WO2024066930A1 - 一种通信方法及装置 - Google Patents

一种通信方法及装置 Download PDF

Info

Publication number
WO2024066930A1
WO2024066930A1 PCT/CN2023/116672 CN2023116672W WO2024066930A1 WO 2024066930 A1 WO2024066930 A1 WO 2024066930A1 CN 2023116672 W CN2023116672 W CN 2023116672W WO 2024066930 A1 WO2024066930 A1 WO 2024066930A1
Authority
WO
WIPO (PCT)
Prior art keywords
strategy
network element
policy
model
information
Prior art date
Application number
PCT/CN2023/116672
Other languages
English (en)
French (fr)
Inventor
曾宇
耿婷婷
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2024066930A1 publication Critical patent/WO2024066930A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/06—Testing, supervising or monitoring using simulated traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/08—Testing, supervising or monitoring using real traffic
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/10—Scheduling measurement reports ; Arrangements for measurement reports

Definitions

  • the embodiments of the present application relate to the field of communication technology, and in particular, to a communication method and device.
  • AI Artificial intelligence
  • the training process of the AI model can be performed on core network devices or access network devices.
  • Access network devices can use the trained AI model to obtain energy-saving strategies, load balancing strategies, etc., and execute the corresponding strategies.
  • each access network device independently determines and executes strategies based on the AI model, which limits the improvement of network performance.
  • the embodiments of the present application provide a communication method and device for improving network performance.
  • a communication method wherein the execution subject of the method may be a first network element, or a component applied in the first network element, such as a chip, a processor, etc.
  • the following description is made by taking the execution subject being the first network element as an example.
  • the first network element obtains a first measurement quantity to be input, and obtains a first AI model.
  • the first network element inputs the first measurement quantity into the first AI model to obtain a first strategy output by the first AI model.
  • the first network element executes the first strategy.
  • the first network element determines that an abnormality occurs in the execution of the first strategy
  • the first network element sends an indication message to the second network element, wherein the indication message is used to indicate that an abnormality occurs in the execution of the first strategy.
  • the first network element executes the first strategy determined based on the first AI model, it can detect the execution status and notify the second network element. In this way, the network elements can perceive the execution status of other network elements executing the strategy based on the AI model, which can improve network performance to a certain extent. In particular, when it is determined that an abnormality occurs in the execution of the first strategy, the situation is notified to other network elements, which can facilitate other network elements to respond based on the abnormality and improve network performance.
  • the first network element is a first access network device
  • the second network element is a core network element, or an operation, administration and maintenance OAM network element
  • the first network element is a first terminal device
  • the second network element is an access network device
  • the first network element is a first access network device
  • the second network element is a second access network device
  • the first network element is a first terminal device
  • the second network element is a second terminal device.
  • the type of the first strategy includes any one of the following: energy saving strategy, load balancing strategy, mobility optimization strategy, channel state information reference signal CSI-RS feedback enhancement strategy, beam management enhancement strategy, and positioning accuracy enhancement strategy.
  • the indication information of the abnormality in the execution of the first policy includes one or more of the following: an indication of the abnormality in the execution, the time of the abnormality in the execution, the reason for the abnormality in the execution, a measurement parameter used to determine the abnormality in the policy, a configuration parameter used to determine the abnormality in the policy, the effective time of the configuration parameter, the correction method expected to be adopted by the first network element for the abnormality in the execution, the information required for the correction method expected to be adopted by the first network element for the abnormality in the execution, the identifier of the first AI model, the parameters of the first AI model, the identifier of the first policy, the first policy or the effective time of the first policy.
  • the second network element can know the relevant situation of the abnormality more clearly so as to make more reasonable countermeasures.
  • the first network element receives first information from the second network element, the first information is used to indicate a correction method allowed for execution anomalies; the first network element corrects the anomalies generated by the first policy based on the correction method.
  • the second network element can serve as a manager or decision maker to notify the first network element of the correction method allowed to be adopted, and the first network element makes reasonable corrections to the anomalies to further improve network performance.
  • the correction method allowed to be adopted includes: executing a second policy, the second policy is a policy obtained based on non-AI, and the second policy is of the same type as the first policy; it can be understood that the correction method allowed to be adopted is to execute a traditional policy (the second policy is a traditional policy) rather than a policy obtained based on AI. Then the first network element corrects the anomaly generated by the first policy based on the correction method, including: the first network element skips the first policy and executes the second policy.
