WO2024007156A1 - 一种通信方法和装置 - Google Patents
一种通信方法和装置 Download PDFInfo
- Publication number
- WO2024007156A1 WO2024007156A1 PCT/CN2022/103941 CN2022103941W WO2024007156A1 WO 2024007156 A1 WO2024007156 A1 WO 2024007156A1 CN 2022103941 W CN2022103941 W CN 2022103941W WO 2024007156 A1 WO2024007156 A1 WO 2024007156A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- network element
- data
- target network
- characteristic data
- event
- Prior art date
Links
- 238000000034 method Methods 0.000 title claims abstract description 135
- 238000004891 communication Methods 0.000 title claims abstract description 120
- 238000000605 extraction Methods 0.000 claims description 107
- 230000004044 response Effects 0.000 claims description 52
- 238000012545 processing Methods 0.000 claims description 33
- 238000004590 computer program Methods 0.000 claims description 19
- 238000003860 storage Methods 0.000 claims description 12
- 230000006870 function Effects 0.000 description 52
- 238000012549 training Methods 0.000 description 47
- 238000013480 data collection Methods 0.000 description 29
- 238000013473 artificial intelligence Methods 0.000 description 24
- 238000010586 diagram Methods 0.000 description 22
- 230000008569 process Effects 0.000 description 18
- 238000013145 classification model Methods 0.000 description 14
- 238000007726 management method Methods 0.000 description 11
- 239000000284 extract Substances 0.000 description 10
- 230000005540 biological transmission Effects 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 6
- 238000013461 design Methods 0.000 description 5
- 238000009826 distribution Methods 0.000 description 5
- 230000009471 action Effects 0.000 description 4
- 230000008859 change Effects 0.000 description 4
- 238000007405 data analysis Methods 0.000 description 4
- 238000012986 modification Methods 0.000 description 4
- 230000004048 modification Effects 0.000 description 4
- 230000000694 effects Effects 0.000 description 3
- 238000011022 operating instruction Methods 0.000 description 3
- 230000007958 sleep Effects 0.000 description 3
- 238000011161 development Methods 0.000 description 2
- 230000018109 developmental process Effects 0.000 description 2
- 238000004134 energy conservation Methods 0.000 description 2
- 238000011156 evaluation Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- RYGMFSIKBFXOCR-UHFFFAOYSA-N Copper Chemical compound [Cu] RYGMFSIKBFXOCR-UHFFFAOYSA-N 0.000 description 1
- 101150119040 Nsmf gene Proteins 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 238000013475 authorization Methods 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008878 coupling Effects 0.000 description 1
- 238000010168 coupling process Methods 0.000 description 1
- 238000005859 coupling reaction Methods 0.000 description 1
- 238000012517 data analytics Methods 0.000 description 1
- 238000013144 data compression Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 230000005611 electricity Effects 0.000 description 1
- 230000014509 gene expression Effects 0.000 description 1
- 230000006266 hibernation Effects 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 239000007788 liquid Substances 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 230000002093 peripheral effect Effects 0.000 description 1
- 230000008054 signal transmission Effects 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
- 238000012546 transfer Methods 0.000 description 1
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 1
Images
Classifications
-
- 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
- H04L41/14—Network analysis or design
Definitions
- the present application relates to the field of communication technology, and in particular, to a communication method and device.
- AI artificial intelligence
- 5G fifth generation mobile communication
- release16 core network
- NWDAF network data analysis functionality
- the application of AI function usually includes the two processes of AI model training and reasoning.
- the purpose of AI model training is to obtain One or more AI models.
- the purpose of AI model reasoning is to apply the trained AI model to new data to obtain prediction results or evaluation results.
- network elements at different layers in the network have different data and information resources.
- Network elements in the upper layer have global data and information, while network elements in the lower layer have local data and information.
- the existing The AI technical solution only supports a single network element (or node) to use its own data and information resources to train the AI model (i.e. functional model), and cannot fully utilize the data and information resources of the upper-layer network element (or node). Therefore, , resulting in lower accuracy of the trained AI model.
- different network elements (or nodes) directly provide and transmit their own original data, data security cannot be guaranteed.
- a communication method and device that interacts between different network elements to obtain a highly accurate functional model while ensuring data security.
- this application provides a communication method.
- the method can be executed by the first network element or by a component of the first network element (such as a processor, a chip, or a chip system, etc.).
- This application does not cover this aspect. Make specific limitations.
- the method may specifically include the following steps: the first network element determines the first characteristic data of the target network event; the first network element receives first information from the second network element, the first information includes the information of the target network event, The second characteristic data of the target network event; the first network element executes the target network event according to a first functional model.
- the first functional model is the first network element based on the first characteristic data and the second characteristic data. owned.
- the first network element determines the first characteristic data (ie, local characteristic data) of the target network event. Then, the first network element receives the first information from the second network element, and the first information includes The information of the target network event and the second characteristic data of the target network event. Further, the first network element can obtain the first functional model based on the first characteristic data and the second characteristic data provided by the second network element.
- the first network element can effectively utilize the second characteristic data of the target network event provided by the second network element to assist the characteristic data of the local target network event while ensuring the security of the data of the second network element.
- a first functional model with higher accuracy is obtained, which can also improve the accuracy of executing the target network event based on the first functional model.
- the method before the first network element receives the first information from the second network element, the method further includes: the first network element sends a first request message to the second network element.
- the first request message Used to request the second characteristic data of the target network event.
- the first network element that needs to obtain the first functional model can effectively request the required second characteristic data from the second network element for the target network event.
- the first request message includes information about the target network event; wherein the information about the target network event includes identification information of the target network event, and also includes characteristic data of the target network event.
- the target network event includes identification information of the target network event, and also includes characteristic data of the target network event.
- the second network element can accurately know each information of the target network event according to the first request message, as well as the requirement information of the first network element requesting the characteristic data of the target network event.
- the method further includes: the first network element receiving a first response message from the second network element, the first response message being used to notify the first network element of the second network element The second characteristic data of the target network event.
- the first network element receives the first response message from the second network element. It can be seen that the second network element accepts the request of the first network element and determines that the second network element can provide the third response message of the target network event. Two characteristic data.
- the first information also includes the first response message.
- the first network element determines the first characteristic data of the target network event, including: the first network element obtains the first data set of the target network event; the first network element uses the first A feature extraction model performs feature extraction on the first data set to obtain the first feature data.
- the first feature extraction model is used by the first network element to extract feature data of the target network event.
- the first network element uses the first feature extraction model (ie, local feature extraction model) to effectively extract features from the first data set (ie, local data set) of the target network event to obtain the first feature data.
- first feature extraction model ie, local feature extraction model
- the first information further includes a first indication, and the first indication is used to indicate whether the first network element returns the adjustment amount of the second characteristic data.
- the second network element can also inform the first network element whether to update the second feature extraction model through the first instruction, and then instruct the first network element to return the adjustment amount of the second feature data, so that the first network element
- the second network element can use the adjustment amount of the second feature data to effectively adjust the second feature extraction model, thereby improving the accuracy of the second feature extraction model.
- the method further includes: the first network element obtains sample data of the target network event, where the sample data includes characteristic data samples and the true value of the target network event; the first network element Input the characteristic data sample into the first functional model to obtain the output value of the target network event; the first network element obtains the second characteristic data based on the output value of the target network event and the true value of the target network event
- the adjustment amount of the second feature data is used to adjust the output layer of the second feature extraction model of the second network element.
- the first network element can effectively obtain the adjustment amount of the second feature data, so that the second network element can adjust the second feature extraction model of the second network element according to the amount of the second feature data.
- the output layer of the second feature extraction model can also be adjusted to other layers of the second feature extraction model. There is no need for the first network element to feed back the adjustment amounts of other layers of the second feature extraction model to the second network element multiple times, thereby saving the need to update the second feature extraction model. Resource overhead and delay of the second feature extraction model.
- the method when the first instruction is used to instruct the first network element to return the adjustment amount of the second characteristic data, the method further includes: the first network element transmits the adjustment amount to the first network element according to the first instruction.
- the second network element sends the adjustment amount of the second characteristic data.
- the first network element when the second network element feeds back the first indication in the first information to the first network element to instruct the first network element to return the adjustment amount of the second characteristic data, the first network element
- the adjustment amount of the second feature data can also be effectively fed back to the second network element, so that the second network element can accurately and effectively adjust the second feature extraction model based on the adjustment amount of the second feature data.
- this application provides a communication method, which can be executed by the second network element or by a component of the second network element (such as a processor, a chip, or a chip system, etc.). This application does not cover this aspect. Make specific limitations.
- the method may specifically include the following steps: the second network element determines the second characteristic data of the target network event; the second network element sends first information to the first network element, the first information includes the information of the target network event and Second characteristic data of the target network event.
- the second network element determines the second characteristic data for the target network event, and then carries the information of the target network event and the second characteristic data in the first information and sends it to the first network element, thereby Under ensuring the data security of the second network element, the first network element can use the second characteristic data provided by the second network element to obtain a first functional model with higher accuracy.
- the method before the second network element determines the second characteristic data of the target network event, the method further includes: the second network element receives a first request message from the first network element, and the first request message Used to request second characteristic data of the target network event.
- the second network element receives the first request message sent by the first network element, so that the target network event can be determined, and the second characteristic data of the target network event can be effectively provided.
- the first request message includes information about the target network event, where the information about the target network event includes identification information of the target network event, and also includes characteristic data of the target network event.
- the target network event includes identification information of the target network event, and also includes characteristic data of the target network event.
- the second network element can accurately know each information of the target network event according to the first request message, as well as the requirement information of the first network element requesting the characteristic data of the target network event.
- the method further includes: the second network element sending a first response message to the first network element, the first response message being used to notify the first network element of the second network element Second characteristic data of the target network event is provided.
- the second network element can effectively inform the first network element that the second characteristic data of the target network event can be provided, so that the first network element determines that the second characteristic data provided by the second network element can be utilized.
- the data were derived from the first functional model.
- the first information also includes the first response message.
- the second network element determines the second characteristic data of the target network event, including: the second network element obtains the second data set of the target network event; the second network element uses the second A feature extraction model performs feature extraction on the second data set to obtain the second feature data.
- the second feature extraction model is used by the second network element to extract feature data of the target network event.
- the second network element can use the second feature extraction model (ie, the local feature extraction model) to perform feature extraction on the second data set of the target network event, so as to effectively obtain the feature data of the target network event.
- the second feature extraction model ie, the local feature extraction model
- the first information further includes a first indication, and the first indication is used to indicate whether the first network element returns the adjustment amount of the second characteristic data.
- the method when the first indication is used to instruct the first network element to return the adjustment amount of the second characteristic data, the method further includes: the second network element receives from the first network element The adjustment amount of the second characteristic data.
- the second network element when the first instruction is used to instruct the first network element to return the adjustment amount of the second characteristic data, the second network element can also effectively receive the adjustment amount of the second characteristic data from the first network element. , and then the second feature extraction model can be optimized.
- the method further includes: the second network element adjusting the output layer of the second feature extraction model of the second network element according to the adjustment amount of the second feature data.
- the first network element only feeds back to the second network element the adjustment amount used to adjust the output layer of the second feature extraction model.
- the second network element can effectively adjust the output layer of the second feature data according to the adjustment amount.
- the output layer of the second feature extraction model of the second network element is adjusted, and other layers of the second feature extraction model can also be adjusted to improve the accuracy of the second feature extraction model.
- the present application provides a communication device, which can be applied to the first network element and has the function of implementing the above-mentioned first aspect or any possible implementation manner of the above-mentioned first aspect.
- the device may include: a communication unit (the communication unit includes a sending unit and/or a receiving unit) and a processing unit.
- the present application provides a communication device, which can be applied to a second network element and has the function of implementing any possible implementation manner of the above-mentioned first aspect or the above-mentioned second aspect.
- the device may include: a communication unit (the communication unit includes a sending unit and/or a receiving unit) and a processing unit.
- the present application also provides a communication device, which can be applied to the first network element.
- the communication device includes a processor for implementing the above-mentioned first aspect or any possible implementation of the above-mentioned first aspect. mode function.
- the communication device further includes a transceiver for implementing the communication function of the communication device.
- the present application also provides a communication device that can be applied to a second network element.
- the communication device includes a processor for implementing the above second aspect or any possible design of the above second aspect. function of the method in .
- the communication device further includes a transceiver for implementing the communication function of the communication device.
- the present application also provides a communication system, which includes a first network element for performing the method provided by the first aspect, and a second network element used for performing the method provided by the second aspect.
- embodiments of the present application also provide a computer storage medium, which stores a software program.
- the software program When the software program is read and executed by one or more processors, it can implement the first aspect or any one of the above. methods provided by three possible implementations, or implement the method provided by the above second aspect or any one of the possible implementations.
- embodiments of the present application further provide a computer program product containing instructions that, when run on a computer, cause the method provided in the above first aspect or any of the possible implementations to be executed, or cause the above third aspect to be executed.
- the method provided by the two aspects or any one of the possible implementations is performed.
- embodiments of the present application further provide a chip system.
- the chip system includes a processor for supporting the first network element to implement the functions involved in the above-mentioned first aspect; or for supporting the second network element to implement the above-mentioned functions.
- the chip system further includes a memory, and the memory is used to save necessary program instructions and data executed by the loading device.
- the chip system may be composed of chips, or may include chips and other discrete devices.
- Figure 1 is a schematic diagram of a reference architecture for data collection and reporting
- Figure 2 is a schematic diagram of a network architecture suitable for the communication method provided by the embodiment of the present application.
- Figure 3 is a schematic diagram of a communication system to which the communication method provided by the embodiment of the present application is applicable;
- Figure 4A is a schematic structural diagram of a communication method provided by an embodiment of the present application.
- Figure 4B is a schematic flow chart of a communication method provided by an embodiment of the present application.
- Figure 5A is a schematic structural diagram of the first embodiment provided by the embodiment of the present application.
- Figure 5B is a schematic flow chart of the first embodiment provided by the embodiment of this application.
- Figure 6A is a schematic structural diagram of the second embodiment provided by the embodiment of the present application.
- Figure 6B is a schematic flow chart of the second embodiment provided by the embodiment of the present application.
- Figure 7A is a schematic structural diagram of a third embodiment provided by the embodiment of the present application.
- Figure 7B is a schematic flow diagram of the third embodiment provided by the embodiment of the present application.
- Figure 8 is a schematic structural diagram of a communication device provided by an embodiment of the present application.
- Figure 9 is a schematic structural diagram of a communication device provided by an embodiment of the present application.
- FIG. 10 is a schematic structural diagram of a chip device provided by an embodiment of the present application.
- the embodiments of the present application provide a communication method and device.
- the method and the device are based on the same or similar technical concepts. Since the principles of the method and the device in solving problems are similar, the implementation of the method and the device can be referred to each other and repeated. No further details will be given.
- AI artificial intelligence
- NWDAF network data analysis function
- the application of AI functions usually includes two processes: AI model training and inference.
- AI model training is to obtain one or more AI models.
- AI model inference is to apply the trained AI model to new data. To obtain prediction results or evaluation results.
- FIG 1 is a schematic diagram of the reference architecture for data collection and reporting formulated in the R17 protocol.
- the UE side data can be the quality of experience (QoE) of the service in the UE or other data, which is not limited here.
- the UE side data may be data obtained by the UE.
- UE side data can also be called UE side data, UE data or UE data (UE data).
- Each network element (or node) in the architecture shown in Figure 1 is briefly described below.
- Data collection AF can receive configuration information from an application service provider (ASP) for configuring data collection and reporting functions.
- ASP application service provider
- the network side can interact with DCAF through the service provisioning AF (Provisioning AF) in ASP, and enable DCAF's data collection and reporting functions.
- DCAF can generate data collection and reporting configuration information, and send the data collection and reporting configuration information to the data collection client side, so that the data collection client collects data and reports data based on the received data collection and reporting configuration information.
- Collection reports; DCAF can also receive data collection reports reported by data collection clients.
- the data collection client can include: direct data collection client (DDCC), indirect data collection client (IDCC) or application server (AS).
- DCAF can process the data in the data collection report and send the processed UE-side data to the subscribing network element or third-party entity according to the subscription request of other network elements or third-party entities.
- DDCC is the data collection functional entity in the UE.
- DDCC can receive data collection and reporting configuration information from DCAF through the R2 interface, and perform data collection and reporting based on the data collection and reporting configuration information.
- the UE application Application, APP
- the application service provider (ASP) can receive UE-side data from UE applications through the R8 interface.
- the R8 interface is a pure application layer implementation. It can be considered that there is an indirect data collection client (IDCC) in the ASP.
- IDCC indirect data collection client
- the service activation AF sends the data collection report to DCAF through the R1 interface.
- the IDCC sends the UE side data collected through the application layer to DCAF through the R3 interface. Prior to this, IDCC will also receive data collection and reporting configuration information from DCAF through the R3 interface.
- the application server (AS) can receive data collection and reporting configuration information from DCAF through the R4 interface, then collect data based on the data collection and reporting configuration information, obtain a data collection report, and send the data collection report to DCAF through the R4 interface.
- This data collection report includes access log information of the application server, etc.
- NWDAF can subscribe to and receive UE-side data from DCAF through the R5 interface, and then analyze the UE-side data.
- ASP may also obtain the data collected by DCAF.
- the event subscription AF Event Consumer AF
- the event subscription AF in ASP can subscribe to the collected data from DCAF through the R6 interface and receive the subscribed data information from DCAF.
- this application provides a communication method, which method includes: a first network element determines first characteristic data of a target network event; the first network element receives first information from a second network element, and the first information includes the The information of the target network event and the second characteristic data of the target network event; the first network element executes the target network event according to the first functional model, and the first functional model is the first network element according to the first characteristic data and the second characteristic data.
- the first network element can use the data and information of the second network element to obtain an accurate first functional model while ensuring the data privacy and security of the second network element.
- the first network element is based on the third network element.
- a functional model executes the target network event, which can also improve execution accuracy.
- Figure 2 shows a schematic diagram of a network architecture to which the communication method provided by the embodiment of the present application can be applied.
