅区域内订阅满足第二条件的AI服务。
S702：第三AI节点向订阅区域内的一个或多个第四AI节点发送订阅信息。
S703：订阅区域内的一个或多个第四AI节点向终端设备发送订阅结果，相对应的，终端设备接收来自一个或多个第四AI节点的订阅结果。其中，订阅结果包括目标AI服务信息，目标AI服务信息对应的AI服务满足第二条件。
对于S701，终端设备通过向第三AI节点发送订阅信息来请求订阅终端设备期望获得的AI服务，即请求订阅满足第二条件的AI服务。本申请实施例中，订阅信息包括以下至少一项信息：
AI服务标识信息，用户标识信息，服务等级信息，订阅区域信息，模型标识信息，模型参数信息和订阅有效期信息。
以下分别对上述订阅信息可能包括的信息进行介绍。
订阅信息中的AI服务标识信息用于标识终端设备请求订阅的AI服务，换言之，终端设备可以通过订阅信息中的AI服务标识信息指示期望获得的AI服务的类型。可选的，本申请实施例中，终端设备请求订阅的AI服务可以是终端设备根据实际需要或者业务场景确定的。可选的，根据终端设备请求订阅的AI服务，订阅信息中可以包括一个或多个AI服务标识信息。示例性的，AI服务的标识信息可以为AI服务的ID，即AI服务ID(AI service ID)。
订阅信息中的用户标识信息用于标识发送该订阅信息的终端设备，用户标识信息也可以称为订阅发起方标识。可选的，本申请实施例中，用户标识信息可以复用现有的用于标识终端设备的信息，例如用户标识信息可以为临时移动用户识别码(temporary mobile subscriber identity，TMSI)。或者，用户标识信息可以为新定义的，用于在AI网络中标识终端设备的标识信息。
订阅信息中的服务等级信息用于表征第二条件，换言之，用于表征终端设备期望获得的AI服务所需满足的条件。可选的，根据实际需求，订阅信息可以包括一个或多个服务等级信息。示例性的，服务等级信息可以为请求订阅的AI服务的性能需要满足的阈值，若第四AI节点的AI服务的性能满足该阈值，则第四AI节点可以将与订阅信息对应的订阅结果反馈给终端设备，否则不反馈订阅结果。
可选的，本申请实施例中，不同的AI服务可以对应不同的第二条件，换言之，订 阅信息中，服务等级信息可以与AI服务的标识信息呈对应关系。例如，AI service ID为1的AI服务对应的服务等级信息可以为性能高于第一阈值，AI service ID为2的AI服务对应的服务等级信息可以为性能高于第二阈值。或者，不同的AI服务也可以对应相同的第二条件，订阅信息中可以仅包括同一服务等级信息，本申请实施例对此不作具体限制。
订阅信息中的订阅区域信息用于表征订阅区域。可选的，本申请实施例中，订阅区域信息可以包括订阅的AI节点的标识信息，第三AI节点根据订阅信息中AI节点的标识信息，向对应的AI节点发送订阅信息。可以理解的是，与订阅信息中每个AI节点的标识信息分别对应的每个AI节点也可以称为订阅区域内的每个AI节点。或者，订阅信息可以包括预先规划的订阅区域的标识信息，第三AI节点根据订阅区域的标识信息，向对应的订阅区域内的所有AI节点发送订阅信息。
可选的，本申请实施例中，订阅区域信息可以是终端设备或者第三AI节点根据实际需要或者业务场景确定的。其中，若由终端设备确定订阅区域信息，终端设备可以向第三AI节点发送包括订阅区域信息的订阅信息，第三AI节点将收到的订阅信息发送给订阅区域内的AI节点。若由第三AI节点确定订阅区域信息，第三AI节点在收到来自终端设备的不包括订阅区域信息的订阅信息后，确定订阅区域信息并携带在订阅信息中，再将携带订阅区域信息的订阅信息发送给订阅区域内的AI节点。
订阅信息中的模型标识信息用于标识订阅信息请求订阅的AI服务对应的神经网络模型。终端设备可以通过模型标识信息指示用于实现请求订阅的AI服务的神经网络模型的种类。
可选的，本申请实施例中，订阅信息中的AI服务的标识信息可以与模型标识信息呈对应关系，同一AI服务的标识信息可以对应一个或多个模型标识信息。
订阅信息中的模型参数信息为订阅信息请求订阅的AI服务对应的神经网络模型的参数信息，用于指示终端设备期望的神经网络模型的参数信息。本申请实施例中，订阅信息中的模型参数信息可以包括一项或多项参数信息，例如包括神经网络模型的size、input format或者output format等信息，本申请实施例对此不作具体限制。示例性的，假设终端设备请求订阅AI服务，终端设备期望用于实现该AI服务的神经网络模型具有100个神经元，则订阅信息中，模型参数信息包括的size信息可以为100。
可选的，本申请实施例中，订阅信息中的模型标识信息可以与模型参数信息呈对应关系。可以理解的是，对于不同的神经网络模型，终端设备期望的参数信息也可以不同。例如，假设模型参数信息包括size信息，终端设备期望用于实现AI服务的神经网络模型1具有100个神经元，期望用于实现AI服务的神经网络模型2具有200个神经元，则订阅信息中，模型标识信息1对应的size信息可以为100，模型标识信息2对应的size信息可以为200。或者，不同的神经网络模型(模型标识信息)也可以对应相同的模型参数信息，本申请实施例对此不作具体限制。
订阅信息中的订阅有效期信息用于表征订阅信息的有效时间，在有效时间内该订阅信息有效，超过有效时间则失效。可选的，订阅有效期信息可以通过时刻表征，例如时刻0(time0，T0)～时刻1(time1，T1)，或者，也可以通过时长表征，例如5分钟。
可选的，本申请实施例中，订阅信息可以由第一标识信息标识。其中，第一标识信息可以为新定义的，专用于标识订阅信息的标识信息。示例性的，第一标识信息的一种可能的形式可以为订阅标识号(subscribe-ID，S-ID)。
可选的，本申请实施例中，第一标识信息可以由第三AI节点生成，该实现方式中，第三AI节点可以在收到来自终端设备的订阅信息后，生成第一标识信息，并将第一标识信息携带在响应消息中发送给终端设备，其中，响应消息用于响应终端设备发送的订阅信息。或者，第一标识信息可以由终端设备生成，该实现方式中，终端设备向第三AI节点发送由第一标识信息标识的订阅信息。