  • the correction method allowed to be adopted includes: executing a third strategy, the third strategy is a strategy determined based on the second AI model, and the third strategy is of the same type as the first strategy; it can be understood that the correction method allowed to be adopted is to adopt a new AI model (the second model is the new AI model) to determine a new AI strategy (the third strategy is the new AI strategy), and the third strategy can be obtained based on the following method: input the second measurement quantity to be input into the second AI model to obtain the third strategy output by the second AI model.
  • the first network element corrects the anomaly generated by the first strategy based on the correction method, including: the first network element skips the first strategy and executes the third strategy.
  • the correction method allowed to be adopted includes: executing a fourth policy, the fourth policy is a policy executed before executing the first policy, and the fourth policy is of the same type as the first policy; it can be understood that the correction method allowed to be adopted is to roll back to the previously executed policy (the fourth policy is the previously executed policy). Then the first network element corrects the anomaly generated by the first policy based on the correction method, including: the first network element skips the first policy and executes the fourth policy.
  • the correction method allowed to be adopted includes: not executing a policy of the same type as the first policy. It can be understood that the correction method allowed to be adopted is to exit the first policy. Then the first network element corrects the anomaly generated by the first policy based on the correction method, including: the first network element exits the first policy.
  • the first information includes at least one of the following: an identifier of the third strategy, the third strategy, an identifier of the second AI model, a parameter of the second AI model, or performance information of the third network element; wherein the performance information of the third network element is used to determine the third strategy.
  • the first network element corrects the anomaly caused by the first policy based on the correction method, including: when the first information includes the identifier of the third policy, executing the third policy indicated by the identifier of the third policy to correct the anomaly caused by the first policy.
  • the first network element corrects the anomaly generated by the first policy based on the correction method, including: when the first information includes the third policy, executing the third policy to correct the anomaly generated by the first policy.
  • the first network element corrects the anomaly caused by the first strategy based on the correction method, including: when the first information includes the identifier of the second AI model, the first acquisition obtains the second AI model based on the identifier, inputs the second measurement amount into the second AI model, obtains the third strategy, and executes the third strategy to correct the anomaly caused by the first strategy, wherein the second measurement amount is the same as the first measurement amount, or the measurement time of the second measurement amount is not earlier than the measurement time of the first measurement amount.
  • the first network element corrects the anomaly caused by the first strategy based on the correction method, including: when the first information includes parameters of the second AI model, inputting the second measurement amount into the second AI model corresponding to the parameter, obtaining the third strategy, and executing the third strategy to correct the anomaly caused by the first strategy, wherein the second measurement amount is the same as the first measurement amount, or the measurement time of the second measurement amount is not earlier than the measurement time of the first measurement amount.
  • the first network element corrects the anomaly generated by the first strategy based on the correction method, including: when the first information includes performance information of the third network element, correcting the first AI model based on the performance information to obtain a second AI model; inputting a third measurement amount into the second AI model to obtain the third strategy; and executing the third strategy to correct the anomaly generated by the first strategy, wherein the third measurement amount is the same as the first measurement amount, or the measurement time of the third measurement amount is not earlier than the measurement time of the first measurement amount.
  • a communication method is provided, and the execution subject of the method can be a second network element, or a component applied in the second network element, such as a chip, a processor, etc.
  • the execution subject is the second network element.
  • the second network element receives indication information from the first network element, and the indication information is used to indicate that an exception occurs when the first network element executes a first strategy, and the first strategy is a strategy output by the first AI model after the first network element inputs a first measurement quantity to be input into the first AI model; the second network element sends second information to the third network element, and the second information is used to indicate that an exception occurs when the first strategy is executed or an exception occurs when a first type of strategy output by the first AI model is executed, and the type of the first strategy is the first type.