- the network architecture may include an access network and a core network. Terminal equipment is connected to the data network (DN) through the access network and core network.
- DN data network
- the terminal device can be a user equipment (UE), a mobile station, a mobile terminal, an application client, etc.
- Terminal devices can be widely used in various scenarios, such as device-to-device (D2D), vehicle to everything (V2X) communication, machine-type communication (MTC), and the Internet of Things (internet of things, IOT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grid, smart furniture, smart office, smart wear, smart transportation, smart city, etc.
- Terminal devices can be mobile phones, tablets, computers with wireless transceiver functions, wearable devices, vehicles, urban air vehicles (such as drones, helicopters, etc.), ships, robots, robotic arms, smart home devices, etc.
- a UE is used as an example of a terminal device for description. The UE appearing anywhere subsequently can also be replaced with a terminal device or other examples of a terminal device.
- the access network is used to implement access-related functions. It can provide network access functions for authorized users in a specific area, and can determine transmission links of different qualities to transmit user data based on user levels, business needs, etc.
- the access network forwards control signals and user data between the UE and the core network.
- the access network may include access network equipment, which may be equipment that provides access for UEs, and may include wireless access network (radio access network, RAN) equipment and wired access network equipment.
- RAN equipment is mainly responsible for wireless resource management, quality of service (QoS) management, data compression and encryption on the air interface side.
- RAN equipment can include various forms of base stations, such as macro base stations, micro base stations (also called small stations), relay stations, access points, balloon stations, etc.
- the names of equipment with base station functions may be different.
- RAN next-generation Node base station
- gNB next-generation Node base station
- LTE long term evolution
- eNB evolved NodeB
- the access network equipment and UE can be fixed-positioned or mobile. Access network equipment and UE 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 satellites in the sky.
- the embodiments of this application do not limit the application scenarios of access network equipment and UE.
- the core network is responsible for maintaining the subscription data of the mobile network and providing functions such as session management, mobility management, policy management, and security authentication for UEs.
- the core network includes but is not limited to one or more of the following network elements: application function (AF) network element, unified data management (UDM) network element, unified data repository (UDR) network element , policy control function (PCF) network element, session management function (SMF) network element, access and mobility management function (AMF) network element, network storage function ( network repository function (NRF) network element, authentication server function (AUSF) network element, network exposure function (NEF) network element, user plane function (UPF) network element, network Data analysis function network element (Network Data Analytics Function, NWDAF) network element.
- AF application function
- UDM unified data management
- UDR unified data repository
- PCF policy control function
- SMF session management function
- AMF access and mobility management function
- NRF network storage function
- AUSF authentication server function
- UPF user plane function
- AMF network element is mainly responsible for mobility management in mobile networks, such as user registration network, user switching, etc.
- the SMF network element is mainly responsible for session management in mobile networks, such as session establishment, modification, release, etc.
- the UPF network element is mainly responsible for forwarding and receiving user data. It can receive user data from the data network and transmit it to the UE through the access network device. It can also receive user data from the UE through the access network device and forward it to the data network.
- UDM network elements include functions such as execution and management of contract data and user access authorization.
- the UDR network element includes access functions for executing contract data, policy data, application data and other types of data.
- NEF network elements are mainly used to support the opening of capabilities and events.
- the AF network element transmits the requirements of the application side to the network side.
- the PCF network element mainly supports providing a unified policy framework to control network behavior, provides policy rules to the control layer network functions, and is also responsible for obtaining user subscription information related to policy decisions.
- PCF network elements can provide policies to AMF network elements and SMF network elements.
- NRF network elements can be used to provide network element discovery functions, provide network element information corresponding to network element types based on requests from other network elements, and provide network element management services.
- the AUSF network element is responsible for authenticating the UE and verifying the legitimacy of the UE.
- NWDAF network elements provide functions such as network data collection and analysis based on technologies such as big data and artificial intelligence.
- DN on which a variety of services can be deployed, can provide data and/or voice services to UEs.
- AF network element UDM network element, UDR network element, PCF network element, SMF network element, AMF network element, NRF network element, AUSF network element, NEF network element, UPF network element and NWDAF network element
- AF network element UDM network element, UDR network element, PCF network element, SMF network element, AMF network element, NRF network element, AUSF network element, NEF network element, UPF network element and NWDAF network element
- Nausf, Nnef, Nnrf, Namf, Npcf, Nsmf, Nudm, Nudr, Naf, and Nnwdaf are the service interfaces provided by the above-mentioned AUSF, NEF, NRF, AMF, PCF, SMF, UDM, UDR, AF, and NWDAF respectively. Used to call corresponding service operations.
- N1, N2, N3, N4 and N6 are interface serial numbers. The meanings of these interface serial numbers are as follows:
- N1 The interface between AMF and UE, which can be used to transmit non-access stratum (NAS) signaling (such as QoS rules from AMF) to UE.
- N2 The interface between AMF and access network equipment, which can be used to transmit wireless bearer control information from the core network side to the access network equipment.
- N3 The interface between the access network equipment and the UPF. It is mainly used to transmit uplink and downlink user plane data between the access network equipment and the UPF.
- N4 The interface between SMF and UPF, which can be used to transfer information between the control plane and the user plane, including controlling the distribution of forwarding rules, QoS rules, traffic statistics rules, etc. for the user plane and reporting of user plane information.
- N6 The interface between UPF and DN, used to transmit uplink and downlink user data flows between UPF and DN.
- the above network elements or functions can be network elements in hardware devices, software functions running on dedicated hardware, or virtualization functions instantiated on a platform (for example, a cloud platform).
- a platform for example, a cloud platform.
- the above network element or function can be implemented by one device, can be implemented by multiple devices, or can be a functional module in one device, which is not specifically limited in the embodiments of this application.
- the first network element in the solution of this application may be a lower layer network element in the above-mentioned network, such as the above-mentioned application function AF network element and UE, which may have local local information and/or data.
- the second network element in the solution of this application It may be an upper-layer network element in the above-mentioned network architecture, such as CN and RAN equipment, and may have global information and/or data in the network.
- the embodiment of the present application also provides a schematic architectural diagram of a communication system to which the communication method is applicable.
- the communication system includes a base station, a core network, and multiple terminal devices UE.
- the base stations are all connected to core network devices, and the multiple base stations are respectively connected to multiple terminal devices.
- the embodiments of this application are applicable to both homogeneous and heterogeneous network scenarios.
- transmission points can be multi-point coordinated transmission between macro base stations and macro base stations, micro base stations and micro base stations, or macro base stations and micro base stations. , applicable to both frequency division duplexing (FDD)/time division duplexing (TDD) systems.
- FDD frequency division duplexing
- TDD time division duplexing
- Embodiments of the present invention are applicable to both low-frequency scenarios (sub 6G) and high-frequency scenarios (above 6G), terahertz, optical communications, etc.
- the first network element in the solution of this application may be each UE in Figure 3, and may have local information and/or data.
- the second network element in the solution of this application may be the base station in Figure 3, and may have local information and/or data. Global information and/or data.
- the first network element in the solution of this application may be each base station in Figure 3, and the second network element in the solution of this application may be the core network in Figure 3.
- this application does not place specific restrictions on the types and physical forms of the first network element and the second network element, as long as the second network element has more global and comprehensive data than the first network element.
- the embodiment of the present application provides a communication method, which can be applied to, but is not limited to, the network architecture shown in Figure 2 and the communication system shown in Figure 3, and the method can be executed by the network elements involved in this application. , or executed by the chip corresponding to the involved network element.
- the network element in this application can be a physical entity network element or a virtual network element. This application does not specifically limit the form of the involved network element.
- the first network element when the first network element is a network element that owns local data, and the second network element is a network element that owns global data, the first network element can also be the above-mentioned terminal device ( or a controller or chip corresponding to a terminal device), and the second network element may be the above-mentioned access network device (or a controller or chip corresponding to the access network device), such as a base station.
- the ordinal numbers such as “first” and “second” mentioned below are used to distinguish multiple objects for ease of description, and are not used to limit the order, timing, and sequence of multiple objects. Priority or importance. This application does not limit the specific forms of the first network element and the second network element in this embodiment.
- Figure 4A shows a structure for executing the communication method provided by the embodiment of the present application.
- Figure 4B shows the flow of a communication method provided by the embodiment of the present application. Referring to Figures 4A and 4B, the specific steps may include the following steps :
- the first network element sends a first request message to the second network element.
- the first request message is used to request the second characteristic data of the target network event.
- the second network element receives the first request message.
- the first network element is a lower-layer network element, or a network element that needs to obtain auxiliary data from the second network element to obtain the first functional model.
- the second network element is an upper-layer network element, or a network element that provides global characteristic data to the first network element to assist the first network element in obtaining a functional model.
- the first request message includes information about the target network event, where the information about the target network event includes identification information of the target network event, and a target that also includes characteristic data of the target network event.
- the information about the target network event includes identification information of the target network event, and a target that also includes characteristic data of the target network event.
- the target network event can be several network event types clearly defined in the protocol, such as load balancing, energy saving, etc.
- the target network event can also be a pre-agreed event between network elements in the network (such as a first network element and a second network element), as long as the second network element can identify the target after receiving the first request message. Network events are sufficient. Therefore, this application does not specifically limit the target network event.
- S402B The second network element sends the first response message to the first network element.
- the first network element receives the first response message.
- the second network element After receiving the first request message from the first network element, the second network element needs to determine whether to accept or reject the request, that is, to accept or reject the second characteristic data of the target network element event provided to the first network element. If it is determined that the request is accepted, the following step S403B is executed. If it is determined that the request is rejected, none of the following steps is executed.
- the second network element determines whether to accept or reject the request for several reasons including but not limited to:
- the second network element determines that the request can be accepted if the local data of the second network element includes data related to the target network event. If the local data of the second network element does not include data related to the target network event, the second network element determines that the request can be rejected.
- the second network element when the second network element itself allows other network elements to access or obtain local characteristic data, it can be determined to accept the above request. If the second network element itself does not allow other network elements to access or obtain local characteristic data, it can be determined that the second network element itself does not allow other network elements to access or obtain local characteristic data. Deny the above request. Or if the second network element can transmit the second characteristic data requested by the first network element, it is determined to accept the above request. If the second network element cannot transmit the second characteristic data requested by the first network element, then reject the above request.
- this application does not specifically limit the reason why the second network element accepts or rejects the first request message of the first network element.
- the first network element can obtain the following first functional model based on local characteristic data, or the first network element can send a request message to the second network element after a period of time. Send the above-mentioned first request message.
- the second network element sends a first response message to the first network element, and the first network element receives the first response message.
- the first response message is used to notify the first network element of the second network element.
- the element provides second characteristic data of the target network event. In this implementation manner, the second network element continues to execute the following step S403B.
- the second network element sends a first response message to the first network element, and the first network element receives the first response message.
- the first response message is used to notify the first network element of the second response message.
- the network element does not provide the second characteristic data of the target network event.
- the second network element does not continue to perform the following step S403B and any subsequent steps.
- S403B The second network element determines the second characteristic data of the target network event.
- the second network element determines the second feature data of the target network event, including: the second network element obtains a second data set of the target network event; the second network element uses the second feature extraction model, perform feature extraction on the second data set to obtain the second feature data, and the second feature extraction model is used by the second network element to extract feature data of the target network event.
- the first network element also determines first characteristic data of the target network event.
- the first network element may determine the first characteristic data of the target network event before the above step S403B, or simultaneously with the above step S403B, or simultaneously with the following step S404B, or after the following step S404B. implement. Therefore, compared with the above step S403B and the following step S404B, this application does not specifically limit the order in which the first network element determines the first characteristic data of the target network event.
- the first network element determines the first feature data of the target network event, including: the first network element obtains the first data set of the target network event; the first network element uses the first feature extraction model, perform feature extraction on the first data set to obtain the first feature data, and the first feature extraction model is used by the first network element to extract feature data of the target network event.
- the second network element sends first information to the first network element, where the first information includes information about the target network event and second characteristic data of the target network event.
- the first information further includes the above-mentioned first response message, which is used to notify the first network element that the second network element provides the second characteristic data of the target network event.
- the first information further includes a first indication, the first indication being used to indicate whether the first network element returns the adjustment amount of the second characteristic data.
- the first network element needs to change the adjustment amount of the second characteristic data after determining the adjustment amount of the second characteristic data.
- the adjustment amount is sent to the second network element. If the first instruction is used to instruct the first network element not to return the adjustment amount of the second characteristic data, the first network element does not need to send the adjustment amount of the second characteristic data to the second network element.
- the first indication is used to indicate that The first network element does not return the adjustment amount of the second characteristic data. If the second feature extraction model of the second network element is immature (ie, the accuracy is low), or the second network element allows the local feature extraction model to be changed, the first indication is used to instruct the first network element to return the The adjustment amount of the second characteristic data.
- the first network element executes the target network event according to the first functional model.
- the first functional model is obtained by the first network element based on the first characteristic data and the second characteristic data.
- the first network element performs feature union on the first feature data and the second feature data, and uses the feature data after feature combination as the input of the functional model to obtain the first functional model.
- the first network element can more accurately execute the target network event by using the first functional model.
- the first functional model may be but is not limited to: a prediction model, a regression model, a classification model, and a clustering model.
- the first functional model is not specifically limited in this application solution and can be obtained according to actual needs.
- S406B The first network element updates the first feature extraction model.
- step S406B when performing the above step S406B, the following steps may be included:
- the first network element determines the adjustment amount of the first characteristic data
- the first network element updates the first feature extraction model according to the adjustment amount of the first feature data.
- the first network element determines the adjustment amount of the first characteristic data, including: the first network element obtains sample data of the target network event, and the sample data includes the characteristic data sample of the first network element and the first network element.
- the real value of , the adjustment amount of the first characteristic data is obtained.
- the adjustment amount of the first feature data is used to adjust the output layer of the first feature extraction model of the first network element.
- the method when the first indication in the first information is used to instruct the first network element to return the adjustment amount of the second characteristic data, the method further includes: the first network element responds to the first indication according to the first indication. , sending the adjustment amount of the second characteristic data to the second network element.
- the first network element determines the adjustment amount of the second characteristic data, including: the first network element obtains sample data of the target network event, and the sample data includes the characteristic data sample and the true value of the target network event. ; The first network element inputs the characteristic data sample into the first functional model to obtain the output value of the target network event; the first network element obtains the second characteristic data according to the output value of the target network event and the true value of the target network event. Adjustment amount. The adjustment amount of the second feature data is used to adjust the output layer of the second feature extraction model of the second network element.
- the first network element sends a first request message to the second network element.
- the first request message is used to request the second characteristic data of the target network event.
- the first network element determines the target network event.
- the first network element can obtain the first functional model according to the first characteristic data and the second characteristic data in the first information.
- the first network element can obtain the first functional model according to the first characteristic data.
- Functional model that executes network events for that target.
- the target network event takes network load balancing as an example.
- the node sNode (equivalent to the first network element in the solution of this application, such as the lower layer network element) has or stores the load of the area.
- Local data such as local data and/or information, access user traffic, user distribution, etc.
- the node sNode needs to be trained to obtain a regression model, which is used to determine the load threshold.
- the node cNode (equivalent to the second network element in the solution of the present application, such as the upper layer network element) has or stores the global load data of the area, such as the load data and/or information, service inflow and/or outflow of the base station in the area.
- the node cNode can assist the node sNode in adjusting the load threshold to obtain the optimal load threshold.
- the specific process of the method in this embodiment is as follows:
- S501B sNode sends a training assistance request to cNode.
- cNode receives the training assistance request from sNode (equivalent to the first request message in the above-mentioned solution of this application).
- This training assistance request is used to request cNode to provide load global characteristics (equivalent to the second feature in the above-mentioned solution of this application). data).
- the training assistance request includes network load information.
- the network load information includes load balance (LB) and hidden feature parameters, such as dimensions (equivalent to the information of the target network event in the above-mentioned solution of this application).
- S502B cNode sends a receipt message to sNode.
- the sNode receives the receipt message (equivalent to the first response message in the above solution of this application).
- the receipt message is used to notify the sNode that the cNode rejects the training assistance request, and the cNode continues to execute the following step S503B.
- the receipt message is used to notify the sNode that the cNode rejects the training assistance request, and the cNode will not continue to perform any of the following steps.
- sNode can only use local load partial data for training to obtain the regression model, or sNode can wait for a period of time and then send a training assistance request to cNode again.
- S503B cNode determines the global characteristics of the load.
- cNode determines the global characteristics of the load, it may include: cNode first collects global data, such as load data of area base stations, business inflow and/or outflow data, etc.; then, cNode extracts a model through local characteristics (equivalent to the above-mentioned application in this application) The second feature extraction model in the solution) extracts the global data to obtain load global features (equivalent to the second feature data in the solution of the present application).
- the sNode can also determine the local characteristics of the load by referring to the way the cNode determines the global characteristics of the load in step S503B.
- sNode determines the local characteristics of the load, it may include: sNode first collects local local data, such as local data and/or information, access user business volume, user distribution, etc.; then, sNode extracts a model (equivalent to The local data is extracted using the first feature extraction model in the above-mentioned solution of the present application) to obtain load local features (equivalent to the first feature data in the above-mentioned solution of the present application).
- local local data such as local data and/or information, access user business volume, user distribution, etc.
- sNode extracts a model (equivalent to The local data is extracted using the first feature extraction model in the above-mentioned solution of the present application) to obtain load local features (equivalent to the first feature data in the above-mentioned solution of the present application).
- S504B cNode sends auxiliary information to sNode.
- the sNode receives the auxiliary information (equivalent to the first information in the solution of the present application), and the auxiliary information includes the global load characteristics of the cNode and the network load information.
- the auxiliary information may include the above-mentioned receipt message, and cNode does not need to send the above-mentioned receipt message and the auxiliary information to sNode respectively, thereby reducing additional transmission overhead. and delay.
- the first auxiliary information also includes first indication information, which is used to indicate whether the sNode returns the adjustment amount of the load global characteristics of the cNode.