对于S702，第三AI节点获取到订阅信息后，向订阅区域内的每个AI节点分别发送订阅信息。接收到订阅信息的AI节点可以根据订阅信息，确定自身能否提供终端设备请求请阅的AI服务，若可以提供终端设备请求订阅的AI服务，则监测自身提供的对应的AI服务是否满足第二条件。本申请实施例中，将可以提供满足第二条件的请求订阅的AI服务的AI节点称为第四AI节点。换言之，第三AI节点可以向订阅区域内的一个或多个第四AI节点发送订阅信息。
可选的，本申请实施例中，第三AI节点可以为订阅区域内的AI节点，该情况中，第三AI节点也可以根据订阅信息监测自身能否提供满足第二条件的对应的AI服务，可以理解的是，若第三AI节点可以提供满足第二条件的对应的AI服务，则第三节点也为一个第四AI节点。或者，第三AI节点也可以不是订阅区域内的AI节点，则第三AI节点无需监测自身能否提供满足第二条件的对应的AI服务。
对于S703，订阅区域内的第四AI节点确定自身提供的AI服务可以满足第二条件后，将满足第二条件的AI服务对应的AI服务信息携带在订阅结果中发送给终端设备，以使终端设备获取期望的AI服务对应的AI服务信息。其中，满足第二条件的AI服务对应的AI服务信息也可以称为目标AI服务信息。
基于本申请实施例提供的AI服务信息获取方法，终端设备可以向AI节点订阅期望的AI服务，并获取到满足条件的AI服务对应的AI服务信息，从而使终端设备可以在AI网络中订阅到满足要求的AI服务。
可选的，本申请实施例中，满足第二条件的AI服务对应的AI服务信息可以参考上文对S301中AI服务信息的介绍。
可选的，本申请实施例中，第四AI节点将订阅结果发送给终端设备的方式有以下两种：
实现方式一：第四AI节点将订阅结果发送给第三AI节点，由第三AI节点将订阅结果发送给终端设备。
实现方式二：第四AI节点将订阅结果直接发送给终端设备。需要说明的是，该实现方式中的“直接”指不需要经过第三AI节点，并非指无需中间设备的转发。
本申请实施例中，第三AI节点或者第四AI节点将订阅结果发送给终端设备的方式，与终端设备的状态相关，以下进行展开介绍。
若终端设备处于连接态(active state)，第三AI节点或者第四AI节点可以通过终端设备当前接入的RAN设备，将订阅结果发送给终端设备。示例性的，终端设备当前接入的RAN设备，可以通过终端专用参数信号(UE-specific)消息，将订阅结果发 送给终端设备。其中，第三AI节点或者第四AI节点将订阅结果发送给终端设备当前接入的RAN设备的具体实现，可以参考上文对终端设备向第一AI节点发送测量报告的具体实现的介绍。
若终端设备处于空闲态(idle state)或者非激活态(inactive state)，第四AI节点将订阅结果发送给终端设备的方式与下文将介绍的本申请实施例提供的一种对终端设备进行移动性管理的方案相关，在此先不展开介绍。
可选的，在S703之后，本申请实施例提供的AI服务获取方法还可以包括S704.a或者S704.b：
S704.a：终端设备首次获取到与订阅信息对应的订阅结果后，告知当前驻留的AI节点，终端设备已经获取订阅结果，由该当前驻留的AI节点通知订阅区域内其他AI节点可以终止订阅任务。
其中，终端设备当前驻留的AI节点，指与终端设备当前相连的网络设备(例如RAN设备)对应的AI节点。终止订阅任务指AI节点可以停止监测自身能否提供满足订阅信息要求的(满足第二条件的)AI服务。
S704.b：订阅区域内，时间上第一个确定自身能提供订阅信息要求的AI服务的第四AI节点，在将订阅结果发送给终端设备的同时，一并通知订阅区域内的其他AI节点可以终止订阅任务。
以上对S701-S703进行了介绍，S701-S703的方案可以使终端设备在AI网络中订阅到满足要求的AI服务。但是，目前的AI网络还存在以下问题。目前，蜂窝系统定义了跟踪区(tracking area，TA)的概念，如图8所示，将多个小区定义为一个跟踪区，用于无线接入网络侧的移动性管理。当终端设备一直在一个跟踪区内移动时，不需要告知无线接入网络，当处于空闲态的终端设备被寻呼时，在终端设备所在的跟踪区的所有小区进行寻呼。但是，无线接入网络与AI网络逻辑上分离，目前还没有机制可以在AI网络侧对终端设备进行移动性管理。因此，本申请实施例还提供一种可以在AI网络侧对终端设备进行移动性管理的方案，以下对该方案进行说明。
可选的，S702中，第三AI节点获取到订阅信息之后，若订阅信息由第一标识信息标识，该AI服务信息获取方法还包括S702.1-S702.3：
S702.1:第三AI节点确定终端设备当前位于第一通知区域，将第一标识信息与第二标识信息相关联；第二标识信息用于标识第一通知区域。
S702.2:第三AI节点向一个或多个第四AI节点发送与第一标识信息关联的第二标识信息。
S702.3：终端设备获取与第一标识信息关联的第二标识信息。
S702.1中，第三AI节点可以根据终端设备的位置信息，确定终端设备当前位于第一通知区域，从而确定用于标识第一通知区域的第二标识信息，并建立第一标识信息与第二标识信息间的关联关系，得到与第一标识信息相关联的第二标识信息。本申请实施例中，通知区域(notification-area，N-Area)可以是新定义的，用于在AI网络中追踪终端设备，进行移动性管理的概念。具体如何根据通知区域对终端设备进行移动性管理在下文进行展开介绍。
可选的，本申请实施例中，通知区域(例如第一通知区域或者在下文介绍的第二 通知区域)是根据小区和/或波束设置或者说规划的。例如，通知区域可以以小区为单位进行规划，3个相邻小区构成一个通知区域。又例如，通知区域可以以波束为单位进行规划，3束波束构成一个通知区域。
可选的，本申请实施例中，用于标识通知区域的标识信息，例如用于标识第一通知区域的第二标识信息，其也可以是新定义的信息。示例性的，用于标识通知区域的标识信息的一种可能的形式可以为通知区域标识号(notification-area-ID，N-ID)。
S702.2中，第三AI节点将第二标识信息，以及第一标识信息与第二标识信息间的关联关系发送给订阅区域内的每个AI节点。可以理解的是，订阅区域内的一个或多个第四AI节点可以接收到第二标识信息，以及第一标识信息与第二标识信息间的关联关系，并根据与第一标识信息相关联的第二标识信息，确定终端设备当前位于第一通知区域。
本申请实施例中，第三AI节点可以同时向一个或多个第四AI节点发送由第一标识信息标识的订阅信息、第二标识信息以及第一标识信息与第二标识信息间的关联关系。或者，第三AI节点也可以先向一个或多个第四AI节点发送由第一标识信息标识的订阅信息，再发送第二标识信息以及第一标识信息与第二标识信息间的关联关系，本申请实施例不作具体限制。
S702.3中，终端设备可以通过以下方式获取与第一标识信息相关联的第二标识信息：
实现方式一：第三AI节点向终端设备发送与第一标识信息关联的第二标识信息。可选的，第三AI节点还可以向终端设备发送第一标识信息与第二标识信息的关联关系。该实现方式中，第三AI节点确定终端设备位于第一通知区域后，可以向第一通知区域内的网络设备发送第二标识信息，从而将第二标识信息发送给终端设备。可选的，第三AI节点可以在向终端设备发送的响应消息中携带第二标识信息，其中，该响应消息用于响应终端设备发送的订阅信息。
实现方式二：终端设备接收来自第一通知区域内的网络设备发送的第二标识信息。可选的，本申请实施例中，通知区域中的网络设备，例如RAN设备，可以广播用于标识该通知区域的标识信息，以使终端设备移动至该通知区域后，可以接收到用于标识该通知区域的标识信息。因此，该实现方式中，终端设备位于第一通知区域时，可以接收到第一通知区域内的RAN设备广播的第二标识信息。
需要说明的是，本申请实施例并不限制S702.3的时序在S702.1和S702.2之后。例如，若终端设备是通过第一通知区域内的RAN设备接收广播的第二标识信息，则终端设备可以在移动至第一通知区域后就获取到第二标识信息，而与第三AI节点何时将第一标识信息与第二标识信息关联无关。
以下展开介绍如何通过通知区域在AI网络中对终端设备进行移动性管理。
进一步地，S702.1-S702.3之后，若终端设备移动至与第一通知区域不同的第二通知区域，该AI服务信息获取方法还包括S702.4-S702.5：
S702.4：终端设备获取到第三标识信息，第三标识信息用于标识终端设备当前位于的第二通知区域。
可选的，终端设备可以接收来自第三网络设备的第三标识信息，其中，第三网络 设备位于第二通知区域中。示例性的，第二通知区域内的RAN设备广播用于标识第二通知区域的第三标识信息，终端设备移动至第二通知区域后，可以接收到RAN设备广播的第三标识信息。
S702.5：终端设备确定第三标识信息与第二标识信息不一致，向一个或多个第四AI节点发送更新请求，更新请求用于请求将第一标识信息与第三标识信息相关联。
S702.5中，终端设备如果确定获取的第三标识信息与之前获取的第二标识信息不一致，则说明终端设备已经移出之前位于的第一通知区域，且移动至第二通知区域。因此，终端设备需要向订阅区域内的每个AI节点(包括一个或多个第四AI节点)发送更新请求，请求将第一标识信息与第三标识信息相关联。
订阅区域内的AI节点接收到更新请求后，根据更新请求将第一标识信息与第二标识信息的关联关系更新为第一标识信息与第三标识信息的关联关系，更新后，第一标识信息与第三标识信息相关联。因此，订阅区域内的AI节点可以根据与第一标识信息相关联的第三标识信息，获知终端设备已经移动至第二通知区域。
为了便于理解如何通过通知区域在AI网络中对终端设备进行移动性管理，示例性的，如图9所示，假设终端设备向第三AI节点(图9中未示出)发送S-ID为S1的订阅信息，该S-ID为1的订阅信息对应的订阅区域包括通知区域1、通知区域2和通知区域3。通知区域1、通知区域2和通知区域3分别对应的N-ID为001、002和003。通知区域1、通知区域2和通知区域3对应一个AI节点2。第三AI节点在收到订阅信息后，确定终端设备当前位于通知区域1(第一通知区域)，则将S1与通知区域1的标识信息001相关联，并将关联后的信息(包括关联关系)以及订阅信息转发给AI节点2。终端设备位于第一通知区域时，接收到通知区域1内的RAN设备(图9中未示出)广播的N-ID：001。之后，终端设备从通知区域1移动至通知区域2，终端设备移动到通知区域2后，接收到通知区域2内的RAN设备(图9中未示出)广播的N-ID：002。终端设备确定002与001不一致，则向第三AI节和AI节点2发送更新请求，请求将S1与002相关联。第三AI节点与AI节点2接收到更新请求后，将与S1相关联的001更新为002。终端设备从通知区域2移动至通知区域3后，并根据通知区域3内的RAN设备广播的N-ID：003，确定003与002不一致，则会再次向第三AI节点与AI节点2发送更新请求，以使第三AI节点与AI节点2将与S1相关联的002更新为003。
可选的，在订阅信息由第一标识信息标识，且第一标识信息与用于标识通知区域的标识信息(例如第二标识信息或者第三标识信息)关联的情况下，第四AI节点可以通过与第一标识信息相关联的用于标识通知区域的标识信息，确定终端设备当前位于的通知区域，从而将订阅结果发送给终端设备。
示例性的，若终端设备处于空闲态或者非激活态，第四AI节点可以根据与第一标识信息相关联的用于标识通知区域的标识信息，确定终端设备当前位于的通知区域，从而指示终端设备当前位于的通知区域中所有RAN设备，根据用户标识信息或者第一标识信息，通过呼叫(paging)的方式寻呼终端设备，使终端设备与当前位于的通知区域内的某一RAN设备重建空口链接，从而使第四AI节点可以通过与终端设备建立空口链接的RAN设备将订阅结果发送至终端设备。
示例性的，若终端设备处于连接态，第四AI节点可以根据与第一标识信息相关联的用于标识通知区域的标识信息，确定终端设备当前位于的通知区域，从而可以更快速地将订阅结果发送给终端设备当前接入的RAN设备。
以下结合示例和附图，对本申请实施例提供的一种示例性的AI服务信息获取方法进行介绍。如图10所示，该方法包括S1001-S1007。
S1001：终端设备通过相连的RAN设备，向AI节点1(第三AI节点)发送订阅信息，该订阅信息也可以称为订阅请求(subscribe request)。
S1001中，订阅信息请求向AI节点2～AI节点N订阅满足第二条件的AI服务1，其中，N为大于等于2的正整数。S1001的具体实现可以参考上文对S701的介绍。
S1002：AI节点1将订阅信息转发给AI节点2～AI节点N。
AI节点1为订阅信息分配S-ID：S1(第一标识信息)，并根据终端设备当前位于的通知区域1的N-ID：001(第二标识信息)，将S1与001相关联。AI节点1将信息001，S1和001的关联关系，以及由S1标识的订阅信息发送给AI节点2～AI节点N。AI节点2～AI节点N接收到订阅信息后，监测自身提供的AI服务是否满足第二条件。