- the first instruction information is used to instruct the sNode not to return the load global features of the cNode. amount of adjustment. If the cNode determines that the local feature extraction model needs to be changed or adjusted, the first instruction information is used to instruct the sNode to return the adjustment amount of the load global feature of the cNode.
- S505B sNode trains a regression model based on the local load characteristics of sNode and the global load characteristics of cNode.
- sNode performs feature union based on the global load characteristics of cNode and the local load characteristics of sNode in the auxiliary information, and uses the combined feature data as the input data of the training model to obtain the regression model (equivalent to the above-mentioned feature of this application)
- this regression model is used to determine the network load threshold.
- the global load characteristics of cNode are 3-dimensional feature data
- the local load characteristics of sNode are 2-dimensional feature data. If the 3 dimensions corresponding to cNode and the 2 dimensions corresponding to sNode are not the same, then the global load characteristics of cNode can be Perform series feature combination with the load local features of sNode to obtain load feature data after feature combination. If the three dimensions corresponding to cNode and the two dimensions corresponding to sNode have the same dimensions, then the data of the same dimensions in the global load characteristics of cNode and the local load characteristics of sNode can be merged and combined.
- S506B sNode determines the network load threshold based on the regression model to execute network events.
- the sNode obtains the sample data of the network load including the characteristic data sample, inputs the characteristic data sample into the regression model, predicts the optimal load threshold, and executes the network event based on the optimal load threshold.
- S507B sNode updates the local feature extraction model of sNode.
- the sNode can use the error value between the optimal load threshold predicted based on the regression model and the real load threshold as the adjustment value of the local load characteristic of the sNode. Further, the sNode adjusts the local feature extraction model of the sNode according to the adjustment value of the load local feature of the sNode to obtain an updated local feature extraction model.
- the updated local feature extraction model can be used for the next round of functional model training, which can be performed with reference to the above steps S501B-S507B, which will not be described in detail here.
- S508B sNode sends the adjustment value of the load global characteristics of the cNode to the cNode.
- first auxiliary information includes first indication information for instructing sNode to return the adjustment value of the global load characteristics of cNode
- sNode will use the optimal load threshold value and the real load threshold value predicted based on the regression model.
- the error value (loss) between them is used as the adjustment value of the cNode's load global characteristics, and the adjustment value of the entire local characteristic data of the cNode is sent to the cNode.
- S509B The cNode updates the local feature extraction model of the cNode based on the adjustment value of the cNode's load global feature.
- steps S508B-S509B are optional steps.
- the target network event takes network load balancing as an example.
- the sNode that needs to predict the load threshold can request the cNode to provide global load characteristics, and the global load characteristics provided by the cNode can be characterized with the local network load characteristics of the sNode. Combined to assist the sNode training to obtain an accurate regression model, which can be used to predict the optimal load threshold.
- sNode can also update the local feature extraction model based on the optimal load threshold and the real load threshold.
- cNode instructs sNode to return the adjustment value of the load global feature
- sNode can also return to cNode.
- the adjustment value of the load global feature is used so that the cNode can update the local feature extraction model based on the adjustment value of the load global feature to ensure that the next round of data feature extraction is more accurate.
- the target network event takes network traffic as an example.
- the UE (equivalent to the first network element in the solution of this application, such as the lower layer network element) has or stores local traffic partial data. For example, channel state information (CSI), traffic type, service type, etc.
- the UE needs to train a regression model, which is used to predict network traffic.
- the base station xNodeB (equivalent to the second network element in the solution of the present application, such as the upper layer network element) has or stores global traffic data, such as load information of the base station, service inflow and/or outflow volume, etc.
- the base station xNodeB can Assist UE in traffic prediction. Referring to Figure 6B, the specific process of the method in this embodiment is as follows:
- S601B The UE sends a training assistance request to the xNodeB.
- xNodeB receives the training assistance request from the UE (equivalent to the first request message in the above-mentioned solution of this application).
- This training assistance request is used to request xNodeB to provide load global characteristics (equivalent to the second feature in the above-mentioned solution of this application). data).
- the training assistance request includes network traffic information, and the network traffic information includes traffic predictions and hidden feature parameters, such as dimensions (equivalent to the information on the target network events in the solution of the present application).
- S602B xNodeB sends a receipt message to the UE.
- the UE receives the receipt message (equivalent to the first response message in the solution of the present application).
- the receipt message is used to notify the UE that the xNodeB accepts the training assistance request, and the xNodeB continues to perform the following step S603B.
- the receipt message is used to notify the UE that the xNodeB rejects the training assistance request, and the xNodeB will not continue to perform any of the following steps.
- the UE can only use local traffic partial data for training to obtain the regression model, or the UE can wait for a period of time and then send a training assistance request to the xNodeB again.
- S603B xNodeB determines the global characteristics of traffic.
- xNodeB determines the global characteristics of traffic, it may include: xNodeB first collects global data, such as process data of area base stations, business inflow and/or outflow volume, etc.; then, xNodeB extracts a model through local features (equivalent to the above-mentioned application) The second feature extraction model in the solution) extracts the global data to obtain the global traffic characteristics (equivalent to the second feature data in the above-mentioned solution of this application).
- the UE may also determine the local characteristics of the traffic by referring to the method in which the xNodeB determines the global characteristics of the traffic in step S603B.
- the UE determines the local characteristics of the load, it may include: the UE first collects local local data, such as CSI, historical traffic, service type, etc.; then, the UE extracts the local characteristics through the local characteristics model (equivalent to the third step in the solution of the present application).
- a feature extraction model extracts the local data to obtain the local traffic features (equivalent to the first feature data in the above-mentioned solution of the present application).
- S604B xNodeB sends auxiliary information to the UE.
- the UE receives the auxiliary information (equivalent to the first information in the solution of the present application), and the auxiliary information includes the global traffic characteristics of the xNodeB and the network traffic information.
- the assistance information may include the above-mentioned receipt message.
- the xNodeB does not need to send the above-mentioned receipt message and the assistance information to the UE separately, thereby reducing additional transmission overhead. and delay.
- the first auxiliary information also includes first indication information, which is used to indicate whether the UE transmits back the adjustment amount of the xNodeB's traffic global characteristics.
- the first indication information is used to instruct the UE not to return the xNodeB's traffic global features. amount of adjustment. If the xNodeB determines that the local feature extraction model needs to be changed or adjusted, the first indication information is used to instruct the UE to return the adjustment amount of the xNodeB's traffic global feature.
- S605B The UE trains a regression model based on the local traffic characteristics of the UE and the global traffic characteristics of the xNodeB.
- the UE performs feature combination based on the xNodeB's traffic global features and the UE's traffic local features in the auxiliary information, and uses the combined features as input data for the training model to obtain a regression model (equivalent to the above solution of this application)
- This regression model is used to predict network traffic.
- the global traffic characteristics of xNodeB are 3-dimensional feature data
- the local traffic characteristics of UE are 2-dimensional feature data. If the 3 dimensions corresponding to xNodeB are different from the 2 dimensions corresponding to UE, then the global traffic characteristics of xNodeB can be Perform series feature combination with the local traffic characteristics of the UE to obtain the traffic characteristics after feature combination. If the three dimensions corresponding to the xNodeB and the two dimensions corresponding to the UE have the same dimensions, then the features of the same dimensions in the global traffic characteristics of the xNodeB and the local traffic characteristics of the UE can be merged and combined.
- S606B The UE determines network traffic according to the regression model to execute network events.
- the UE obtains sample data of network traffic including feature data samples, inputs the feature data samples into the regression model, predicts the traffic value, and executes the network event based on the predicted traffic value.
- S607B The UE updates the local feature extraction model of the UE.
- the UE may use the error value between the traffic value predicted based on the regression model and the real traffic value as the adjustment value of the local traffic characteristic of the UE. Further, the UE adjusts the local feature extraction model of the sNode according to the adjustment value of the UE's traffic local feature to obtain an updated local feature extraction model.
- the updated local feature extraction model can be used for the next round of functional model training, which can be performed with reference to the above steps S601B-S607B, which will not be described in detail here.
- S608B The UE sends the adjustment values of all local features of the xNodeB to the xNodeB.
- first auxiliary information includes first indication information used to instruct the UE to return the adjustment value of the global traffic characteristics of the xNodeB
- the UE will calculate the error value between the traffic value predicted based on the regression model and the actual traffic value ( loss), as the adjustment value of all local features of the xNodeB, and sends the adjustment value of all local features of the xNodeB to xNodeB.
- xNodeB updates the local feature extraction model of xNodeB based on the adjustment values of all local features of the xNodeB.
- the target network event takes network traffic as an example.
- the UE that needs to predict the network traffic value can request the xNodeB to provide global characteristics of the network traffic.
- the global characteristics of the network traffic provided by the xNodeB can be combined with the local characteristics of the UE's network traffic.
- Features are combined to assist the UE in training to obtain an accurate regression model, which is used to predict traffic values.
- the UE can also update the local feature extraction model based on the predicted traffic value and the real traffic value.
- the xNodeB instructs the UE to return the adjustment value of the traffic global feature
- the UE can also return the adjustment value of the traffic global feature to the xNodeB. value, so that the xNodeB can update the local feature extraction model based on the adjusted value of the traffic global feature to ensure higher accuracy in the next round of data feature extraction.
- the target network event takes sleep energy saving as an example.
- the access network RAN (equivalent to the first network element in the solution of this application, such as the lower layer network element) has or stores a local local Data, such as the load of the access network RAN (such as a base station), busy and idle time periods, user distribution information and/or data, etc.
- the RAN needs to be trained to obtain a classification model, which is used to predict the working status of the RAN (that is, whether to enter a dormant energy-saving state).
- the core network CN (equivalent to the second network element in the solution of the present application, such as the upper layer network element) has or stores global network data, such as service data of area base stations, user business traffic information and/or data, etc. CN can assist RAN in predicting dormant energy-saving states.
- the specific process of the method in this embodiment is as follows:
- S701B The RAN sends a training assistance request to the CN.
- the CN receives the training assistance request from the RAN (equivalent to the first request message in the above-mentioned solution of the present application), and the training assistance request is used to request the CN to provide network global features (equivalent to the second feature in the above-mentioned solution of the present application). data).
- the training assistance request includes dormant energy saving information.
- the dormant energy saving information includes energy conservation (EC) and hidden feature parameters, such as dimensions (equivalent to the information of the target network event in the above-mentioned solution of this application).
- S702B CN sends a receipt message to RAN.
- the RAN receives the receipt message (equivalent to the first response message in the solution of the present application).
- the receipt message is used to notify the RAN that the CN accepts the training assistance request, and the CN continues to execute the following step S703B.
- the receipt message is used to notify the RAN that the CN rejects the training assistance request, and the CN will not continue to perform any of the following steps.
- the RAN can use local load partial data for training to obtain a regression model, or the RAN can wait for a period of time and then send a training assistance request to the CN again.
- S703B CN determines the global characteristics of the network.
- CN determines the global characteristics of the network, it may include: CN first collects global data of the network, such as the service status of base stations in the area, business inflow and/or outflow information and/or data, etc.; then, CN extracts the model through local characteristics (Equivalent to the second feature extraction model in the above-mentioned solution of this application) Feature extraction is performed on the global data of the network to obtain global network features (equivalent to the second feature data in the above-mentioned solution of this application).
- global data of the network such as the service status of base stations in the area, business inflow and/or outflow information and/or data, etc.
- Feature extraction is performed on the global data of the network to obtain global network features (equivalent to the second feature data in the above-mentioned solution of this application).
- the RAN may also determine local features by referring to the way in which the CN determines global features of the network in step S703B.
- RAN when it determines local features, it may include: RAN first collects local local data, such as base station load, busy and idle time periods, user distribution information and/or data, etc.; then, RAN extracts a model (equivalent to The first feature extraction model in the above-mentioned solution of the present application) performs feature extraction on the local data to obtain local features (equivalent to the first feature data in the above-mentioned solution of the present application).
- local local data such as base station load, busy and idle time periods, user distribution information and/or data, etc.
- RAN extracts a model (equivalent to The first feature extraction model in the above-mentioned solution of the present application) performs feature extraction on the local data to obtain local features (equivalent to the first feature data in the above-mentioned solution of the present application).
- S704B CN sends auxiliary information to RAN.
- the RAN receives the auxiliary information (equivalent to the first information in the solution of the present application), and the auxiliary information includes the CN's network global characteristics and dormant energy saving information.
- the assistance information may include the above-mentioned receipt message.
- the CN does not need to send the above-mentioned receipt message and the assistance information to the RAN respectively, thereby reducing additional transmission overhead. and delay.
- the first auxiliary information also includes first indication information, which is used to indicate whether the RAN returns the adjustment amount of the network global characteristics of the CN.
- the first indication information is used to instruct the RAN not to return the global network features of the CN. amount of adjustment. If the CN determines that the local feature extraction model needs to be changed or adjusted, the first indication information is used to instruct the RAN to return the adjustment amount of the network global features of the CN.
- S705B RAN trains a classification model based on the local characteristics of RAN and the global network characteristics of CN.
- the RAN performs feature combination based on the global network features of the CN and the local features of the RAN in the auxiliary information, and uses the combined features as input data for the training model to obtain a classification model (equivalent to the above solution in this application)
- this classification model is used to predict whether to enter the hibernation energy-saving state.
- the global features of CN are 3-dimensional features
- the local features of RAN are 2-dimensional features. If the 3 dimensions corresponding to CN are different from the 2 dimensions corresponding to RAN, then the global features of CN and the local features of RAN can be Perform a series of feature combinations to obtain the features after feature combination. If the three dimensions corresponding to CN are the same as the two dimensions corresponding to RAN, then the global features of CN and the local features of RAN can be combined into a merged feature.
- S706B RAN determines whether to enter the dormant energy-saving state based on the classification model.
- the RAN obtains the sample data of the network including feature samples, inputs the feature samples into the classification model, and predicts that the RAN will enter the dormant energy-saving state, and then the RAN will enter the dormant energy-saving state.
- S707B RAN updates the local feature extraction model of RAN.
- the RAN can predict the working state of the RAN based on the above classification model (such as predicting that the RAN will enter the dormant energy-saving state) and the actual working state of the RAN (such as the RAN failing to enter the dormant energy-saving state). error, adjust the local feature extraction model of RAN to obtain an updated local feature extraction model.
- the updated local feature extraction model of RAN can be used for the next round of functional model training, which can be performed with reference to the above steps S701B-S707B, which will not be described in detail here.
- S708B The RAN sends the adjustment value of the network global characteristics of the CN to the CN.
- the RAN sends the adjustment value of the network global characteristics of the CN to the CN, that is, the above-mentioned prediction based on the classification model is obtained.
- S709B The CN updates the local feature extraction model of the CN according to the adjustment value of the global feature of the CN.
- CN updates the local feature extraction model of CN based on the error between the working status of RAN predicted by the above classification model and the actual working status of RAN.
- the target network event takes sleep energy saving as an example.
- the RAN can request the CN to provide global network characteristics, and then the RAN combines the global network characteristics provided by the CN with the local local characteristics of the RAN. Features are combined to assist the RAN training to obtain an accurate classification model.
- This classification model can be used to predict whether the RAN can enter a dormant energy-saving state.
- the RAN can also update the local feature extraction model based on the predicted dormant energy-saving state and the real working state.
- the RAN when the CN indicates the adjustment value of the global feature of the RAN backhaul network, the RAN also returns the adjustment value of the global feature of the network to the CN. , that is, the error between the working status of the RAN and the real working status of the RAN is predicted based on the classification model, so that the CN can predict the working status of the RAN and the error between the real working status of the RAN based on the classification model, and the local feature extraction model Updates are made to ensure that the next round of extracted data features is more accurate.
- embodiments of the present application provide a communication device, which can be used to perform the operations performed by the first network element in the above method embodiment.
- the communication device may also be a first network element, a processor of the first network element, or a chip.
- the device includes modules or units that perform one-to-one correspondence with the methods/operations/steps/actions described in the first network element in the above embodiments.
- the modules or units may be hardware circuits, software, or a combination of hardware circuits.
- the communication device may have a structure as shown in FIG. 8 .
- the device 800 includes a communication unit 801 and a processing unit 802; wherein the processing unit 802 is used to determine the first characteristic data of the target network event; the communication unit 801 is used to obtain the data from the second network event.
- the processing unit 802 receives the first information, which includes the information of the target network event and the second characteristic data of the target network event; the processing unit 802 is also configured to execute the For a target network event, the first functional model is obtained based on the first characteristic data and the second characteristic data.
- the communication device 800 may also include a storage unit (not shown in Figure 8), which may be used to store a computer program for executing the method/operation/step/action described by the first network element and /or instructions and/or information, etc.
- a storage unit not shown in Figure 8
- the communication device 800 may also include a storage unit (not shown in Figure 8), which may be used to store a computer program for executing the method/operation/step/action described by the first network element and /or instructions and/or information, etc.
- the communication unit is further configured to: before receiving the first information from the second network element, send a first request message to the second network element, where the first request message is used to Requesting second characteristic data of the target network event.
- the first request message includes information about the target network event, where the information about the target network event includes identification information of the target network event, and also includes the target network event information.
- the target network event information One or more of the target dimensions of the characteristic data of the network event, response delay, data accuracy, and quality of service QoS of the target network event.
- the communication unit 801 is further configured to: receive a first response message from the second network element, where the first response message is used to notify the first network element of the The second network element provides second characteristic data of the target network event.
- the first information further includes the first response message.
- the processing unit 802 when determining the first characteristic data of the target network event, is specifically configured to: obtain the first data set of the target network event through the communication unit 801; use The first feature extraction model performs feature extraction on the first data set to obtain the first feature data.
- the first feature extraction model is used by the first network element to extract feature data of the target network event.
- the first information further includes a first indication, and the first indication is used to indicate whether the first network element returns the adjustment amount of the second characteristic data.
- the processing unit 802 is further configured to obtain sample data of the target network event through the communication unit 801, where the sample data includes characteristic data samples and the target network event.