S1003：AI节点1将订阅信息的响应消息发送给终端设备。该响应消息也可以称为订阅响应(subscribe response)。
S1003中，AI节点1将与S1相关联的001信息携带在响应消息中发送给终端设备。
S1004：终端设备接收相连的RAN设备广播(broadcast)的消息。
S1004中，终端设备从通知区域1移动至通知区域2，并与通知区域2内的RAN设备相连。通知区域2内的RAN设备广播通知区域2的N-ID：002(第三标识信息)，终端设备与通知区域2内的RAN设备相连后，接收到该RAN设备广播的002。
S1005：终端设备向AI节点1和AI节点2～AI节点N发送更新请求，该更新请求也可以称为通知区域更新(N-Area update)请求。
S1005中，终端设备确定002与目前与S1关联的001不一致，则向AI节点1和AI节点2～AI节点N发送更新请求，请求将S1与002相关联。
S1006：AI节点1和AI节点2～AI节点N更新与S1关联的信息，AI节点2～N中有一个或多个AI节点确定自身提供的AI服务可以满足第二条件。
S1006中，AI节点1和AI节点2～AI节点N将与S1关联的001更新为002。且在此期间，AI节点2～AI节点N中有一个或多个AI节点确定自身提供的AI服务可以满足第二条件，即AI节点2～AI节点N中存在一个或多个第四AI节点。
S1002-S1006的具体实现可以参考上文对S702、S702.1-S702.5的介绍。
S1007：AI节点2～AI节点N中的一个或多个第四AI节点，将订阅结果发送给终端设备。
AI节点2～AI节点N中的一个或多个第四AI节点根据与S1相关联的002，确定终端设备当前位于通知区域2，将订阅结果发送给通知区域2内的RAN设备，从而通过通知区域002内的RAN设备将订阅结果发送给终端设备。S1007的具体实现可以参考上文对S703的介绍。
在本申请的各个实施例中，如果没有特殊说明以及逻辑冲突，不同的实施例之间 的术语和/或描述具有一致性、且可以相互引用，不同的实施例中的技术特征根据其内在的逻辑关系可以组合形成新的实施例。
可以理解的是，以上各个实施例中，由第一AI节点/第三AI节点实现的方法和/或步骤，也可以由第一AI节点/第三AI节点对应的网络设备，或者由可用于第一AI节点/第三AI节点对应的网络设备的部件(例如芯片或者电路)实现，由终端设备实现的方法和/或步骤，也可以由可用于终端设备的部件实现。
上述主要从各个装置之间交互的角度对本申请实施例提供的方案进行了介绍。相应的，本申请实施例还提供了通信装置，该通信装置用于实现上述各种方法。该通信装置可以为上述方法实施例中的第一AI节点/第三AI节点，或者上述方法实施例中的第一AI节点/第三AI节点对应的网络设备，或者包含上述第一AI节点/第三AI节点对应的网络设备的装置，或者为可用于第一AI节点/第三AI节点对应的网络设备的部件；或者，该通信装置可以为上述方法实施例中的终端设备，或者包含上述终端设备的装置，或者为可用于终端设备的部件。可以理解的是，该通信装置为了实现上述功能，其包含了执行各个功能相应的硬件结构和/或软件模块。本领域技术人员应该很容易意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，本申请能够以硬件或硬件和计算机软件的结合形式来实现。某个功能究竟以硬件还是计算机软件驱动硬件的方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
本申请实施例可以根据上述方法实施例中对通信装置进行功能模块的划分，例如，可以对应各个功能划分各个功能模块，也可以将两个或两个以上的功能集成在一个处理模块中。上述集成的模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。需要说明的是，本申请实施例中对模块的划分是示意性的，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式。
比如，以通信装置为上述方法实施例中的第一AI节点/第三AI节点为例。图11示出了一种通信装置110的结构示意图。该通信装置110包括发送模块1101和接收模块1102。所述发送模块1101和接收模块1102，也可以称为收发模块。所述发送模块1101和接收模块1102也可以称为收发单元，收发单元用以实现发送和/或接收功能，例如可以是收发电路，收发机，收发器或者通信接口。
其中，发送模块1101，可以用于执行上述方法实施例中由第一AI节点/第三AI节点执行的发送类的步骤，接收模块1102，可以用于执行上述方法实施例中由第一AI节点/第三AI节点执行的接收类的步骤。
例如，发送模块1101，用于向终端设备发送测量配置信息；测量配置信息用于指示测量待测量AI节点的AI服务；接收模块1102，用于接收来自终端设备的测量报告；测量报告包括候选AI服务信息；候选AI服务信息为根据测量配置信息对待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；第一条件携带在测量配置信息中。
可选的，通信装置110还可以包括处理模块1103，处理模块1103，可以用于执行上述方法实施例中由第一AI节点/第三AI节点执行的除接收和发送类步骤之外的其他 步骤。
其中，上述方法实施例涉及的各步骤的所有相关内容均可以援引到对应功能模块的功能描述，在此不再赘述。
在本实施例中，该通信装置110以采用集成的方式划分各个功能模块的形式来呈现。这里的“模块”可以指特定ASIC，电路，执行一个或多个软件或固件程序的处理器和存储器，集成逻辑电路，和/或其他可以提供上述功能的器件。在一个简单的实施例中，本领域的技术人员可以想到该通信装置110可以采用图2所示的通信装置200的形式。
比如，图2所示的通信装置200中的处理器201可以通过调用存储器203中存储的计算机执行指令，使得通信装置200执行上述方法实施例中的AI服务信息获取方法。
具体的，图11中的发送模块1101、接收模块1102和处理模块1103的功能/实现过程可以通过图2所示的通信装置200中的处理器201调用存储器203中存储的计算机执行指令来实现。或者，图11中的处理模块1103的功能/实现过程可以通过图2所示的通信装置200中的处理器201调用存储器203中存储的计算机执行指令来实现，图11中的发送模块1101和接收模块1102的功能/实现过程可以通过图2所示的通信装置200中的通信接口204来实现。
由于本实施例提供的通信装置110可执行上述的AI服务信息获取方法，因此其所能获得的技术效果可参考上述方法实施例，在此不再赘述。
或者，比如，以通信装置为上述方法实施例中的终端设备为例。图12示出了一种通信装置120的结构示意图。该通信装置120包括发送模块1201和接收模块1202。所述发送模块1201和接收模块1202，也可以合并称为收发模块。所述发送模块1201和接收模块1202也可以称为收发单元，收发单元用以实现发送和/或接收功能，例如可以是收发电路，收发机，收发器或者通信接口。
其中，发送模块1201，可以用于执行上述方法实施例中由终端设备执行的发送类的步骤，接收模块1202，可以用于执行上述方法实施例中由终端设备执行的接收类的步骤。
例如，接收模块1202，用于接收来自第一AI节点的测量配置信息；测量配置信息用于指示测量待测量AI节点的AI服务；发送模块1201，用于向第一AI节点发送测量报告；测量报告包括候选AI服务信息；候选AI服务信息为根据测量配置信息对待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；第一条件携带在测量配置信息中。
又例如，发送模块1201，用于向第三AI节点发送订阅信息；订阅信息用于请求在订阅区域内订阅满足第二条件的AI服务；接收模块1202，用于接收来自订阅区域内的一个或多个第四AI节点的订阅结果；订阅结果包括目标AI服务信息，目标AI服务信息对应的AI服务满足第二条件。
可选的，通信装置120还可以包括处理模块1203，处理模块1203，可以用于执行上述方法实施例中由终端设备执行的除接收和发送类步骤之外的其他步骤。
其中，上述方法实施例涉及的各步骤的所有相关内容均可以援引到对应功能模块的功能描述，在此不再赘述。
在本实施例中，该通信装置120以采用集成的方式划分各个功能模块的形式来呈现。这里的“模块”可以指特定ASIC，电路，执行一个或多个软件或固件程序的处理器和存储器，集成逻辑电路，和/或其他可以提供上述功能的器件。在一个简单的实施例中，本领域的技术人员可以想到该通信装置120可以采用图2所示的通信装置200的形式。
比如，图2所示的通信装置200中的处理器201可以通过调用存储器203中存储的计算机执行指令，使得通信装置200执行上述方法实施例中的AI服务信息获取方法。
具体的，图12中的发送模块1201、接收模块1202和处理模块1203的功能/实现过程可以通过图2所示的通信装置200中的处理器201调用存储器203中存储的计算机执行指令来实现。或者，图12中的处理模块1203的功能/实现过程可以通过图2所示的通信装置200中的处理器201调用存储器203中存储的计算机执行指令来实现，图12中的发送模块1201和接收模块1202的功能/实现过程可以通过图2所示的通信装置200中的通信接口204来实现。
由于本实施例提供的通信装置120可执行上述的AI服务信息获取方法，因此其所能获得的技术效果可参考上述方法实施例，在此不再赘述。
可选的，本申请实施例还提供了一种通信装置(例如，该通信装置可以是芯片或芯片系统)，该通信装置包括处理器，用于实现上述任一方法实施例中的方法。在一种可能的设计中，该通信装置还包括存储器。该存储器，用于保存必要的程序指令和数据，处理器可以调用存储器中存储的程序代码以指令该通信装置执行上述任一方法实施例中的方法。当然，存储器也可以不在该通信装置中。在另一种可能的设计中，该通信装置还包括接口电路，该接口电路为代码/数据读写接口电路，该接口电路用于接收计算机执行指令(计算机执行指令存储在存储器中，可能直接从存储器读取，或可能经过其他器件)并传输至该处理器。该通信装置是芯片系统时，可以由芯片构成，也可以包含芯片和其他分立器件，本申请实施例对此不作具体限定。
可选的，本申请实施例还提供一种通信装置，该通信装置可以包括处理器和接口电路，该接口电路，用于与该通信装置之外的其他模块通信，该处理器可以用于执行计算机程序或指令，以使该通信装置实现上述任一方法实施例中的方法。在一些场景下，该通信装置可以为芯片或芯片系统。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件程序实现时，可以全部或部分地以计算机程序产品的形式来实现。该计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、服务器或者数据中心通过有线(例如同轴电缆、光纤、数字用户线(digital subscriber line，DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存取的任何可用介质或者是包含一个或多个可以用介质集成的服务器、数据中心等数据存储设备。所述可用介质 可以是磁性介质(例如，软盘、硬盘、磁带)，光介质(例如，DVD)、或者半导体介质(例如固态硬盘(solid state disk，SSD))等。本申请实施例中,计算机可以包括前面所述的装置。
尽管在此结合各实施例对本申请进行了描述，然而，在实施所要求保护的本申请过程中，本领域技术人员通过查看所述附图、公开内容、以及所附权利要求书，可理解并实现所述公开实施例的其他变化。在权利要求中，“包括”(comprising)一词不排除其他组成部分或步骤，“一”或“一个”不排除多个的情况。单个处理器或其他单元可以实现权利要求中列举的若干项功能。相互不同的从属权利要求中记载了某些措施，但这并不表示这些措施不能组合起来产生良好的效果。
尽管结合具体特征及其实施例对本申请进行了描述，显而易见的，在不脱离本申请的精神和范围的情况下，可对其进行各种修改和组合。相应地，本说明书和附图仅仅是所附权利要求所界定的本申请的示例性说明，且视为已覆盖本申请范围内的任意和所有修改、变化、组合或等同物。显然，本领域的技术人员可以对本申请进行各种改动和变型而不脱离本申请的精神和范围。这样，倘若本申请的这些修改和变型属于本申请权利要求及其等同技术的范围之内，则本申请也意图包含这些改动和变型在内。