- the communication unit 801 when the first indication is used to instruct the first network element to return the adjustment amount of the second characteristic data, the communication unit 801 is further configured to: according to the first Instruct to send the adjustment amount of the second characteristic data to the second network element.
- embodiments of the present application provide a communication device, which can be used to perform the operations performed by the second network element in the above method embodiment.
- the communication device may also be a second network element, a processor of the second network element, or a chip.
- the device includes modules or units that perform one-to-one correspondence with the methods/operations/steps/actions described in the second network element in the above embodiments.
- the modules or units may be hardware circuits, software, or a combination of hardware circuits.
- the communication device may have a structure as shown in FIG. 8 .
- the device 800 includes a communication unit 801 and a processing unit 802; wherein the processing unit 802 is used to determine the second characteristic data of the target network event; the communication unit 801 is used to provide the first network
- the first information includes the information of the target network event and the second characteristic data of the target network event.
- the communication device 800 may also include a storage unit (not shown in Figure 8), which may be used to store a computer program for executing the method/operation/step/action described by the second network element and /or instructions and/or information, etc.
- a storage unit not shown in Figure 8
- the communication unit 801 is further configured to receive a first request message from the first network element before the processing unit 802 determines the second characteristic data of the target network event.
- the first request message is used to request the second characteristic data of the target network event.
- the first request message includes information about the target network event; wherein the information about the target network event includes identification information of the target network event, and also includes the target network event information.
- the target network event information includes information about the target network event; wherein the information about the target network event includes identification information of the target network event, and also includes the target network event information.
- the communication unit 801 is further configured to: send a first response message to the first network element, where the first response message is used to notify the first network element of the The second network element provides second characteristic data of the target network event.
- the first information further includes the first response message.
- the processing unit 802 when determining the second characteristic data of the target network event, is specifically configured to: obtain the second data set of the target network event through the communication unit 801; use The second feature extraction model performs feature extraction on the second data set to obtain the second feature data.
- the second feature extraction model is used by the second network element to extract feature data of the target network event.
- the first information further includes a first indication, and the first indication is used to indicate whether the first network element returns the adjustment amount of the second characteristic data.
- the communication unit 801 when the first indication is used to instruct the first network element to return the adjustment amount of the second characteristic data, the communication unit 801 is also used to: from the first network element The network element receives the adjustment amount of the second characteristic data.
- the processing unit 802 is further configured to adjust the output layer of the second feature extraction model of the second network element according to the adjustment amount of the second feature data.
- the communication device 900 includes: a transceiver 901 , a processor 902 and a memory 903 .
- the transceiver 901, the processor 902 and the memory 903 are connected through a bus 904 to realize data exchange.
- the processor 902 and the memory 903 can be integrated together.
- the communication line 904 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc.
- PCI peripheral component interconnect
- EISA extended industry standard architecture
- the communication line 904 can be divided into an address bus, a data bus, a control bus, etc. For ease of presentation, only one thick line is used in Figure 9, but it does not mean that there is only one bus or one type of bus.
- the transceiver 901 in the communication device 900 includes the sending and/or receiving functions of the communication unit 801 in the communication device 800 .
- the transceiver 901 is used to support the sending and receiving of information, data, etc. between the communication device 800 and the second network element in the above embodiment.
- Memory 903 is used for program code and data of communications device 900 .
- the processor 902 is used to call the program code and data stored in the memory 903, and execute the methods involving the first network element in the methods shown in Figures 4A-4B, 5A-5B, 6A-6B, and 7A-7B. processing and/or other processes for the techniques described herein.
- the communication device 900 may also include other interfaces, such as optical fiber link interfaces, Ethernet interfaces, microwave link interfaces, copper wire interfaces, etc., to realize interaction between the communication device 900 and the second network element.
- other interfaces such as optical fiber link interfaces, Ethernet interfaces, microwave link interfaces, copper wire interfaces, etc.
- the processor 902 may be a central processing unit, ASIC, FPGA or CPLD.
- the communication device 900 shown in FIG. 9 only includes a transceiver 901, a processor 902 and a memory 903.
- the number of the transceiver 901, the processor 902, and the memory 903 may be one or multiple.
- the communication device 900 shown in Figure 9 can also implement the second network in the method provided by the embodiments corresponding to Figures 4A-4B, 5A-5B, 6A-6B, and 7A-7B.
- the communication device 900 shown in FIG. 9 may be the same device as the communication device 800 shown in FIG. 8 described above. Therefore, the implementation of the communication device 900 that is not described in detail may refer to the relevant descriptions in the methods provided by the corresponding embodiments of FIGS. 4A-4B, 5A-5B, 6A-6B, and 7A-7B or refer to the above. Relevant description in the communication device 900 shown in Figure 9. No more details will be given here.
- FIG. 10 is a schematic structural diagram of a chip provided by an embodiment of the present application.
- the chip 1000 includes an interface circuit 1001 and one or more processors 1002 .
- the chip 1000 may also include a bus. in:
- the processor 1002 may be an integrated circuit chip with signal processing capabilities. During the implementation process, each step of the above eye tracking method can be completed by instructions in the form of hardware integrated logic circuits or software in the processor 1002 .
- the above-mentioned processor 1002 can be a general-purpose processor, a digital communicator (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. .
- DSP digital communicator
- ASIC application-specific integrated circuit
- FPGA field programmable gate array
- a general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
- the interface circuit 1001 can be used to send or receive data, instructions or information.
- the processor 1002 can use the data, instructions or other information received by the interface circuit 1001 to process, and can send the processed information through the interface circuit 1001.
- the chip also includes a memory 1003, which may include read-only memory and random access memory, and provide operating instructions and data to the processor.
- memory 1003 may also include non-volatile random access memory (NVRAM).
- NVRAM non-volatile random access memory
- the memory stores executable software modules or data structures
- the processor can perform corresponding operations by calling operating instructions stored in the memory (the operating instructions can be stored in the operating system).
- the chip can be used in the first network element involved in the embodiment of this application.
- the interface circuit 1001 may be used to output execution results of the processor 1002.
- interface circuit 1001 and the processor 1002 can be realized through hardware design, software design, or a combination of software and hardware, which are not limited here.
- embodiments of the present application also provide a computer-readable storage medium, which stores some instructions. When these instructions are called and executed by the computer, the computer can complete the above method embodiments and method implementation. Examples of any of the possible designs involved.
- the computer-readable storage medium is not limited. For example, it may be RAM (network device random-access memory), ROM (read-only memory), etc.
- this application also provides a computer program product, which when called and executed by a computer can complete the method embodiments and the methods involved in any possible design of the above method embodiments.
- this application also provides a chip, which may include a processor and an interface circuit, for completing the above method embodiments and any possible implementation of the method embodiments.
- a chip which may include a processor and an interface circuit, for completing the above method embodiments and any possible implementation of the method embodiments.