Claims (39)

  1. 一种人工智能AI服务信息获取方法，其特征在于，所述方法，包括：
    第一AI节点向终端设备发送测量配置信息；所述测量配置信息用于指示测量待测量AI节点的AI服务；
    所述第一AI节点接收来自所述终端设备的测量报告；所述测量报告包括候选AI服务信息；所述候选AI服务信息为根据所述测量配置信息对所述待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；所述第一条件携带在所述测量配置信息中。
  2. 根据权利要求1所述的方法，其特征在于，所述候选AI服务信息包括AI服务标识信息和性能信息，所述AI服务标识信息用于标识所述候选AI服务信息对应的AI服务，所述性能信息用于表征所述候选AI服务信息对应的AI服务的性能，所述方法还包括：
    所述第一AI节点根据所述AI服务标识信息和所述性能信息，分别将所述候选AI服务信息对应的一个或多个AI服务中，每个AI服务对应的候选AI服务信息中的目标AI服务信息添加至本地AI服务列表中，其中，所述目标AI服务信息为所述每个AI服务对应的候选AI服务信息中，AI服务的性能最优的候选AI服务信息。
  3. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    所述第一AI节点将所述候选AI服务信息中的目标AI服务信息添加至所述本地AI服务列表中，其中，所述目标AI服务信息为所述候选AI服务信息中，与本地AI服务列表中记录的AI服务信息不同的候选AI服务信息。
  4. 根据权利要求2或3所述的方法，其特征在于，所述候选AI服务信息包括AI节点标识信息，所述AI节点标识信息用于标识所述候选AI服务信息对应的待测量AI节点，所述方法还包括：
    所述第一AI节点根据所述本地AI服务列表中的一个或多个AI服务信息包括的AI节点标识信息，确定与所述一个或多个AI服务信息对应的AI节点；
    所述第一AI节点获取所述一个或多个AI服务信息对应的AI节点的神经网络模型；所述神经网络模型用于实现AI服务。
  5. 根据权利要求4所述的方法，其特征在于，所述候选AI服务信息包括模型标识信息，所述模型标识信息用于标识所述候选AI服务信息对应的神经网络模型，所述第一AI节点获取所述一个或多个AI服务信息对应的AI节点的神经网络模型，包括：
    所述第一AI节点从所述一个或多个AI服务信息对应的AI节点上获取与所述一个或多个AI服务信息包括的模型标识信息对应的神经网络模型。
  6. 根据权利要求4所述的方法，其特征在于，所述第一AI节点获取所述一个或多个AI服务信息对应的AI节点的神经网络模型，包括：
    所述第一AI节点从所述一个或多个AI服务信息对应的AI节点上获取预配置的神经网络模型。
  7. 根据权利要求1-6任一项所述的方法，其特征在于，所述测量配置信息包括所述待测量AI节点的标识信息和需要测量的所述AI服务的标识信息。
  8. 根据权利要求1-7任一项所述的方法，其特征在于，所述候选AI服务信息包括 以下至少一项信息：
    AI服务标识信息、AI节点标识信息、模型标识信息、模型参数信息、性能信息和时效信息；其中，所述AI服务标识信息用于标识所述候选AI服务信息对应的AI服务，所述AI节点标识信息用于标识所述候选AI服务信息对应的待测量AI节点，所述模型标识信息用于标识所述候选AI服务信息对应的神经网络模型，所述模型参数信息为所述候选AI服务信息对应的神经网络模型的参数信息，所述性能信息用于表征所述候选AI服务信息对应的AI服务的性能，所述时效信息用于表征所述候选AI服务信息的有效时间。
  9. 根据权利要求1-8任一项所述的方法，其特征在于，所述第一AI节点接收来自所述终端设备的测量报告，包括：
    所述第一AI节点通过第二网络设备接收来自所述终端设备的测量报告，所述第二网络设备为与所述待测量AI节点中的第二AI节点对应的网络设备；其中，所述候选AI服务信息为与所述第二AI节点对应的候选AI服务信息。
  10. 根据权利要求9所述的方法，其特征在于，所述测量报告还包括所述第一AI节点的标识信息；所述第一AI节点的标识信息用于所述第二网络设备将所述测量报告发送给所述第一AI节点。
  11. 一种人工智能AI服务信息获取方法，其特征在于，所述方法，包括：
    第三AI节点接收来自终端设备的订阅信息；所述订阅信息用于请求在订阅区域内订阅满足第二条件的AI服务；
    所述第三AI节点向所述订阅区域内的一个或多个第四AI节点发送所述订阅信息，以使得所述第四AI节点根据所述订阅信息，向所述终端设备发送订阅结果；所述订阅结果包括目标AI服务信息，所述目标AI服务信息对应的AI服务满足所述第二条件。
  12. 根据权利要求11所述的方法，其特征在于，所述订阅信息由第一标识信息标识。
  13. 根据权利要求12所述的方法，其特征在于，所述第一标识信息由所述第三AI节点生成；
    或者，所述第一标识信息由所述终端设备生成。
  14. 根据权利要求12或13所述的方法，其特征在于，在所述第三AI节点接收来自终端设备的订阅信息之后，所述方法还包括：
    所述第三AI节点确定所述终端设备当前位于第一通知区域，将所述第一标识信息与第二标识信息相关联；所述第二标识信息用于标识所述第一通知区域；
    所述第三AI节点将与所述第一标识信息相关联的第二标识信息发送给所述一个或多个第四AI节点。
  15. 根据权利要求14所述的方法，其特征在于，所述第一通知区域是根据小区和/或波束设置的。
  16. 根据权利要求11-15任一项所述的方法，其特征在于，所述订阅信息包括以下至少一项信息：
    用于标识请求订阅的AI服务的AI服务标识信息，用于标识所述终端设备的用户标识信息，用于表征所述第二条件的服务等级信息，用于表征所述订阅区域的订阅区 域信息，模型标识信息，模型参数信息和订阅有效期信息；其中，所述模型标识信息用于标识所述订阅信息请求订阅的AI服务对应的神经网络模型，所述模型参数信息为所述订阅信息请求订阅的AI服务对应的神经网络模型的参数信息，所述订阅有效期信息用于表征所述订阅信息的有效时间。
  17. 根据权利要求16所述的方法，其特征在于，所述订阅区域信息由所述终端设备确定；
    或者，所述订阅区域信息由所述第三AI节点确定。
  18. 一种人工智能AI服务信息获取方法，其特征在于，所述方法，包括：
    终端设备向第三AI节点发送订阅信息；所述订阅信息用于请求在订阅区域内订阅满足第二条件的AI服务；
    所述终端设备接收来自所述订阅区域内的一个或多个第四AI节点的订阅结果；所述订阅结果包括目标AI服务信息，所述目标AI服务信息对应的AI服务满足所述第二条件。
  19. 根据权利要求18所述的方法，其特征在于，所述订阅信息由第一标识信息标识。
  20. 根据权利要求19所述的方法，其特征在于，所述第一标识信息由所述第三AI节点生成；
    或者，所述第一标识信息由所述终端设备生成。
  21. 根据权利要求19或20所述的方法，其特征在于，在所述终端设备向第三AI节点发送订阅信息之后，所述方法还包括：
    所述终端设备获取与所述第一标识信息关联的第二标识信息；所述第二标识信息用于标识所述终端设备当前位于的第一通知区域。
  22. 根据权利要求21所述的方法，其特征在于，所述方法还包括：
    所述终端设备获取第三标识信息；所述第三标识信息用于标识所述终端设备当前位于的第二通知区域；
    所述终端设备确定所述第三标识信息与所述第二标识信息不一致，向所述一个或多个第四AI节点发送更新请求，所述更新请求用于请求将所述第一标识信息与所述第三标识信息相关联。
  23. 根据权利要求22所述的方法，其特征在于，所述终端设备获取第三标识信息包括：
    所述终端设备接收来自第三网络设备的第三标识信息；所述第三网络设备位于所述第二通知区域中。
  24. 根据权利要求21-23任一项所述的方法，其特征在于，所述第一通知区域是根据小区和/或波束设置的。
  25. 根据权利要求18-24任一项所述的方法，其特征在于，所述订阅信息包括以下至少一项信息：
    用于标识请求订阅的AI服务的AI服务标识信息，用于标识所述终端设备的用户标识信息，用于表征所述第二条件的服务等级信息，用于表征所述订阅区域的订阅区域信息，模型标识信息，模型参数信息和订阅有效期信息；其中，所述模型标识信息 用于标识所述订阅信息请求订阅的AI服务对应的神经网络模型，所述模型参数信息为所述订阅信息请求订阅的AI服务对应的神经网络模型的参数信息，所述订阅有效期信息用于表征所述订阅信息的有效时间。
  26. 根据权利要求25所述的方法，其特征在于，所述订阅区域信息由所述终端设备确定；
    或者，所述订阅区域信息由所述第三AI节点确定。
  27. 一种AI服务信息获取方法，其特征在于，所述方法包括：
    终端设备接收来自第一AI节点的测量配置信息；所述测量配置信息用于指示测量待测量AI节点的AI服务；
    所述终端设备向所述第一AI节点发送测量报告；所述测量报告包括候选AI服务信息；所述候选AI服务信息为根据所述测量配置信息对所述待测量AI节点进行测量得到的一个或多个AI服务信息中满足第一条件的AI服务信息；第一条件携带在所述测量配置信息中。
  28. 根据权利要求27所述的方法，其特征在于，所述测量配置信息包括所述待测量AI节点的标识信息和需要测量的所述AI服务的标识信息。
  29. 根据权利要求27或28所述的方法，其特征在于，所述候选AI服务信息包括以下至少一项信息：
    AI服务标识信息、AI节点标识信息、模型标识信息、模型参数信息、性能信息和时效信息；其中，所述AI服务标识信息用于标识所述候选AI服务信息对应的AI服务，所述AI节点标识信息用于标识所述候选AI服务信息对应的待测量AI节点，所述模型标识信息用于标识所述候选AI服务信息对应的神经网络模型，所述模型参数信息为所述候选AI服务信息对应的神经网络模型的参数信息，所述性能信息用于表征所述候选AI服务信息对应的AI服务的性能，所述时效信息用于表征所述候选AI服务信息的有效时间。
  30. 根据权利要求27-29任一项所述的方法，其特征在于，所述终端设备向所述第一AI节点发送测量报告，包括：
    所述终端设备通过第二网络设备向所述第一AI节点发送所述测量报告，所述第二网络设备为与所述待测量AI节点中的第二AI节点对应的网络设备；其中，所述候选AI服务信息为与所述第二AI节点对应的候选AI服务信息。
  31. 根据权利要求30所述的方法，其特征在于，所述测量报告还包括所述第一AI节点的标识信息；所述第一AI节点的标识信息用于所述第二网络设备将所述测量报告发送给所述第一AI节点。
  32. 一种通信装置，其特征在于，包括：处理器以及存储器，所述存储器用于存储计算机执行指令，所述处理器用于执行所述存储器存储的所述指令；当所述指令被所述处理器运行时，使得所述通信装置执行权利要求1-10中任一项所述的方法，或者，使得所述通信装置执行权利要求11-17中任一项所述的方法，或者，使得所述通信装置实现权利要求18-26中任一项所述的方法，或者，使得所述通信装置实现权利要求27-31中任一项所述的方法。
  33. 一种通信装置，其特征在于，所述通信装置包括用于执行权利要求1-10中任一 项所涉及的各个步骤的单元，或者，所述通信装置包括用于执行权利要求11-17中任一项所涉及的各个步骤的单元，或者，所述通信装置包括用于执行权利要求18-26中任一项所涉及的各个步骤的单元，或者，所述通信装置包括用于执行权利要求27-31中任一项所涉及的各个步骤的单元。
  34. 一种通信装置，其特征在于，包括：处理器和接口电路，所述接口电路用于与所述通信装置之外的模块通信；所述处理器用于运行计算机程序或指令以执行如权利要求1-10中任一项所述的方法；或者，所述处理器用于运行计算机程序或指令以执行如权利要求11-17中任一项所述的方法；或者，所述处理器用于运行计算机程序或指令以执行如权利要求18-26中任一项所述的方法；或者，所述处理器用于运行计算机程序或指令以执行如权利要求27-31中任一项所述的方法。
  35. 一种通信装置，其特征在于，包括：处理器；所述处理器用于与存储器耦合，并读取所述存储器中的指令之后，根据所述指令执行如权利要求1-10中任一项所述的方法；或者，执行如权利要求11-17中任一项所述的方法；或者，执行如权利要求18-26中任一项所述的方法；或者，执行如权利要求27-31中任一项所述的方法。
  36. 一种计算机程序产品，其特征在于，所述计算机程序产品包含指令，当所述计算机程序产品在通信装置上运行时，使得所述通信装置执行如权利要求1-10中任一项所述的方法；或者，使得所述通信装置执行如权利要求11-17中任一项所述的方法；或者，使得所述通信装置执行如权利要求18-26中任一项所述的方法；或者，使得所述通信装置执行如权利要求27-31中任一项所述的方法。
  37. 一种计算机可读存储介质，其特征在于，其上存储有计算机程序，当所述计算机程序被计算机执行时使得所述计算机执行权利要求1-10中任一项所述的方法，或者，使得所述计算机执行权利要求11-17中任一项所述的方法，或者，使得所述计算机执行权利要求18-26中任一项所述的方法，或者，使得所述计算机执行权利要求27-31中任一项所述的方法。
  38. 一种通信系统，其特征在于，包括：用于执行权利要求11-17中任一项所述的方法的通信装置；以及用于执行权利要求18-26中任一项所述的方法的通信装置。
  39. 一种通信系统，其特征在于，包括：用于执行权利要求1-10中任一项所述的方法的通信装置；以及用于执行权利要求27-31中任一项所述的方法的通信装置。
PCT/CN2023/085246 2022-04-01 2023-03-30 一种ai服务信息获取方法、装置及系统 Ceased WO2023186048A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP23778413.7A EP4492856A4 (en) 2022-04-01 2023-03-30 AI SERVICE INFORMATION ACQUISITION METHOD, APPARATUS AND SYSTEM
US18/901,120 US20250024276A1 (en) 2022-04-01 2024-09-30 Ai service information obtaining method, apparatus, and system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202210349458.2A CN116939687A (zh) 2022-04-01 2022-04-01 一种ai服务信息获取方法、装置及系统
CN202210349458.2 2022-04-01