- method where "coupling" means that two components are directly or indirectly combined with each other. This combination can be fixed or movable. This combination can allow flowing liquid, electricity, electrical signals or other types of signals to be transmitted between the two components. communicate between components.
- At least one involved in the embodiments of this application includes one or more; where multiple means greater than or equal to two.
- words such as “first” and “second” are only used for the purpose of distinguishing the description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating. Or suggestive order.
- each functional module in each embodiment of the present application may be integrated into one processing unit. In the device, it can exist physically alone, or two or more modules can be integrated into one module.
- the above integrated modules can be implemented in the form of hardware or software function modules.
- Embodiments of the present application provide a computer-readable storage medium storing a computer program.
- the computer program includes instructions for executing the above method embodiments.
- Embodiments of the present application provide a computer program product containing instructions that, when run on a computer, cause the computer to execute the above method embodiments.
- each functional module in each embodiment of the present application may be integrated into one processing unit. In the device, it can exist physically alone, or two or more modules can be integrated into one module.
- the above integrated modules can be implemented in the form of hardware or software function modules.
- embodiments of the present application may be provided as methods, systems, or computer program products. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment that combines software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) having computer-usable program code embodied therein.
- computer-usable storage media including, but not limited to, disk storage, CD-ROM, optical storage, etc.
- These computer program instructions may also be stored in a computer-readable memory that causes a computer or other programmable data processing apparatus to operate in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including the instruction means, the instructions
- the device implements the functions specified in a process or processes of the flowchart and/or a block or blocks of the block diagram.
- These computer program instructions may also be loaded onto a computer or other programmable data processing device, causing a series of operating steps to be performed on the computer or other programmable device to produce computer-implemented processing, thereby executing on the computer or other programmable device.
- Instructions provide steps for implementing the functions specified in a process or processes of a flowchart diagram and/or a block or blocks of a block diagram.
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
本申请公开了一种通信方法和装置,该方法包括:第一网元确定目标网络事件的第一特征数据;该第一网元从第二网元接收第一信息,该第一信息中包括该目标网络事件的信息、该目标网络事件的第二特征数据(S404B);该第一网元根据第一功能模型,执行该目标网络事件,该第一功能模型是该第一网元根据该第一特征数据和该第二特征数据得到的(S405B)。通过该方法,第一网元在保证第二网元的数据隐私安全性的情况下,可以利用第二网元的数据和信息得到准确的第一功能模型,同时该第一网元基于该第一功能模型执行该目标网络事件,也可提高执行的准确度。
Description
本申请涉及通信技术领域,尤其涉及一种通信方法和装置。
随着通信技术的发展,越来越多的应用将通过人工智能(artificial intelligence,AI)实现智能化,例如第五代移动通信(5th generation mobile communication,5G)网络协议R16(release16)核心网(core network,CN)开始研究网络数据分析功能((network data analysis functionality,NWDAF)支持网络的AI功能。AI功能的应用通常包括AI模型的训练和推理两个过程,AI模型的训练的目的是获得一个或者多个AI模型,AI模型的推理的目的是将训练的AI模型应用于新的数据中,以得到预测结果或评价结果。
通常网络中不同层的网元(或节点)拥有的数据和信息资源不同,处于上层的网元拥有全局性的数据和信息,而下层的网元拥有局部性的数据和信息,而现有关于AI的技术方案仅支持单个网元(或节点)使用自身拥有的数据和信息资源进行AI模型(即功能模型)的训练,并不能充分利用上层网元(或节点)的数据和信息资源,因此,导致训练得到AI模型的精确度较低。然而,若不同的网元(或节点)之间直接提供并传输各自的原始数据时,从而无法保证数据的安全性。
因此,亟需提供一种通信方法,在保证数据的安全性下,不同网元之间进行交互以得到准确度较高的功能模型。
发明内容
一种通信方法和装置,在保证数据的安全性下,不同网元之间进行交互以得到准确度较高的功能模型。
第一方面,本申请提供一种通信方法,该方法可以由第一网元执行,也可以由第一网元的部件(例如处理器、芯片、或芯片***等)执行,本申请对此不做具体限定。该方法具体可包括以下步骤:第一网元确定目标网络事件的第一特征数据;该第一网元从第二网元接收第一信息,该第一信息中包括该目标网络事件的信息、该目标网络事件的第二特征数据;该第一网元根据第一功能模型,执行该目标网络事件,该第一功能模型是该第一网元根据该第一特征数据和该第二特征数据得到的。
在本申请的方案中,第一网元确定目标网络事件的第一特征数据(即本地特征数据),然后,该第一网元从第二网元接收第一信息,该第一信息中包括该目标网络事件的信息和该目标网络事件的第二特征数据,进一步的,该第一网元可以根据该第一特征数据和第二网元提供的第二特征数据得到第一功能模型。通过该方案,第一网元在保证第二网元的数据的安全性下,可以有效地利用该第二网元提供的目标网络事件的第二特征数据,以辅助本地目标网络事件的特征数据得到准确性较高的第一功能模型,进而也可以提高基于该第一功能模型执行该目标网络事件的准确度。
在一种可能的实施方式中,该第一网元从第二网元接收第一信息之前,还包括:该第一网元向该第二网元发送第一请求消息,该第一请求消息用于请求该目标网络事件的第二 特征数据。
通过该实施方式,需要得到第一功能模型的第一网元针对目标网络事件,可以有效地向第二网元请求所需的第二特征数据。
在一种可能的实施方式中,该第一请求消息中包括该目标网络事件的信息;其中,该目标网络事件的信息包括该目标网络事件的标识信息,以及还包括该目标网络事件的特征数据的目标维度、响应时延、数据精度、该目标网络事件的服务质量QoS中的一项或多项。
通过该实施方式,该第二网元根据该第一请求消息,可以准确地知晓该目标网络事件的各信息,以及该第一网元请求目标网络事件的特征数据的需求信息。
在一种可能的实施方式中,该方法还包括:该第一网元从该第二网元接收第一响应消息,该第一响应消息用于向该第一网元通知该第二网元该所述目标网络事件的第二特征数据。
通过该实施方式,第一网元从第二网元接收该第一响应消息,可知该第二网元接受该第一网元的请求,并确定该第二网元可以提供目标网络事件的第二特征数据。
在一种可能的实施方式中,该第一信息还包括该第一响应消息。
通过该实施方式,可以避免第二网元分别向第二网元发送第一响应消息和第一信息而产生额外的资源开销和时延。
在一种可能的实施方式中,该第一网元确定目标网络事件的第一特征数据,包括:该第一网元获取该目标网络事件的第一数据集;该第一网元利用第一特征提取模型,对该第一数据集进行特征提取,得到该第一特征数据,该第一特征提取模型用于该第一网元提取该目标网络事件的特征数据。
通过该实施方式,该第一网元利用第一特征提取模型(即本地特征提取模型),有效地对目标网络事件的第一数据集(即本地数据集)进行特征提取,以得到第一特征数据。
在一种可能的实施方式中,该第一信息中还包括第一指示,该第一指示用于指示该第一网元是否回传所述第二特征数据的调整量。
通过该实施方式,该第二网元还可以通过第一指示向第一网元告知是否更新第二特征提取模型,进而指示该第一网元回传第二特征数据的调整量,以使得第二网元可利用该第二特征数据的调整量,对该第二特征提取模型进行有效地调整,进而可提高该第二特征提取模型的准确度。
在一种可能的实施方式中,该方法还包括:该第一网元获取该目标网络事件的样本数据,该样本数据中包含特征数据样本和该目标网络事件的真实值;该第一网元将所述特征数据样本输入该第一功能模型,得到该目标网络事件的输出值;该第一网元根据该目标网络事件的输出值和该目标网络事件的真实值,得该第二特征数据的调整量,该第二特征数据的调整量用于调整该第二网元的第二特征提取模型的输出层。
通过该实施方式,该第一网元可以有效地得到第二特征数据的调整量,以使得该第二网元可以根据该第二特征数据的量,调整第二网元的第二特征提取模型的输出层,还可以调整该第二特征提取模型的其它层,无需该第一网元多次向该第二网元反馈第二特征提取模型的其它层的调整量,从而可以节省更新该第二特征提取模型的资源开销和时延。
在一种可能的实施方式中,该第一指示用于指示该第一网元回传该第二特征数据的调整量时,该方法还包括:该第一网元根据该第一指示,向该第二网元发送该第二特征数据的调整量。
通过该实施方式,当第二网元向第一网元反馈第一信息中的第一指示,用于指示该第一网元回传该第二特征数据的调整量时,该第一网元还能有效地向该第二网元反馈该第二特征数据的调整量,以使得该第二网元可以基于该第二特征数据的调整量对第二特征提取模型进行准确且有效地调整。
第二方面,本申请提供一种通信方法,该方法可以由第二网元执行,也可以由第二网元的部件(例如处理器、芯片、或芯片***等)执行,本申请对此不做具体限定。该方法具体可包括以下步骤:第二网元确定目标网络事件的第二特征数据;该第二网元向第一网元发送第一信息,该第一信息中包括该目标网络事件的信息和该目标网络事件的第二特征数据。
在本申请的方案中,第二网元针对目标网络事件确定第二特征数据,然后,将该目标网络事件的信息和该第二特征数据携带在第一信息中发送给第一网元,从而在保证该第二网元的数据安全性下,使得该第一网元可以利用该第二网元提供的第二特征数据得到准确性更高的第一功能模型。
在一种可能的实施方式中,该第二网元确定目标网络事件的第二特征数据之前,还包括:该第二网元从该第一网元接收第一请求消息,该第一请求消息用于请求所述目标网络事件的第二特征数据。
通过该实施方式,该第二网元接收第一网元发送的第一请求消息,从而可以确定该目标网络事件,进而可有效地提供该目标网络事件的第二特征数据。
在一种可能的实施方式中,该第一请求消息中包括该目标网络事件的信息,其中,该目标网络事件的信息包括该目标网络事件的标识信息,以及还包括该目标网络事件的特征数据的目标维度、响应时延、数据精度、该目标网络事件的服务质量QoS中的一项或多项。
通过该实施方式,该第二网元根据该第一请求消息,可以准确地知晓该目标网络事件的各信息,以及该第一网元请求目标网络事件的特征数据的需求信息。
在一种可能的实施方式中,该方法还包括:该第二网元向该第一网元发送第一响应消息,该第一响应消息用于向该第一网元通知该第二网元提供该目标网络事件的第二特征数据。
通过该实施方式,该第二网元可以有效地向第一网元告知可以提供该目标网络事件的第二特征数据,使得该第一网元确定可以利用该第二网元提供的第二特征数据得到第一功能模型。
在一种可能的实施方式中,该第一信息还包括该第一响应消息。
通过该实施方式,可以避免第二网元分别向第一网元反馈第一信息和第一响应消息而产生的额外的资源开销和时延。
在一种可能的实施方式中,该第二网元确定目标网络事件的第二特征数据,包括:该第二网元获取该目标网络事件的第二数据集;该第二网元利用第二特征提取模型,对该第二数据集进行特征提取,得到该第二特征数据,该第二特征提取模型用于该第二网元提取该目标网络事件的特征数据。
通过该实施方式,第二网元可以利用第二特征提取模型(即本地特征提取模型)对该目标网络事件的第二数据集进行特征提取,以有效地得到该目标网络事件的特征数据。
在一种可能的实施方式中,该第一信息中还包括第一指示,该第一指示用于指示该第一网元是否回传该第二特征数据的调整量。
通过该实施方式,可以有效地指示第一网元是否需要向第二网元反馈第二特征数据的调整量。
在一种可能的实施方式中,该第一指示用于指示该第一网元回传该第二特征数据的调整量时,该方法还包括:该第二网元从该第一网元接收该第二特征数据的调整量。
通过该实施方式,当第一指示用于指示第一网元回传第二特征数据的调整量时,该第二网元还可以从第一网元有效地接收到第二特征数据的调整量,进而可以优化该第二特征提取模型。
在一种可能的实施方式中,该方法还包括:该第二网元根据该第二特征数据的调整量,对该第二网元的第二特征提取模型的输出层进行调整。
通过该实施方式,该第一网元仅向第二网元反馈用于调整第二特征提取模型的输出层的调整量,该第二网元可以根据该第二特征数据的调整量,有效地对该第二网元的第二特征提取模型的输出层进行调整,同时还可以对该第二特征提取模型的其它层进行调整,以提高该第二特征提取模型的准确度。
第三方面,本申请实施提供一种通信装置,该通信装置可应用于第一网元,具有实现上述第一方面或上述第一方面的任意一种可能的实施方式的功能。该装置可以包括:通信单元(通信单元包括发送单元和/或接收单元)、处理单元。
第四方面,本申请实施提供一种通信装置,该通信装置可应用于第二网元,具有实现上述第一方面或上述第二方面的任意一种可能的实施方式的功能。该装置可以包括:通信单元(通信单元包括发送单元和/或接收单元)、处理单元。
第五方面,本申请还提供一种通信装置,该通信装置可应用于第一网元,该通信装置包括处理器,用于实现上述第一方面或上述第一方面的任意一种可能的实施方式的功能。可选地,该通信装置还包括收发器,用于实现该通信装置的通信功能。
第六方面,本申请还提供一种通信装置,该通信装置可应用于第二网元,该通信装置包括处理器,用于实现上述第二方面或上述第二方面的任意一种可能的设计中的方法的功能。可选地,该通信装置还包括收发器,用于实现该通信装置的通信功能。
第七方面,本申请还提供一种通信***,该***包括用于执行上述第一方面提供的方法的第一网元,以及用于执行上述第二方面提供的方法的第二网元。
第八方面,本申请实施例还提供一种计算机存储介质,该存储介质中存储软件程序,该软件程序在被一个或多个处理器读取并执行时可实现上述第一方面或其中任意一种可能的实施方式提供的方法,或者实现上述第二方面或其中任意一种可能的实施方式提供的方法。
第九方面,本申请实施例还提供一种包含指令的计算机程序产品,当其在计算机上运行时,使得上述第一方面或其中任一种可能的实施方式提供的方法执行,或者使得上述第二方面或其中任一种可能的实施方式提供的方法执行。
第十方面,本申请实施例还提供一种芯片***,该芯片***包括处理器,用于支持第一网元实现上述第一方面中所涉及的功能;或者用于支持第二网元实现上述第二面中所涉及的功能。
在一种可能的设计中,所述芯片***还包括存储器,所述存储器,用于保存装载装置执行的必要的程序指令和数据。该芯片***,可以由芯片构成,也可以包含芯片和其他分立器件。
上述第三方面和第五方面以及其中任一种可能的实施方式可以达到的技术效果,可以参照上述第一方面和第二方面以及其中任意一种可能的实施方式,上述第三方面和第六方面以及其中任一种可能的实施方式可以达到的技术效果,可以参照上述第二方面和第二方面以及其中任意一种可能的实施方式,这里不再重复赘述。
图1为一种数据收集与上报的参考架构示意图;
图2为本申请实施例提供的通信方法所适用的一种网络架构的示意图;
图3为本申请实施例提供的通信方法所适用的一种通信***的示意图;
图4A为本申请实施例提供的一种通信方法的结构示意图;
图4B为本申请实施例提供的一种通信方法的流程示意图;
图5A为本申请实施例提供的第一个实施例的结构示意图;
图5B为本申请实施例提供的第一个实施例的流程示意图;
图6A为本申请实施例提供的第二个实施例的结构示意图;
图6B为本申请实施例提供的第二个实施例的流程示意图;
图7A为本申请实施例提供的第三个实施例的结构示意图;
图7B为本申请实施例提供的第三个实施例的流程示意图;
图8为本申请实施例提供的一种通信装置的结构示意图;
图9为本申请实施例提供的一种通信装置的结构示意图;
图10为本申请实施例提供的一种芯片装置的结构示意图。
本申请实施例提供了一种通信方法和装置,其中,方法和装置是基于相同或相似技术构思的,由于方法及装置解决问题的原理相似,因此,方法与装置的实施可以相互参见,重复之处不再赘述。
随着通信技术的发展,越来越多的应用将通过人工智能(artificial intelligence,AI)实现智能化,例如从5G网络协议R16核心网CN开始研究网络数据分析功能NWDAF支持网络的AI功能。AI功能的应用通常包括AI模型的训练和推理两个过程,AI模型的训练的目的是获得一个或者多个AI模型,AI模型的推理的目的是将训练的AI模型应用于新的数据中,以得到预测结果或评价结果。
图1为R17协议中制定的数据收集与上报的参考架构示意图,UE侧数据可以是UE中的业务的体验质量(quality of experience,QoE)或者其它的数据,在此不做限定。UE侧数据可以为由UE获取的数据。UE侧数据也可以称为UE侧的数据、UE的数据或UE数据(UE data)。
下面对图1所示架构中的各个网元(或节点)进行简单的描述。
数据收集AF(data collection AF,DCAF),可以接收来自应用服务提供商(application service provider,ASP)的用于配置数据收集与上报功能的配置信息。示例性的,网络侧可以通过ASP中的业务开通AF(Provisioning AF)与DCAF进行交互,并开通DCAF的数据收集与上报功能。然后,DCAF可以生成数据收集与上报配置信息,并将数据收集与上 报配置信息发往数据收集客户端侧,以使数据收集客户端根据接收到的数据收集与上报配置信息,收集数据并上报数据收集报告;DCAF还可以接收数据收集客户端上报的数据收集报告。
其中,数据收集客户端可以包括:直接数据收集客户端(direct data collection client,DDCC),间接数据收集客户端(indirect data collection client,IDCC)或应用服务器(application server,AS)。DCAF可以对数据收集报告中的数据进行处理,并根据其它网元或第三方实体的订阅请求,将处理后的UE侧数据发往订阅网元或第三方实体。DDCC是UE中的数据收集功能实体,DDCC能够通过R2接口接收来自DCAF的数据收集与上报配置信息,并根据该数据收集与上报配置信息进行数据收集与上报。UE应用(Application,APP),可以将相关的UE侧数据通过R7接口分享给DDCC。应用服务提供商(ASP),可以通过R8接口从UE应用处接收UE侧数据。其中,R8接口属于纯应用层实现。ASP内可以认为存在一个间接数据收集客户端(IDCC),业务开通AF通过R1接口将数据收集报告发送到DCAF,IDCC将通过应用层收集到的UE侧数据通过R3接口发往DCAF。在此之前,IDCC也会通过R3接口接收来自DCAF的数据收集与上报配置信息。应用服务器(AS),可以通过R4接口接收来自DCAF的数据收集与上报配置信息,然后根据数据收集与上报配置信息进行数据收集,得到数据收集报告,并通过R4接口将数据收集报告发往DCAF,该数据收集报告包括应用服务器的访问日志信息等。NWDAF,可以通过R5接口向DCAF订阅并接收UE侧数据,然后对该UE侧数据进行分析。ASP也可能会获取DCAF收集的数据,比如ASP内的事件订阅AF(Event Consumer AF)可以通过R6接口向DCAF订阅所收集的数据,并从DCAF接收所订阅的数据信息。
然而,通过上述NWDAF收集并分析应用层的数据,以实现网络的AI功能收集数据时,由于网络中不同层的网元(或节点)拥有的数据和信息资源不同,处于上层的网元拥有全局性的数据和信息,而下层的网元拥有局部性的数据和信息,而现有关于AI的技术方案仅支持单个网元(或节点)使用自身拥有的数据和信息资源进行AI模型(即功能模型)的训练,并不能充分利用其它网元(或节点)的数据和信息资源,因此,导致训练得到AI模型的精确度较低。另外,若不同的网元(或节点)之间可以直接提供并传输各自的原始数据时,从而无法保证数据的安全性。
因此,本申请提供一种通信方法,该方法包括:第一网元确定目标网络事件的第一特征数据;该第一网元从第二网元接收第一信息,该第一信息中包括该目标网络事件的信息、该目标网络事件的第二特征数据;该第一网元根据第一功能模型,执行该目标网络事件,该第一功能模型是该第一网元根据该第一特征数据和该第二特征数据得到的。通过该方法,第一网元在保证第二网元的数据隐私安全性的情况下,可以利用第二网元的数据和信息得到准确的第一功能模型,同时该第一网元基于该第一功能模型执行该目标网络事件,也可提高执行的准确度。
需要注意的是,本申请的方案可以适用于5G通信***,或者将要应用的6G通信***,或者未来演进的通信***,或者其他通信***等,本申请对此不做限定。本申请实施例场景是以信号传输为背景。
图2示出了本申请实施例提供的通信方法可以适用的一种网络架构的示意图。如图2所示,该网络架构中可以包括接入网以及核心网。终端设备通过接入网和核心网接入数据网络(data network,DN)。
终端设备可以为用户设备(user equipment,UE)、移动台、移动终端、应用客户端等。终端设备可以广泛应用于各种场景,例如,设备到设备(device-to-device,D2D)、车物(vehicle to everything,V2X)通信、机器类通信(machine-type communication,MTC)、物联网(internet of things,IOT)、虚拟现实、增强现实、工业控制、自动驾驶、远程医疗、智能电网、智能家具、智能办公、智能穿戴、智能交通、智慧城市等。终端设备可以是手机、平板电脑、带无线收发功能的电脑、可穿戴设备、车辆、城市空中交通工具(如无人驾驶机、直升机等)、轮船、机器人、机械臂、智能家居设备等。附图和以下实施例中以UE作为终端设备的一个示例进行说明,后续任意地方出现的UE也可以替换为终端设备或终端设备的其它示例。
接入网用于实现接入有关的功能,可以为特定区域的授权用户提供入网功能,并能够根据用户的级别,业务的需求等确定不同质量的传输链路以传输用户数据。接入网在UE与核心网之间转发控制信号和用户数据。接入网可以包括接入网设备,接入网设备可以是为UE提供接入的设备,可以包括无线接入网(radio access network,RAN)设备和有线接入网设备。RAN设备,主要负责空口侧的无线资源管理、服务质量(quality of service,QoS)管理、数据压缩和加密等功能。RAN设备可以包括各种形式的基站,例如宏基站,微基站(也可称为小站),中继站,接入点,气球站等。在采用不同的无线接入技术的***中,具备基站功能的设备的名称可能会有所不同,例如,在5G***中,称为RAN或者下一代基站(next-generation Node basestation,gNB),在长期演进(long term evolution,LTE)***中,称为演进的节点B(evolved NodeB,eNB或eNodeB)。
接入网设备和UE可以是固定位置的,也可以是可移动的。接入网设备和UE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。本申请的实施例对接入网设备和UE的应用场景不做限定。
核心网负责维护移动网络的签约数据,为UE提供会话管理、移动性管理、策略管理以及安全认证等功能。核心网中包括但不限于以下一个或多个网元:应用功能(application function,AF)网元、统一数据管理(unified data management,UDM)网元、统一数据库(unified data repository,UDR)网元、策略控制功能(policy control function,PCF)网元、会话管理功能(session management function,SMF)网元、接入与移动性管理功能(access and mobility management function,AMF)网元、网络存储功能(network repository function,NRF)网元、鉴权服务器功能(authentication server function,AUSF)网元、网络开放功能(network exposure function,NEF)网元、用户面功能(user plane function,UPF)网元、网络数据分析功能网元(Network Data Analytics Function,NWDAF)网元。
AMF网元,主要负责移动网络中的移动性管理,例如用户注册网络、用户切换等。SMF网元,主要负责移动网络中的会话管理,例如会话建立、修改、释放等。UPF网元,主要负责用户数据的转发和接收,可以从数据网络接收用户数据,通过接入网络设备传输给UE;还可以通过接入网设备从UE接收用户数据,转发至数据网络。UDM网元,包含执行管理签约数据、用户接入授权等功能。UDR网元,包含执行签约数据、策略数据、应用数据等类型数据的存取功能。NEF网元,主要用于支持能力和事件的开放。AF网元,传递应用侧对网络侧的需求。PCF网元,主要支持提供统一的策略框架来控制网络行为,提供策略规则给控制层网络功能,同时负责获取与策略决策相关的用户签约信息。PCF网元可以向AMF网元、SMF网元提供策略等。NRF网元,可用于提供网元发现功能,基于其它网元 的请求,提供网元类型对应的网元信息,以及提供网元管理服务等。AUSF网元,负责对UE进行鉴权,验证UE的合法性。NWDAF网元,具有提供基于大数据和人工智能等技术的网络数据收集和分析等功能。DN,其上可部署多种业务,可为UE提供数据和/或语音等服务。
其中,AF网元、UDM网元、UDR网元、PCF网元、SMF网元、AMF网元、NRF网元、AUSF网元、NEF网元、UPF网元、NWDAF网元,也可以分别简称为AF、UDM、UDR、PCF、SMF、AMF、NRF、AUSF、NEF、UPF、NWDAF,如图1中所示。
图2中Nausf、Nnef、Nnrf、Namf、Npcf、Nsmf、Nudm、Nudr、Naf、Nnwdaf分别为上述AUSF、NEF、NRF、AMF、PCF、SMF、UDM、UDR、AF、NWDAF提供的服务化接口,用于调用相应的服务化操作。N1、N2、N3、N4以及N6为接口序列号,这些接口序列号的含义如下:
N1:AMF与UE之间的接口,可以用于向UE传递非接入层(non access stratum,NAS)信令(如包括来自AMF的QoS规则)等。N2:AMF与接入网设备之间的接口,可以用于传递核心网侧至接入网设备的无线承载控制信息等。N3:接入网设备与UPF之间的接口,主要用于传递接入网设备与UPF间的上下行用户面数据。N4:SMF与UPF之间的接口,可以用于控制面与用户面之间传递信息,包括控制面向用户面的转发规则、QoS规则、流量统计规则等的下发以及用户面的信息上报。N6:UPF与DN的接口,用于传递UPF与DN之间的上下行用户数据流。
可以理解的是,上述网元或者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软件功能,或者是平台(例如,云平台)上实例化的虚拟化功能。作为一种可能的实现方法,上述网元或者功能可以由一个设备实现,也可以由多个设备共同实现,还可以是一个设备内的一个功能模块,本申请实施例对此不作具体限定。
本申请方案中的第一网元可以为上述网络中的下层网元,例如上述的应用功能AF网元、UE,可以具有本地局部的信息和/或数据,本申请方案中的第二网元可以为上述网络架构中的上层网元,例如CN、RAN设备,可以具有网络中全局性的信息和/或数据。
此外,本申请实施例还提供了一种通信方法所适用的通信***的架构示意图。如图3所示,该通信***中包括基站、核心网,以及多个终端设备UE,基站均连接核心网设备,并且该多个基站分别连接多个终端设备。本申请实施例对于同构网络与异构网络的场景均适用,同时对于传输点也无限制,可以是宏基站与宏基站、微基站与微基站和宏基站与微基站间的多点协同传输,对频分双工(frequency division duplexing,FDD)/时分双工(time division duplexing,TDD)***均适用。本发明实施例即适用于低频场景(sub 6G),也适用于高频场景(6G以上),太赫兹,光通信等。
本申请方案中的第一网元可以为图3中的各UE,可以具有本地局部的信息和/或数据,本申请方案中的第二网元可以为图3中的基站,可以具有网络中全局性的信息和/或数据。或者本申请方案中的第一网元可以为图3中的各基站,本申请方案中的第二网元可以为图3中的核心网。
因此,本申请对第一网元和第二网元的类型和物理形态不做具体约束,只要满足第二网元比第一网元拥有更具全局性和综合性的数据即可。
下面结合具体实施例介绍本申请的技术方案。
本申请实施例提供了一种通信方法,该方法可适用于但不限于图2所示的网络架构中,以及图3所示的通信***中,并且该方法可以由本申请涉及到的网元执行,或者由涉及到的网元对应的芯片执行,本申请中的网元可以为物理上的实体网元,也可以是虚拟的网元,本申请对涉及的网元的形态不做具体限定。
此外,在本申请实施例中,该第一网元为拥有局部性数据的网元,第二网元为拥有全局性数据的网元时,该第一网元还可以为上述的终端设备(或终端设备对应的控制器、芯片),该第二网元可以为上述的接入网设备(或接入网设备对应的控制器、芯片),例如基站。并且,需要说明的是,下文中提及的“第一”、“第二”等序数词是用于对多个对象进行区分,以便于描述,并不用于限定多个对象的顺序、时序、优先级或者重要程度。针对执行该实施例中第一网元和第二网元的具体形态,本申请不做限定。
图4A示出了本申请实施例提供的一种执行该通信方法的结构,图4B示出了本申请实施例提供的一种通信方法的流程,参考图4A和图4B,具体可以包括如下步骤:
S401B:第一网元向第二网元发送第一请求消息,该第一请求消息用于请求该目标网络事件的第二特征数据。
相应的,该第二网元接收该第一请求消息。
在本申请方案中,第一网元为下层网元,或者为需要从第二网元得到辅助数据,以得到第一功能模型的网元。第二网元为上层网元,或者为向第一网元提供全局性的特征数据,以辅助该第一网元得到功能模型的网元。
在一种实施方式中,该第一请求消息中包括该目标网络事件的信息,其中,该目标网络事件的信息包括该目标网络事件的标识信息,以及还包括该目标网络事件的特征数据的目标维度、响应时延、数据精度、该目标网络事件的服务质量QoS中的一项或多项。
需要注意的是,该目标网络事件可以为协议中明确定义的几种网络事件类型,例如负载均衡、节能等。该目标网络事件也可以为网络中网元(如第一网元和第二网元)之间预先约定的事件,只要其中的第二网元在收到该第一请求消息之后可以识别该目标网络事件即可,因此,本申请对该目标网络事件不做具体的限定。
S402B:第二网元向第一网元发送第一响应消息。
相应的,该第一网元接收该第一响应消息。
第二网元在接收该第一网元的第一请求消息之后,需要确定是否接受或拒绝该请求,即接收或拒绝向第一网元提供该目标网元事件的第二特征数据。若确定接受该请求,则执行下述步骤S403B,若确定拒绝该请求,则不执行下述任何步骤。
示例性的,第二网元确定是否接受或拒绝该请求可以包括但不限于以下几种原因:
该第二网元根据该第一请求消息中的目标网络事件的信息,若第二网元的本地数据中包括该目标网络事件相关的数据,则该第二网元确定可以接受该请求,若该第二网元的本地数据不包括该目标网络事件相关的数据,则该第二网元确定可以拒绝该请求。
或者该第二网元本身允许其它网元访问或获取本地的特征数据时,可以确定接受上述的请求,若该第二网元本身不允许其它网元访问或获取本地的特征数据时,可以确定拒绝上述的请求。又或者若该第二网元能传输第一网元所请求的第二特征数据,则确定接受上述的请求,若该第二网元不能传输该第一网元所请求的第二特征数据,则拒绝上述的请求。
因此,本申请对该第二网元接受或拒绝第一网元的第一请求消息的原因,不做具体限定。
当第二网元拒绝该第一请求消息后,第一网元可以根据本地的特征数据,得到下述的第一功能模型,或者该第一网元隔一段时间后,再向第二网元发送上述的第一请求消息。
在一种实施方式中,第二网元向第一网元发送第一响应消息,该第一网元接收该第一响应消息,该第一响应消息用于向第一网元通知第二网元提供目标网络事件的第二特征数据。在该实施方式下,该第二网元继续执行下述步骤S403B。
在另一种实施方式中,第二网元向第一网元发送第一响应消息,该第一网元接收该第一响应消息,该第一响应消息用于向第一网元通知第二网元不提供目标网络事件的第二特征数据。然而,在该实施方式下,该第二网元不继续执行下述步骤S403B以及后续任何步骤。
S403B:第二网元确定目标网络事件的第二特征数据。
在一种实施方式中,该第二网元确定目标网络事件的第二特征数据,包括:该第二网元获取该目标网络事件的第二数据集;该第二网元利用第二特征提取模型,对该第二数据集进行特征提取,得到该第二特征数据,该第二特征提取模型用于该第二网元提取该目标网络事件的特征数据。
此外,第一网元也确定目标网络事件的第一特征数据。
需要注意的是,第一网元确定目标网络事件的第一特征数据可以在上述步骤S403B之前执行,或与上述步骤S403B同时执行,或与下述步骤S404B同时执行,又或者在下述步骤S404B之后执行。因此,相比上述步骤S403B和下述步骤S404B,本申请对第一网元确定目标网络事件的第一特征数据的执行先后顺序不做具体限定。
在一种实施方式中,该第一网元确定目标网络事件的第一特征数据,包括:该第一网元获取该目标网络事件的第一数据集;该第一网元利用第一特征提取模型,对该第一数据集进行特征提取,得到该第一特征数据,该第一特征提取模型用于该第一网元提取该目标网络事件的特征数据。
S404B:第二网元向第一网元发送第一信息,该第一信息中包括目标网络事件的信息和该目标网络事件的第二特征数据。
在一种实施方式中,该第一信息还包括上述的第一响应消息,该第一响应消息用于向第一网元通知该第二网元提供该目标网络事件的第二特征数据。通过该实施方式,避免第一网元向第一网元分别传输第一响应消息和第一信息所产生的额外的资源开销和时延。
在一种实施方式中,该第一信息中还包括第一指示,该第一指示用于指示该第一网元是否回传该第二特征数据的调整量。
若该第一指示用于指示该第一网元回传该第二特征数据的调整量时,该第一网元在确定该第二特征数据的调整量之后,需要将该第二特征数据的调整量发送给第二网元。若该第一指示用于指示该第一网元不回传该第二特征数据的调整量时,该第一网元不需要向第二网元发送第二特征数据的调整量。
需要注意的是,如果第二网元的第二特征提取模型已经很成熟(即精准度很高),或者第二网元不允许本地的特征提取模型被更改,该第一指示用于指示该第一网元不回传该第二特征数据的调整量。如果第二网元的第二特征提取模型不成熟(即精准度较低),或者第二网元允许本地的特征提取模型被更改,该第一指示用于指示该第一网元回传该第二特征数据的调整量。
S405B:第一网元根据第一功能模型,执行目标网络事件,该第一功能模型是该第一 网元根据第一特征数据和第二特征数据得到的。
在一种实施方式中,该第一网元对该第一特征数据和第二特征数据进行特征联合,将特征联合后的特征数据作为功能模型的输入,得到第一功能模型。
通过该实施方式,第一网元使用该第一功能模型,可以更准确地执行该目标网络事件。其中,该第一功能模型可以为但不限于:预测模型、回归模型、分类模型、聚类模型。在本申请方案对该第一功能模型不做具体限定,可以按照实际需求得到。
S406B:该第一网元对第一特征提取模型进行更新。
在一种实施方式中,在执行上述步骤S406B时,可以包括以下步骤:
该第一网元确定第一特征数据的调整量;
该第一网元根据该第一特征数据的调整量对第一特征提取模型进行更新。
在一种实施方式中,该第一网元确定第一特征数据的调整量,包括:第一网元获取目标网络事件的样本数据,该样本数据中包含第一网元的特征数据样本和该目标网络事件的真实值;第一网元将该第一网元的特征数据样本输入第一功能模型,得到目标网络事件的输出值;第一网元根据目标网络事件的输出值和目标网络事件的真实值,得第一特征数据的调整量。其中,该第一特征数据的调整量用于调整第一网元的第一特征提取模型的输出层。
在一种实施方式中,该第一信息中的第一指示用于指示该第一网元回传第二特征数据的调整量时,该方法还包括:该第一网元根据该第一指示,向该第二网元发送该第二特征数据的调整量。
在一种实施方式中,该第一网元确定第二特征数据的调整量,包括:第一网元获取目标网络事件的样本数据,该样本数据中包含特征数据样本和目标网络事件的真实值;第一网元将该特征数据样本输入第一功能模型,得到该目标网络事件的输出值;第一网元根据目标网络事件的输出值和目标网络事件的真实值,得第二特征数据的调整量。其中,该第二特征数据的调整量用于调整第二网元的第二特征提取模型的输出层。
在本申请方案中,第一网元向第二网元发送第一请求消息,该第一请求消息用于请求目标网络事件的第二特征数据,然后,该第一网元在确定目标网络事件的第一特征数据和收到第一信息之后,该第一网元可以根据该第一特征数据和第一信息中的第二特征数据得到第一功能模型,该第一网元根据该第一功能模型,执行该目标网络事件。通过该方案,在保证第二网元的数据隐私安全的情况下,可以获取该第二网元的第二特征数据,并且联合该第一网元的第一特征数据得到第一功能模型,也可以提高利用该第一功能模型执行该目标网络事件的准确度。
下面的几个具体实施例中,针对不同的目标网络事件,进一步的详细阐述上述本申请方案提出的一种通信方法。
实施例一:
在该实施例一中,参考图5A所示,目标网络事件以网络负载均衡为例,节点sNode(相当于本申请方案中的第一网元,如下层网元)具有或存储有片区的负载局部数据,例如本地数据和/或信息、接入用户业务量、用户分布等。且该节点sNode需要训练得到回归模型,该回归模型用于决策负载门限。其中,节点cNode(相当于上述本申请方案中的第二网元,如上层网元)具有或存储有片区的负载全局数据,例如片区基站的负载数据和/或信息、业 务流入和/或流出量等,该节点cNode可以辅助节点sNode进行负载门限值的调整,以得到最优的负载门限值。参考图5B所示,该实施例方法的具体流程如下:
S501B:sNode向cNode发送训练辅助请求。
相应的,cNode从sNode接收该训练辅助请求(相当于上述本申请方案中的第一请求消息),该训练辅助请求用于请求cNode提供负载全局特征(相当于上述本申请方案中的第二特征数据)。
其中,该训练辅助请求中包括网络负载的信息,该网络负载的信息包括负载均衡(load balance,LB)、隐藏特征参数,例如维度(相当于上述本申请方案中的目标网络事件的信息)。
S502B:cNode向sNode发送回执消息。
相应的,该sNode接收该回执消息(相当于上述本申请方案中的第一响应消息)。
若cNode确定接受该训练辅助请求时,该回执消息用于向sNode通知该cNode拒绝该训练辅助请求,cNode继续执行下述步骤S503B。
若cNode确定拒绝该训练辅助请求时,该回执消息用于向sNode通知该cNode拒绝该训练辅助请求,cNode不再继续执行下述任何步骤。此时,sNode可以只利用本地的负载局部数据进行训练得到回归模型,或者sNode等待一段时间后再次向cNode发送训练辅助请求。
S503B:cNode确定负载全局特征。
示例性地,cNode确定负载全局特征时,可以包括:cNode先收集全局数据,如片区基站的负载数据、业务流入和/或流出数据等;然后,cNode通过本地特征提取模型(相当于上述本申请方案中的第二特征提取模型)对该全局数据进行提取,得到负载全局特征(相当于上述本申请方案中的第二特征数据)。
同时,sNode也可以参考该步骤S503B中的cNode确定负载全局特征的方式,确定负载局部特征。
示例性地,sNode确定负载局部特征时,可以包括:sNode先收集本地的局部数据,如本地数据和/或信息、接入用户业务量、用户分布等;然后,sNode通过本地特征提取模型(相当于上述本申请方案中的第一特征提取模型)对该局部数据进行提取,得到负载局部特征(相当于上述本申请方案中的第一特征数据)。
S504B:cNode向sNode发送辅助信息。
相应的,该sNode接收该辅助信息(相当于上述本申请方案中的第一信息),该辅助信息中包括cNode的负载全局特征和网络负载的信息。
在一种实施方式中,若cNode接受sNode的训练辅助请求时,该辅助信息中可以包括上述的回执消息,cNode无需向sNode分别发送上述的回执消息和该辅助信息,从而可以减少额外的传输开销和时延。
在一种实施方式中,该第一辅助信息中还包括第一指示信息,该第一指示信息用于指示sNode是否回传cNode的负载全局特征的调整量。
示例性地,若cNode确定不需要更改或调整本地的特征提取模型(相当于本申请方案中的第二特征提取模型)时,该第一指示信息用于指示sNode不回传cNode的负载全局特征的调整量。若cNode确定需要更改或调整本地的特征提取模型时,该第一指示信息用于指示sNode回传cNode的负载全局特征的调整量。
S505B:sNode根据sNode的负载局部特征和cNode的负载全局特征,训练得到回归 模型。
示例性地,sNode根据该辅助信息中的cNode的负载全局特征和sNode的负载局部特征进行特征联合,并将联合后的特征数据,作为训练模型的输入数据,得到回归模型(相当于上述本申请方案中的第一功能模型),该回归模型用于决策网络负载门限。
例如,cNode的负载全局特征属于3维度的特征数据,sNode的负载局部特征为2维度的特征数据,若cNode对应的3维度与sNode对应的2维度均不相同,那么可以将cNode的负载全局特征和sNode的负载局部特征进行串联式的特征联合,得到特征联合后的负载特征数据。若cNode对应的3维度与sNode对应的2维度存在维度相同,那么可以将cNode的负载全局特征和sNode的负载局部特征中相同维度的数据进行合并式的特征联合。
S506B:sNode根据该回归模型,确定网络负载门限值,以执行网络事件。