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/901,120 Continuation US20250024276A1 (en) 2022-04-01 2024-09-30 Ai service information obtaining method, apparatus, and system

Publications (1)

Publication Number Publication Date
WO2023186048A1 true WO2023186048A1 (zh) 2023-10-05

Family

ID=88199462

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/085246 Ceased WO2023186048A1 (zh) 2022-04-01 2023-03-30 一种ai服务信息获取方法、装置及系统

Country Status (4)

Country Link
US (1) US20250024276A1 (zh)
EP (1) EP4492856A4 (zh)
CN (1) CN116939687A (zh)
WO (1) WO2023186048A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US12379835B2 (en) 2023-06-13 2025-08-05 Monday.com Ltd. Digital processing systems and methods for enhanced data representation
US12619833B2 (en) 2023-08-14 2026-05-05 Monday.com Ltd. Digital processing systems and methods for implementing and managing artificial intelligence functionalities in applications
US12639402B2 (en) 2023-08-14 2026-05-26 Monday.com Ltd. Ai-driven integration system for enhanced SaaS platform management and cross-platform synchronization
US12694636B2 (en) 2024-12-13 2026-07-28 Monday.com Ltd. Systems and methods for configuring and/or participating in virtual events

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10381006B1 (en) * 2018-11-26 2019-08-13 Accenture Global Solutions Limited Dialog management system for using multiple artificial intelligence service providers
US20210365804A1 (en) * 2021-06-11 2021-11-25 Intel Corporation Dynamic ai model transfer reconfiguration to minimize performance, accuracy and latency disruptions
CN114245404A (zh) * 2021-12-14 2022-03-25 上海优咔网络科技有限公司 一种人工智能边缘计算切换方法

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11240690B2 (en) * 2019-05-24 2022-02-01 Parallel Wireless, Inc. Streaming media quality of experience prediction for network slice selection in 5G networks
US11696119B2 (en) * 2019-12-16 2023-07-04 Qualcomm Incorporated Neural network configuration for wireless communication system assistance

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10381006B1 (en) * 2018-11-26 2019-08-13 Accenture Global Solutions Limited Dialog management system for using multiple artificial intelligence service providers
US20210365804A1 (en) * 2021-06-11 2021-11-25 Intel Corporation Dynamic ai model transfer reconfiguration to minimize performance, accuracy and latency disruptions
CN114245404A (zh) * 2021-12-14 2022-03-25 上海优咔网络科技有限公司 一种人工智能边缘计算切换方法

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
See also references of EP4492856A4
ZTE, CHINA UNICOM, CMCC: "Solution to AI based UE Trajectory prediction", 3GPP TSG-RAN WG3 #113-E, R3-213759, 6 August 2021 (2021-08-06), XP052035530 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US12379835B2 (en) 2023-06-13 2025-08-05 Monday.com Ltd. Digital processing systems and methods for enhanced data representation
US12430825B2 (en) 2023-06-13 2025-09-30 Monday.com Ltd. Digital processing systems and methods for enhanced data representation
US12573105B2 (en) 2023-06-13 2026-03-10 Monday.com Ltd. Digital processing systems and methods for enhanced data representation
US12586268B2 (en) 2023-06-13 2026-03-24 Monday.com Ltd. Digital processing systems and methods for enhanced data representation
US12619833B2 (en) 2023-08-14 2026-05-05 Monday.com Ltd. Digital processing systems and methods for implementing and managing artificial intelligence functionalities in applications
US12639402B2 (en) 2023-08-14 2026-05-26 Monday.com Ltd. Ai-driven integration system for enhanced SaaS platform management and cross-platform synchronization
US12645759B2 (en) 2023-08-14 2026-06-02 Monday.com Ltd. Intent-based interactions in software platforms
US12694636B2 (en) 2024-12-13 2026-07-28 Monday.com Ltd. Systems and methods for configuring and/or participating in virtual events

Also Published As

Publication number Publication date
EP4492856A1 (en) 2025-01-15
US20250024276A1 (en) 2025-01-16
EP4492856A4 (en) 2025-07-16
CN116939687A (zh) 2023-10-24

Similar Documents

Publication Publication Date Title
US20230232213A1 (en) Information transmission methods and apparatuses, and communication devices and storage medium
WO2022252859A1 (zh) 一种算力资源调度的方法以及相关装置
CN110971421B (zh) 订阅更新方法、设备及系统
EP4492856A1 (en) Ai service information acquisition method, apparatus and system
EP4542953A1 (en) Method and apparatus for determining artificial intelligence (ai) model
CN114303347A (zh) 通信网络中与机器学习相关的方法、装置和机器可读介质
CN117768875A (zh) 一种通信方法及装置
WO2022165721A1 (zh) Ran域的模型共享方法及装置
US20230388908A1 (en) Network slice-based communication method and apparatus
US12549936B2 (en) Systems and methods to optimize registration and session establishment in a wireless network
WO2022082356A1 (zh) 一种通信方法及装置
WO2023056809A1 (zh) 定位方法、装置及系统
CN110324152A (zh) 策略和计费控制规则获取方法、装置及系统
WO2019120073A1 (zh) 数据传输方法、设备及系统
WO2023169423A1 (zh) 拆分推理方法及装置
CN120186637A (zh) 一种通信方法及装置
WO2024027803A1 (zh) 通信方法、装置及系统
WO2021134614A1 (zh) 通信方法、设备及系统
WO2023272670A1 (zh) 一种网络连接的方法、装置及系统
CN105578559A (zh) 一种超密集组网下的接入方法、设备及系统
WO2025227977A1 (zh) 通信方法及装置
WO2025124528A1 (zh) 一种通信方法及通信装置
WO2016112705A1 (zh) 一种支持多种无线网络接入的方法、装置及存储介质
JP2024073081A (ja) 配信サーバ、配信サーバの制御方法、及び配信サーバの制御プログラム
WO2023005440A1 (zh) 通信方法、通信装置及通信系统

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23778413

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 2023778413

Country of ref document: EP

ENP Entry into the national phase

Ref document number: 2023778413

Country of ref document: EP

Effective date: 20241011

NENP Non-entry into the national phase

Ref country code: DE