示例性地,sNode获取网络负载的样本数据包含特征数据样本,将该特征数据样本输入该回归模型,预测得到最优负载门限值,并基于该最优负载门限值,执行网络事件。
S507B:sNode对sNode的本地特征提取模型进行更新。
示例性地,参考图5A所示,sNode可以将上述基于回归模型预测得到的最优负载门限和真实负载门限值之间的误差值,作为该sNode的负载局部特征的调整值。进一步的,sNode根据该sNode的负载局部特征的调整值,对sNode的本地特征提取模型进行调整,得到更新后的本地特征提取模型。
更新后的本地特征提取模型可以用于下一轮的功能模型的训练,可参考上述步骤S501B-S507B执行,此处不再具体赘述。
S508B:sNode向cNode发送该cNode的负载全局特征的调整值。
若上述的第一辅助信息中包括第一指示信息用于指示sNode回传cNode的负载全局特征的调整值时,sNode将上述基于回归模型预测得到的最优负载门限值和真实负载门限值之间的误差值(loss),作为该cNode的负载全局特征的调整值,并向cNode发送该cNode的全局部特征数据的调整值。
S509B:cNode根据该cNode的负载全局特征的调整值,对cNode的本地特征提取模型进行更新。
上述步骤S508B-S509B为可选的步骤。
通过该实施例一,目标网络事件以网络负载均衡为例,需要预测负载门限值的sNode可以向cNode请求提供负载全局特征,将cNode提供的负载全局特征可以与sNode的网络负载局部特征进行特征联合,以辅助该sNode训练得到准确的回归模型,以用于预测得到最优的负载门限值。另外,sNode还可以基于该最优的负载门限值和真实负载门限值,对本地特征提取模型进行更新,同时当cNode指示sNode回传负载全局特征的调整值时,sNode还可以向cNode返回负载全局特征的调整值,以使得该cNode可基于该负载全局特征的调整值对本地特征提取模型进行更新,以保证下一轮提取数据特征的准确性更高。
实施例二:
在该实施例二中,参考图6A所示,目标网络事件以网络流量为例,UE(相当于本申请方案中的第一网元,如下层网元)具有或存储有本地流量局部数据,例如信道状态信息(channel state information,CSI)、流量类型、业务类型等。该UE需要训练得到回归模型,该回归模型用于预测网络流量。其中,基站xNodeB(相当于上述本申请方案中的第二网元, 如上层网元)具有或存储有流量全局数据,例如基站的负载信息、业务流入和/或流出量等,该基站xNodeB可以辅助UE进行流量的预测。参考图6B所示,该实施例方法的具体流程如下:
S601B:UE向xNodeB发送训练辅助请求。
相应的,xNodeB从UE接收该训练辅助请求(相当于上述本申请方案中的第一请求消息),该训练辅助请求用于请求xNodeB提供负载全局特征(相当于上述本申请方案中的第二特征数据)。
其中,该训练辅助请求中包括网络流量的信息,该网络流量的信息包括流量预测、隐藏特征参数,例如维度(相当于上述本申请方案中的目标网络事件的信息)。
S602B:xNodeB向UE发送回执消息。
相应的,该UE接收该回执消息(相当于上述本申请方案中的第一响应消息)。
若xNodeB确定接受该训练辅助请求时,该回执消息用于向UE通知该xNodeB接受该训练辅助请求,xNodeB继续执行下述步骤S603B。
若cxNodeB确定拒绝该训练辅助请求时,该回执消息用于向UE通知该xNodeB拒绝该训练辅助请求,xNodeB不再继续执行下述任何步骤。此时,UE可以只利用本地的流量局部数据进行训练得到回归模型,或者UE等待一段时间后再次向xNodeB发送训练辅助请求。
S603B:xNodeB确定流量全局特征。
示例性地,xNodeB确定流量全局特征时,可以包括:xNodeB先收集全局数据,如片区基站的流程数据、业务流入和/或流出量等;然后,xNodeB通过本地特征提取模型(相当于上述本申请方案中的第二特征提取模型)对该全局数据进行提取,得到流量全局特征(相当于上述本申请方案中的第二特征数据)。
同时,UE也可以参考该步骤S603B中的xNodeB确定流量全局特征的方式,确定流量局部特征。
示例性地,UE确定负载局部特征时,可以包括:UE先收集本地的局部数据,如CSI、历史流量、业务类型等;然后,UE通过本地特征提取模型(相当于上述本申请方案中的第一特征提取模型)对该局部数据进行提取,得到流量局部特征(相当于上述本申请方案中的第一特征数据)。
S604B:xNodeB向UE发送辅助信息。
相应的,该UE接收该辅助信息(相当于上述本申请方案中的第一信息),该辅助信息中包括xNodeB的流量全局特征和网络流量的信息。
在一种实施方式中,若xNodeB接受UE的训练辅助请求时,该辅助信息中可以包括上述的回执消息,xNodeB无需向UE分别发送上述的回执消息和该辅助信息,从而可以减少额外的传输开销和时延。
在一种实施方式中,该第一辅助信息中还包括第一指示信息,该第一指示信息用于指示UE是否回传xNodeB的流量全局特征的调整量。
示例性地,若xNodeB确定不需要更改或调整本地的特征提取模型(相当于本申请方案中的第二特征提取模型)时,该第一指示信息用于指示UE不回传xNodeB的流量全局特征的调整量。若xNodeB确定需要更改或调整本地的特征提取模型时,该第一指示信息用于指示UE回传xNodeB的流量全局特征的调整量。
S605B:UE根据UE的流量局部特征和xNodeB的流量全局特征,训练得到回归模型。
示例性地,UE根据该辅助信息中的xNodeB的流量全局特征和UE的流量局部特征进行特征联合,并将联合后的特征,作为训练模型的输入数据,得到回归模型(相当于上述本申请方案中的第一功能模型),该回归模型用于预测网络流量。
例如,xNodeB的流量全局特征属于3维度的特征数据,UE的流量局部特征为2维度的特征数据,若xNodeB对应的3维度与UE对应的2维度均不相同,那么可以将xNodeB的流量全局特征和UE的流量局部特征进行串联式的特征联合,得到特征联合后的流量特征。若xNodeB对应的3维度与UE对应的2维度存在维度相同,那么可以将xNodeB的流量全局特征和UE的流量局部特征中相同维度的特征进行合并式的特征联合。
S606B:UE根据该回归模型,确定网络流量,以执行网络事件。
示例性地,UE获取网络流量的样本数据包含特征数据样本,将该特征数据样本输入该回归模型,预测得到流量值,并基于预测的流量值,执行网络事件。
S607B:UE对UE的本地特征提取模型进行更新。
示例性地,参考图6A所示,UE可以将上述基于回归模型预测得到的流量值和真实流量值之间的误差值,作为该UE的流量局部特征的调整值。进一步的,UE根据该UE的流量局部特征的调整值,对sNode的本地特征提取模型进行调整,得到更新后的本地特征提取模型。
更新后的本地特征提取模型可以用于下一轮的功能模型的训练,可参考上述步骤S601B-S607B执行,此处不再具体赘述。
S608B:UE向xNodeB发送该xNodeB的全局部特征的调整值。
若上述的第一辅助信息中包括第一指示信息用于指示UE回传xNodeB的流量全局特征的调整值时,UE将上述基于回归模型预测得到的流量值和真实流量值之间的误差值(loss),作为该xNodeB的全局部特征的调整值,并向xNodeB发送该xNodeB的全局部特征的调整值。
S609B:xNodeB根据该xNodeB的全局部特征的调整值,对xNodeB的本地特征提取模型进行更新。
上述步骤S608B-S609B为可选的步骤。
通过该实施例二,目标网络事件以网络流量为例,需要预测网络流量值的UE可以向xNodeB请求提供网络流量的全局特征,将xNodeB提供的网络流量的全局特征可以与UE的网络流量的局部特征进行特征联合,以辅助该UE训练得到准确的回归模型,以用于预测得到流量值。另外,UE还可以基于该预测的流量值和真实流量值,对本地特征提取模型进行更新,同时当xNodeB指示UE回传流量全局特征的调整值时,UE还可以向xNodeB返回流量全局特征的调整值,以使得该xNodeB可基于该流量全局特征的调整值对本地特征提取模型进行更新,以保证下一轮提取数据特征的准确性更高。
实施例三:
在该实施例三中,参考图7A所示,目标网络事件以休眠节能为例,接入网RAN(相当于本申请方案中的第一网元,如下层网元)具有或存储有本地局部数据,例如该接入网RAN(如基站)的负载、忙闲时间段、用户分布的信息和/或数据等。该RAN需要训练得到分类模型,该分类模型用于预测RAN的工作状态(即是否进入休眠节能状态)。其中,核心网CN(相当于上述本申请方案中的第二网元,如上层网元)具有或存储有网络全局数据,例 如片区基站的服务数据、用户业务流量信息和/或数据等,该CN可以辅助RAN进行休眠节能状态的预测。参考图7B所示,该实施例方法的具体流程如下:
S701B:RAN向CN发送训练辅助请求。
相应的,CN从RAN接收该训练辅助请求(相当于上述本申请方案中的第一请求消息),该训练辅助请求用于请求CN提供网络全局特征(相当于上述本申请方案中的第二特征数据)。
其中,该训练辅助请求中包括休眠节能的信息,该休眠节能的信息包括节能(energy conservation,EC)、隐藏特征参数,例如维度(相当于上述本申请方案中的目标网络事件的信息)。
S702B:CN向RAN发送回执消息。
相应的,该RAN接收该回执消息(相当于上述本申请方案中的第一响应消息)。
若上述CN确定接受RAN的训练辅助请求时,该回执消息用于向RAN通知该CN接受该训练辅助请求,CN继续执行下述步骤S703B。
若上述CN确定拒绝RAN的训练辅助请求时,该回执消息用于向RAN通知该CN拒绝该训练辅助请求,CN不再继续执行下述任何步骤。此时,RAN可以利用本地的负载局部数据进行训练得到回归模型,或者RAN等待一段时间后再次向CN发送训练辅助请求。
S703B:CN确定网络全局特征。
示例性地,CN确定网络全局特征时,可以包括:CN先收集网络全局数据,如片区基站的服务情况、业务流入和/或流出量信息和/或数据等;然后,CN通过本地特征提取模型(相当于上述本申请方案中的第二特征提取模型)对该网络全局数据进行特征提取,得到网络全局特征(相当于上述本申请方案中的第二特征数据)。
同时,RAN也可以参考该步骤S703B中的CN确定网络全局特征的方式,确定局部特征。
示例性地,RAN确定局部特征时,可以包括:RAN先收集本地的局部数据,如基站的负载、忙闲时间段、用户分布信息和/或数据等;然后,RAN通过本地特征提取模型(相当于上述本申请方案中的第一特征提取模型)对该局部数据进行特征提取,得到局部特征(相当于上述本申请方案中的第一特征数据)。
S704B:CN向RAN发送辅助信息。
相应的,该RAN接收该辅助信息(相当于上述本申请方案中的第一信息),该辅助信息中包括CN的网络全局特征和休眠节能的信息。
在一种实施方式中,若CN接受RAN的训练辅助请求时,该辅助信息中可以包括上述的回执消息,CN无需向RAN分别发送上述的回执消息和该辅助信息,从而可以减少额外的传输开销和时延。
在一种实施方式中,该第一辅助信息中还包括第一指示信息,该第一指示信息用于指示RAN是否回传CN的网络全局特征的调整量。
示例性地,若CN确定不需要更改或调整本地的特征提取模型(相当于本申请方案中的第二特征提取模型)时,该第一指示信息用于指示RAN不回传CN的网络全局特征的调整量。若CN确定需要更改或调整本地的特征提取模型时,该第一指示信息用于指示RAN回传CN的网络全局特征的调整量。
S705B:RAN根据RAN的局部特征和CN的网络全局特征,训练得到分类模型。
示例性地,RAN根据该辅助信息中的CN的网络全局特征和RAN的局部特征进行特 征联合,并将联合后的特征,作为训练模型的输入数据,得到分类模型(相当于上述本申请方案中的第一功能模型),该分类模型用于预测是否进入休眠节能状态。
例如,CN的全局特征属于3维度的特征,RAN的局部特征为2维度的特征,若CN对应的3维度与RAN对应的2维度均不相同,那么可以将CN的全局特征和RAN的局部特征进行串联式的特征联合,得到特征联合后的特征。若CN对应的3维度与RAN对应的2维度相同,那么可以将CN的全局特征和RAN的局部特征进行合并式的特征联合。
S706B:RAN根据该分类模型,确定是否进入休眠节能状态。
示例性地,RAN获取网络的样本数据包含特征样本,将该特征样本输入该分类模型,预测进入休眠节能状态,进而RAN进入休眠节能状态。
S707B:RAN对RAN的本地特征提取模型进行更新。
示例性地,参考图7A所示,RAN可以根据上述基于分类模型预测得到RAN的工作状态,(如预测RAN进入休眠节能状态)和RAN真实的工作状态(如RAN未能进入休眠节能状态)的误差,对RAN的本地特征提取模型进行调整,得到更新后的本地特征提取模型。
RAN的更新后的本地特征提取模型可以用于下一轮的功能模型的训练,可参考上述步骤S701B-S707B执行,此处不再具体赘述。
S708B:RAN向CN发送该CN的网络全局特征的调整值。
若上述的第一辅助信息中包括第一指示信息用于指示RAN回传CN的网络全局特征的调整值时,RAN向CN发送该CN的网络全局特征的调整值,即上述基于分类模型预测得到RAN的工作状态和RAN真实的工作状态的误差。
S709B:CN根据该CN的全局特征的调整值,对CN的本地特征提取模型进行更新。
CN根据上述基于分类模型预测得到RAN的工作状态和RAN真实的工作状态的误差,对CN的本地特征提取模型进行更新。
上述步骤S708B-S709B为可选的步骤。
通过该实施例三,目标网络事件以休眠节能为例,RAN需要预测是否能进入休眠节能状态时,RAN可以向CN请求提供网络全局特征,然后RAN将CN提供的网络全局特征与RAN的本地局部特征进行特征联合,以辅助该RAN训练得到准确的分类模型,该分类模型可以用于预测RAN是否能进入休眠节能状态。另外,RAN还可以基于预测休眠节能状态和真实的工作状态,对本地特征提取模型进行更新,同时当CN指示RAN回传网络全局特征的调整值时,RAN还向CN返回网络全局特征的调整值,即基于分类模型预测得到RAN的工作状态和RAN真实的工作状态的误差,以使得该CN可根据该基于分类模型预测得到RAN的工作状态和RAN真实的工作状态的误差,对本地特征提取模型进行更新,以保证下一轮提取数据特征的准确性更高。
基于同一技术构思,本申请实施例提供一种通信装置,该通信装置可以用于执行上述方法实施例中由第一网元所执行的操作。该通信装置还可以为第一网元、第一网元的处理器、或芯片。该装置包括执行上述实施例中第一网元所描述的方法/操作/步骤/动作所一一对应的模块或单元,该模块或单元可以是硬件电路,也可是软件,也可以是硬件电路结合软件实现。该通信装置可以具有如图8所示的结构。
参考图8所示,该装置800包括通信单元801和处理单元802;其中,所述处理单元 802,用于确定目标网络事件的第一特征数据;所述通信单元801,用于从第二网元接收第一信息,所述第一信息中包括所述目标网络事件的信息、所述目标网络事件的第二特征数据;所述处理单元802,还用于根据第一功能模型,执行所述目标网络事件,所述第一功能模型是根据所述第一特征数据和所述第二特征数据得到的。
可选地,该通信装置800还可包括存储单元(图8中未画出),所述存储单元可以用于存储执行该第一网元所描述的方法/操作/步骤/动作的计算机程序和/或指令和/或信息等。
在一种可能的实施方式中,所述通信单元,还用于:从第二网元接收第一信息之前,向所述第二网元发送第一请求消息,所述第一请求消息用于请求所述目标网络事件的第二特征数据。
在一种可能的实施方式中,所述第一请求消息中包括所述目标网络事件的信息,其中,所述目标网络事件的信息包括所述目标网络事件的标识信息,以及还包括所述目标网络事件的特征数据的目标维度、响应时延、数据精度、所述目标网络事件的服务质量QoS中的一项或多项。
在一种可能的实施方式中,所述通信单元801,还用于:从所述第二网元接收第一响应消息,所述第一响应消息用于向所述第一网元通知所述第二网元提供所述目标网络事件的第二特征数据。
在一种可能的实施方式中,所述第一信息还包括所述第一响应消息。
在一种可能的实施方式中,所述处理单元802,在确定目标网络事件的第一特征数据时,具体用于:通过所述通信单元801获取所述目标网络事件的第一数据集;利用第一特征提取模型,对所述第一数据集进行特征提取,得到所述第一特征数据,所述第一特征提取模型用于所述第一网元提取所述目标网络事件的特征数据。
在一种可能的实施方式中,所述第一信息中还包括第一指示,所述第一指示用于指示所述第一网元是否回传所述第二特征数据的调整量。
在一种可能的实施方式中,所述处理单元802,还用于:通过所述通信单元801获取所述目标网络事件的样本数据,所述样本数据中包含特征数据样本和所述目标网络事件的真实值;将所述特征数据样本输入所述第一功能模型,得到所述目标网络事件的输出值;根据所述目标网络事件的输出值和所述目标网络事件的真实值,得所述第二特征数据的调整量,所述第二特征数据的调整量用于调整所述第二网元的第二特征提取模型的输出层。
在一种可能的实施方式中,所述第一指示用于指示所述第一网元回传所述第二特征数据的调整量时,所述通信单元801还用于:根据所述第一指示,向所述第二网元发送所述第二特征数据的调整量。
基于同一技术构思,本申请实施例提供一种通信装置,该通信装置可以用于执行上述方法实施例中由第二网元所执行的操作。该通信装置还可以为第二网元、第二网元的处理器、或芯片。该装置包括执行上述实施例中第二网元所描述的方法/操作/步骤/动作所一一对应的模块或单元,该模块或单元可以是硬件电路,也可是软件,也可以是硬件电路结合软件实现。该通信装置也可以具有如图8所示的结构。
参考图8所示,该装置800包括通信单元801和处理单元802;其中,所述处理单元802,用于确定目标网络事件的第二特征数据;所述通信单元801,用于向第一网元发送第一信息,所述第一信息中包括所述目标网络事件的信息和所述目标网络事件的第二特征数 据。
可选地,该通信装置800还可包括存储单元(图8中未画出),所述存储单元可以用于存储执行该第二网元所描述的方法/操作/步骤/动作的计算机程序和/或指令和/或信息等。
在一种可能的实施方式中,所述通信单元801,还用于:在所述处理单元802确定目标网络事件的第二特征数据之前,从所述第一网元接收第一请求消息,所述第一请求消息用于请求所述目标网络事件的第二特征数据。
在一种可能的实施方式中,所述第一请求消息中包括所述目标网络事件的信息;其中,所述目标网络事件的信息包括所述目标网络事件的标识信息,以及还包括所述目标网络事件的特征数据的目标维度、响应时延、数据精度、所述目标网络事件的服务质量QoS中的一项或多项。
在一种可能的实施方式中,所述通信单元801,还用于:向所述第一网元发送第一响应消息,所述第一响应消息用于向所述第一网元通知所述第二网元提供所述目标网络事件的第二特征数据。
在一种可能的实施方式中,所述第一信息还包括所述第一响应消息。
在一种可能的实施方式中,所述处理单元802,在确定目标网络事件的第二特征数据时,具体用于:通过所述通信单元801获取所述目标网络事件的第二数据集;利用第二特征提取模型,对所述第二数据集进行特征提取,得到所述第二特征数据,所述第二特征提取模型用于所述第二网元提取所述目标网络事件的特征数据。
在一种可能的实施方式中,所述第一信息中还包括第一指示,所述第一指示用于指示所述第一网元是否回传所述第二特征数据的调整量。
在一种可能的实施方式中,所述第一指示用于指示所述第一网元回传所述第二特征数据的调整量时,所述通信单元801还用于:从所述第一网元接收所述第二特征数据的调整量。
在一种可能的实施方式中,所述处理单元802,还用于:根据所述第二特征数据的调整量,对所述第二网元的第二特征提取模型的输出层进行调整。
基于同一发明构思,本申请实施例还提供了一种通信设备,该通信设备采用图4A-图4B、图5A-图5B、图6A-图6B、图7A-图7B对应的实施例提供的方法中第一网元执行的步骤,可以是与图8所示的通信装置800相同的设备。参阅图9所示,通信设备900包括:收发器901、处理器902和存储器903。其中,收发器901、处理器902和存储器903通过总线连接904,以便实现数据交换。可选的,处理器902和存储器903可以集成在一起。
其中,通信线路904可以是外设部件互连标准(peripheral component interconnect,简称PCI)总线或扩展工业标准结构(extended industry standard architecture,简称EISA)总线等。所述通信线路904可以分为地址总线、数据总线、控制总线等。为便于表示,图9中仅用一条粗线表示,但并不表示仅有一根总线或一种类型的总线。
应理解,其中,该通信设备900中收发器901包括上述通信装置800中的通信单元801的发送和/或接收功能。该收发器901用于支持通信装置800与上述实施例中的第二网元之间收发信息、数据等。存储器903用于通信设备900的程序代码和数据。处理器902用于调用存储器903中存储的程序代码和数据,执行图4A-图4B、图5A-图5B、图6A-图6B、图7A-图7B所示方法中涉及第一网元的处理过程和/或用于本申请所描述的技术的其他过 程。
此外,通信设备900还可以包括其他接口,例如光纤链路接口、以太网接口、微波链路接口、铜线接口等,用以实现通信设备900与第二网元的交互。
可选的,处理器902可以是中央处理器、ASIC、FPGA或CPLD。
需要说明的是,图9示出的通信设备900中仅包含一个收发器901、一个处理器902和一个存储器903。实际实现时,收发器901、处理器902和存储器903的数量可以为一个,也可以为多个。
同样说明的是,图9示出的通信设备900也可以实现图4A-图4B、图5A-图5B、图6A-图6B、图7A-图7B对应的实施例提供的方法中第二网元执行的方法。图9示出的通信设备900也可以是与上述图8所示的通信装置800相同的设备。因此,通信设备900未详细描述的实现方式可以参照图4A-图4B、图5A-图5B、图6A-图6B、图7A-图7B对应的实施例提供的方法中的相关描述或者参考上述图9所示的通信设备900中的相关描述。此处不再具体赘述。
图10为本申请实施例提供的一种芯片的装置结构示意图。该芯片1000包括接口电路1001和一个或多个处理器1002。可选的,所述芯片1000还可以包含总线。其中:
处理器1002可能是一种集成电路芯片,具有信号的处理能力。在实现过程中,上述眼球跟踪方法的各步骤可以通过处理器1002中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器1002可以是通用处理器、数字通信器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或者其它可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。
接口电路1001可以用于数据、指令或者信息的发送或者接收,处理器1002可以利用接口电路1001接收的数据、指令或者其它信息,进行加工,可以将加工完成信息通过接口电路1001发送出去。
可选的,芯片还包括存储器1003,存储器1003可以包括只读存储器和随机存取存储器,并向处理器提供操作指令和数据。存储器1003的一部分还可以包括非易失性随机存取存储器(NVRAM)。
可选的,存储器存储了可执行软件模块或者数据结构,处理器可以通过调用存储器存储的操作指令(该操作指令可存储在操作***中),执行相应的操作。
可选的,芯片可以使用在本申请实施例涉及的第一网元中。可选的,接口电路1001可用于输出处理器1002的执行结果。关于本申请的一个或多个实施例提供的通信方法可参考前述各个实施例,这里不再赘述。
需要说明的,接口电路1001、处理器1002各自对应的功能既可以通过硬件设计实现,也可以通过软件设计来实现,还可以通过软硬件结合的方式来实现,这里不作限制。
基于与上述方法实施例相同构思,本申请实施例还提供了一种计算机可读存储介质,其上存储有一些指令,这些指令被计算机调用执行时,可以使得计算机完成上述方法实施例、方法实施例的任意一种可能的设计中所涉及的方法。本申请实施例中,对计算机可读存储介质不做限定,例如,可以是RAM(网络设备random-access memory,随机存取存储器)、ROM(read-only memory,只读存储器)等。
基于与上述方法实施例相同构思,本申请还提供一种计算机程序产品,该计算机程序产品在被计算机调用执行时可以完成方法实施例以及上述方法实施例任意可能的设计中所涉及的方法。
基于与上述方法实施例相同构思,本申请还提供一种芯片,该芯片可以包括处理器以及接口电路,用于完成上述方法实施例、方法实施例的任意一种可能的实现方式中所涉及的方法,其中,“耦合”是指两个部件彼此直接或间接地结合,这种结合可以是固定的或可移动性的,这种结合可以允许流动液、电、电信号或其它类型信号在两个部件之间进行通信。
需要注意的是,本申请实施例涉及的至少一个,包括一个或者多个;其中,多个是指大于或者等于两个。另外,需要理解的是,在本申请的描述中,“第一”、“第二”等词汇,仅用于区分描述的目的,而不能理解为指示或暗示相对重要性,也不能理解为指示或暗示顺序。
以下实施例中所使用的术语只是为了描述特定实施例的目的,而并非旨在作为对本申请的限制。如在本申请的说明书和所附权利要求书中所使用的那样,单数表达形式“一种”、“所述”、“上述”、“该”和“这一”旨在也包括例如“一个或多个”这种表达形式,除非其上下文中明确地有相反指示。还应当理解,在本申请实施例中,“一个或多个”是指一个或两个以上(包含两个);“和/或”,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的情况,其中A、B可以是单数或者复数。字符“/”一般表示前后关联对象是一种“或”的关系。
在本说明书中描述的参考“一个实施例”或“一些实施例”等意味着在本申请的一个或多个实施例中包括结合该实施例描述的特定特征、结构或特点。由此,在本说明书中的不同之处出现的语句“在一个实施例中”、“在一些实施例中”、“在其他一些实施例中”、“在另外一些实施例中”等不是必然都参考相同的实施例,而是意味着“一个或多个但不是所有的实施例”,除非是以其他方式另外特别强调。术语“包括”、“包含”、“具有”及它们的变形都意味着“包括但不限于”,除非是以其他方式另外特别强调。
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,另外,在本申请各个实施例中的各功能模块可以集成在一个处理器中,也可以是单独物理存在,也可以两个或两个以上模块集成在一个模块中。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。
本申请实施例提供了一种计算机可读存储介质,存储有计算机程序,该计算机程序包括用于执行上述方法实施例的指令。
本申请实施例提供了一种包含指令的计算机程序产品,当其在计算机上运行时,使得计算机执行上述方法实施例。
需要注意的是,为描述方便和简洁,上述提供的任一种通信装置中相关内容的解释及有益效果均可参考上文提供的对应的通信方法实施例的效果,此处不再赘述。
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,另外,在本申请各个实施例中的各功能模块可以集成在一个处理器中,也可以是单独物理存在,也可以两个或两个以上模块集成在一个模块中。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。
本领域内的技术人员应明白,本申请的实施例可提供为方法、***、或计算机程序产 品。因此,本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且,本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请的方法、设备(***)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
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保护范围。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 (41)
- 一种通信方法,其特征在于,包括:第一网元确定目标网络事件的第一特征数据;所述第一网元从第二网元接收第一信息,所述第一信息中包括所述目标网络事件的信息、所述目标网络事件的第二特征数据;所述第一网元根据第一功能模型,执行所述目标网络事件,所述第一功能模型是所述第一网元根据所述第一特征数据和所述第二特征数据得到的。
- 根据权利要求1所述的方法,其特征在于,所述第一网元从第二网元接收第一信息之前,还包括:所述第一网元向所述第二网元发送第一请求消息,所述第一请求消息用于请求所述目标网络事件的第二特征数据。
- 根据权利要求2所述的方法,其特征在于,所述第一请求消息中包括所述目标网络事件的信息;其中,所述目标网络事件的信息包括所述目标网络事件的标识信息,以及还包括所述目标网络事件的特征数据的目标维度、响应时延、数据精度、所述目标网络事件的服务质量QoS中的一项或多项。
- 根据权利要求2或3所述的方法,其特征在于,所述方法还包括:所述第一网元从所述第二网元接收第一响应消息,所述第一响应消息用于向所述第一网元通知所述第二网元提供所述目标网络事件的第二特征数据。
- 根据权利要求4所述的方法,其特征在于,所述第一信息还包括所述第一响应消息。
- 根据权利要求1至5任一项所述的方法,其特征在于,所述第一网元确定目标网络事件的第一特征数据,包括:所述第一网元获取所述目标网络事件的第一数据集;所述第一网元利用第一特征提取模型,对所述第一数据集进行特征提取,得到所述第一特征数据,所述第一特征提取模型用于所述第一网元提取所述目标网络事件的特征数据。
- 根据权利要求1至6任一项所述的方法,其特征在于,所述第一信息中还包括第一指示,所述第一指示用于指示所述第一网元是否回传所述第二特征数据的调整量。
- 根据权利要求1至7任一项所述的方法,其特征在于,所述方法还包括:所述第一网元获取所述目标网络事件的样本数据,所述样本数据中包含特征数据样本和所述目标网络事件的真实值;所述第一网元将所述特征数据样本输入所述第一功能模型,得到所述目标网络事件的输出值;所述第一网元根据所述目标网络事件的输出值和所述目标网络事件的真实值,得所述第二特征数据的调整量,所述第二特征数据的调整量用于调整所述第二网元的第二特征提取模型的输出层。
- 根据权利要求7或8所述的方法,其特征在于,所述第一指示用于指示所述第一网元回传所述第二特征数据的调整量时,所述方法还包括:所述第一网元根据所述第一指示,向所述第二网元发送所述第二特征数据的调整量。
- 一种通信方法,其特征在于,包括:第二网元确定目标网络事件的第二特征数据;所述第二网元向第一网元发送第一信息,所述第一信息中包括所述目标网络事件的信息和所述目标网络事件的第二特征数据。
- 根据权利要求10所述的方法,其特征在于,所述第二网元确定目标网络事件的第二特征数据之前,还包括:所述第二网元从所述第一网元接收第一请求消息,所述第一请求消息用于请求所述目标网络事件的第二特征数据。
- 根据权利要求11所述的方法,其特征在于,所述第一请求消息中包括所述目标网络事件的信息;其中,所述目标网络事件的信息包括所述目标网络事件的标识信息,以及还包括所述目标网络事件的特征数据的目标维度、响应时延、数据精度、所述目标网络事件的服务质量QoS中的一项或多项。
- 根据权利要求11或12所述的方法,其特征在于,所述方法还包括:所述第二网元向所述第一网元发送第一响应消息,所述第一响应消息用于向所述第一网元通知所述第二网元提供所述目标网络事件的第二特征数据。
- 根据权利要求13所述的方法,其特征在于,所述第一信息还包括所述第一响应消息。
- 根据权利要求10至14任一项所述的方法,其特征在于,所述第二网元确定目标网络事件的第二特征数据,包括:所述第二网元获取所述目标网络事件的第二数据集;所述第二网元利用第二特征提取模型,对所述第二数据集进行特征提取,得到所述第二特征数据,所述第二特征提取模型用于所述第二网元提取所述目标网络事件的特征数据。
- 根据权利要求10至15任一项所述的方法,其特征在于,所述第一信息中还包括第一指示,所述第一指示用于指示所述第一网元是否回传所述第二特征数据的调整量。
- 根据权利要求16所述的方法,其特征在于,所述第一指示用于指示所述第一网元回传所述第二特征数据的调整量时,所述方法还包括:所述第二网元从所述第一网元接收所述第二特征数据的调整量。
- 根据权利要求17所述的方法,其特征在于,所述方法还包括:所述第二网元根据所述第二特征数据的调整量,对所述第二网元的第二特征提取模型的输出层进行调整。
- 一种通信装置,其特征在于,包括通信单元和处理单元;所述处理单元,用于确定目标网络事件的第一特征数据;所述通信单元,用于从第二网元接收第一信息,所述第一信息中包括所述目标网络事件的信息、所述目标网络事件的第二特征数据;所述处理单元,用于根据第一功能模型,执行所述目标网络事件,所述第一功能模型是根据所述第一特征数据和所述第二特征数据得到的。
- 根据权利要求19所述的装置,其特征在于,所述第一网元从第二网元接收第一信息之前,还包括:所述第一网元向所述第二网元发送第一请求消息,所述第一请求消息用于请求所述目标网络事件的第二特征数据。
- 根据权利要求20所述的装置,其特征在于,所述第一请求消息中包括所述目标网络事件的信息;其中,所述目标网络事件的信息包括所述目标网络事件的标识信息,以及 还包括所述目标网络事件的特征数据的目标维度、响应时延、数据精度、所述目标网络事件的服务质量QoS中的一项或多项。
- 根据权利要求20或21所述的装置,其特征在于,所述通信单元还用于:从所述第二网元接收第一响应消息,所述第一响应消息用于向所述第一网元通知所述第二网元提供所述目标网络事件的第二特征数据。
- 根据权利要求22所述的装置,其特征在于,所述第一信息还包括所述第一响应消息。
- 根据权利要求19至23任一项所述的装置,其特征在于,所述处理单元在确定目标网络事件的第一特征数据时,具体用于:通过所述通信单元获取所述目标网络事件的第一数据集;利用第一特征提取模型,对所述第一数据集进行特征提取,得到所述第一特征数据,所述第一特征提取模型用于所述第一网元提取所述目标网络事件的特征数据。
- 根据权利要求19至24任一项所述的装置,其特征在于,所述第一信息中还包括第一指示,所述第一指示用于指示所述第一网元是否回传所述第二特征数据的调整量。
- 根据权利要求19至25任一项所述的装置,其特征在于,所述通信单元还用于获取所述目标网络事件的样本数据,所述样本数据中包含特征数据样本和所述目标网络事件的真实值;所述处理单元,还用于将所述特征数据样本输入所述第一功能模型,得到所述目标网络事件的输出值;根据所述目标网络事件的输出值和所述目标网络事件的真实值,得所述第二特征数据的调整量,所述第二特征数据的调整量用于调整所述第二网元的第二特征提取模型的输出层。
- 根据权利要求25或26所述的装置,其特征在于,所述第一指示用于指示所述第一网元回传所述第二特征数据的调整量时,所述通信单元还用于:根据所述第一指示,向所述第二网元发送所述第二特征数据的调整量。
- 一种通信装置,其特征在于,包括通信单元和处理单元;所述处理单元,用于确定目标网络事件的第二特征数据;所述通信单元,用于向第一网元发送第一信息,所述第一信息中包括所述目标网络事件的信息和所述目标网络事件的第二特征数据。
- 根据权利要求28所述的装置,其特征在于,所述通信单元,还用于:在所述处理单元确定目标网络事件的第二特征数据之前,从所述第一网元接收第一请求消息,所述第一请求消息用于请求所述目标网络事件的第二特征数据。
- 根据权利要求29所述的装置,其特征在于,所述第一请求消息中包括所述目标网络事件的信息;其中,所述目标网络事件的信息包括所述目标网络事件的标识信息,以及还包括所述目标网络事件的特征数据的目标维度、响应时延、数据精度、所述目标网络事件的服务质量QoS中的一项或多项。
- 根据权利要求29或30所述的装置,其特征在于,所述通信单元,还用于:向所述第一网元发送第一响应消息,所述第一响应消息用于向所述第一网元通知所述第二网元提供所述目标网络事件的第二特征数据。
- 根据权利要求31所述的装置,其特征在于,所述第一信息还包括所述第一响应消息。
- 根据权利要求28至32任一项所述的装置,其特征在于,所述处理单元,在确定目标网络事件的第二特征数据时,具体用于:通过所述通信单元获取所述目标网络事件的第二数据集;利用第二特征提取模型,对所述第二数据集进行特征提取,得到所述第二特征数据,所述第二特征提取模型用于所述第二网元提取所述目标网络事件的特征数据。
- 根据权利要求28至33任一项所述的装置,其特征在于,所述第一信息中还包括第一指示,所述第一指示用于指示所述第一网元是否回传所述第二特征数据的调整量。
- 根据权利要求34所述的装置,其特征在于,所述第一指示用于指示所述第一网元回传所述第二特征数据的调整量时,所述通信单元还用于:从所述第一网元接收所述第二特征数据的调整量。
- 根据权利要求35所述的装置,其特征在于,所述处理单元还用于:根据所述第二特征数据的调整量,对所述第二网元的第二特征提取模型的输出层进行调整。
- 一种通信装置,其特征在于,包括处理器和接口电路,所述接口电路用于接收来自所述通信装置之外的其它通信装置的信号并传输至所述处理器或将来自所述处理器的信号发送给所述通信装置之外的其它通信装置,所述处理器通过逻辑电路或执行代码指令用于实现如权利要求1至9中任一项所述的方法,或用于实现如权利要求10至18中任一项所述的方法。
- 一种计算机程序产品,其特征在于,包括计算机程序,当所述计算机程序被通信装置执行时,实现如权利要求1至18中任一项所述的方法。
- 一种计算机可读存储介质,其特征在于,所述存储介质中存储有计算机可读程序或指令,当所述计算机程序或指令被通信装置执行时,实现如权利要求1至18中任一项所述方法。
- 一种通信***,其特征在于,包括用于执行权利要求1至9中任一项所述方法的第一网元和用于执行权利要求10至18中任一项所述方法的第二网元。
- 一种芯片,其特征在于,所述芯片与存储器耦合,所述芯片读取所述存储器中存储的计算机程序,执行权利要求1-18任一项所述的方法。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2022/103941 WO2024007156A1 (zh) | 2022-07-05 | 2022-07-05 | 一种通信方法和装置 |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
PCT/CN2022/103941 WO2024007156A1 (zh) | 2022-07-05 | 2022-07-05 | 一种通信方法和装置 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2024007156A1 true WO2024007156A1 (zh) | 2024-01-11 |
Family
ID=89454732
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2022/103941 WO2024007156A1 (zh) | 2022-07-05 | 2022-07-05 | 一种通信方法和装置 |
Country Status (1)
Country | Link |
---|---|
WO (1) | WO2024007156A1 (zh) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180089587A1 (en) * | 2016-09-26 | 2018-03-29 | Google Inc. | Systems and Methods for Communication Efficient Distributed Mean Estimation |
CN110430068A (zh) * | 2018-04-28 | 2019-11-08 | 华为技术有限公司 | 一种特征工程编排方法及装置 |
CN113839797A (zh) * | 2020-06-23 | 2021-12-24 | 华为技术有限公司 | 数据处理方法和装置 |
WO2022061940A1 (zh) * | 2020-09-28 | 2022-03-31 | 华为技术有限公司 | 一种模型数据传输方法及通信装置 |
-
2022
- 2022-07-05 WO PCT/CN2022/103941 patent/WO2024007156A1/zh unknown
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20180089587A1 (en) * | 2016-09-26 | 2018-03-29 | Google Inc. | Systems and Methods for Communication Efficient Distributed Mean Estimation |
CN110430068A (zh) * | 2018-04-28 | 2019-11-08 | 华为技术有限公司 | 一种特征工程编排方法及装置 |
CN113839797A (zh) * | 2020-06-23 | 2021-12-24 | 华为技术有限公司 | 数据处理方法和装置 |
WO2021258986A1 (zh) * | 2020-06-23 | 2021-12-30 | 华为技术有限公司 | 数据处理方法和装置 |
WO2022061940A1 (zh) * | 2020-09-28 | 2022-03-31 | 华为技术有限公司 | 一种模型数据传输方法及通信装置 |
Non-Patent Citations (1)
Title |
---|
"3rd Generation Partnership Project; Technical Specification Group Services and System Aspects; Study on enablers for network automation for the 5G System (5GS); Phase 2 (Release 17)", 3GPP TR 23.700-91, no. V0.4.0, 1 July 2020 (2020-07-01), pages 1 - 186, XP051925861 * |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US20230209390A1 (en) | Intelligent Radio Access Network | |
US20230179490A1 (en) | Artificial intelligence-based communication method and communication apparatus | |
US11601878B2 (en) | Methods and systems for intelligent AMF assignment to minimize re-direction | |
EP4132066A1 (en) | Communication method, apparatus and system | |
Zhou et al. | Software-defined hyper-cellular architecture for green and elastic wireless access | |
TWI745851B (zh) | 藉由聯合網路及雲端資源管理之服務傳遞 | |
US20220338106A1 (en) | Slice control method and apparatus | |
WO2022041285A1 (zh) | 一种模型数据传输方法及通信装置 | |
CN108777883A (zh) | 一种无线接入网络切片的构建方法和装置 | |
WO2022226713A1 (zh) | 策略确定的方法和装置 | |
US20230189057A1 (en) | Service traffic steering method and apparatus | |
WO2023069534A1 (en) | Using ai-based models for network energy savings | |
US20240057139A1 (en) | Optimization of deterministic and non-deterministic traffic in radio-access network (ran) | |
WO2024007156A1 (zh) | 一种通信方法和装置 | |
WO2022213897A1 (zh) | 通信方法及通信装置 | |
WO2022042483A1 (zh) | 信息处理方法、装置、服务器及车联网终端 | |
US11595286B2 (en) | Systems and methods for modifying parameters of a wireless network based on granular energy efficiency metrics | |
KR20230137998A (ko) | Af 세션에 대한 외부 매개변수 프로비저닝을 위한 새로운 방법 | |
WO2016201860A1 (zh) | 数据传输方法及装置 | |
WO2024140290A1 (zh) | 一种计算任务的处理方法及相关装置 | |
WO2023201733A1 (zh) | 无线通信的方法和设备 | |
WO2023078183A1 (zh) | 一种数据收集方法及通信装置 | |
WO2023246267A1 (zh) | 通信方法、通信装置和*** | |
WO2024093739A1 (zh) | 一种通信方法及装置 | |
WO2023185452A1 (zh) | 一种通信方法和通信装置 |
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: 22949744 Country of ref document: EP Kind code of ref document: A1 |