  • Network elements can sense the execution status of other network elements executing strategies based on AI models, which can improve network performance to a certain extent. Specifically, when the first network element determines that an abnormality occurs in the execution of the first strategy, it notifies the second network element of the situation, and the second network element responds based on the abnormality, that is, notifies the third network element that an abnormality occurs when the first strategy is executed or an abnormality occurs when the first type of strategy obtained based on AI is executed, so that the third network element can respond and improve network performance.
  • the first network element is a first access network device
  • the second network element is a core network element or an operation, administration and maintenance (OAM) network element
  • the third network element is a second access network device
  • the first network element is a first terminal device
  • the second network element is an access network device.
  • the third network element is a second terminal device
  • the first network element is a terminal device
  • the second network element is a first access network device
  • the third network element is a second access network device.
  • the type of the first strategy includes any one of the following: energy saving strategy, load balancing strategy, mobility optimization strategy, channel state information reference signal CSI-RS feedback enhancement strategy, beam management enhancement strategy, and positioning accuracy enhancement strategy.
  • the indication information of an exception in the execution of the first policy includes one or more of the following: an indication of the execution exception, the time of the execution exception, the cause of the execution exception, a measurement parameter used to determine whether the policy has an exception, a configuration parameter used to determine whether the policy has an exception, the effective time of the configuration parameter, the correction method that the first network element expects to adopt for the execution exception, information required for the correction method that the first network element expects to adopt for the execution exception, the identifier of the first AI model, the parameters of the first AI model, the identifier of the first policy, and the effective time of the first policy.
  • the third network element satisfies any of the following conditions: the third network element is executing the strategy output by the first AI model; or, the third network element is executing the first strategy; or, the third network element is executing a fifth strategy, and the fifth strategy is of the same type as the first strategy.
  • the second network element determines a correction method that is allowed for the execution exception; and sends first information to the first network element, where the first information is used to indicate the correction method that is allowed for the execution exception.
  • the allowed correction methods include: executing a second strategy, where the second strategy is a strategy obtained based on non-AI, and the second strategy is of the same type as the first strategy; or, executing a third strategy, where the third strategy is a strategy determined based on a second AI model, and the third strategy is of the same type as the first strategy; or, executing a fourth strategy, where the fourth strategy is a strategy executed before executing the first strategy, and the fourth strategy is of the same type as the first strategy.
  • the first information includes at least one of the following: an identifier of the third strategy, the third strategy, an identifier of the second AI model, a parameter of the second AI model, or performance information of the third network element; wherein the performance information of the third network element is used to determine the third strategy.
  • the performance information of the third network element is used to determine the third strategy, including: the performance information of the third network element is used to correct the first AI model, and the corrected first AI model is used to determine the third strategy.
  • the second network element receives performance information from the third network element.
  • a communication method wherein the execution subject of the method may be a third network element, or a component applied in the third network element, such as a chip, a processor, etc.
  • the execution subject being the third network element as an example.
  • the third network element receives second information from the second network element, wherein the second information is used to indicate that an exception occurs when a first strategy is executed or an exception occurs when a first type of strategy output by a first AI model is executed, and the first strategy belongs to the first type; wherein the first strategy is obtained based on the first AI model; the performance information of the third network element is sent to the second network element, and/or it is determined whether an exception occurs in the strategy executed by the third network element; wherein the performance information is used by the first network element to correct the first strategy or correct the first AI model.
  • Network elements can sense the execution status of other network elements' policies based on AI models, which can improve network performance to a certain extent.
  • AI models which can improve network performance to a certain extent.
  • the situation is notified to a third network element, which can facilitate the third network element to respond based on the abnormality and improve network performance.
  • the second network element is a core network element, or an operation, administration and maintenance OAM network element
  • the third network element is a second access network device
  • the second network element is an access network device
  • the third network element is a second terminal device
  • the second network element is a first access network device
  • the third network element is a second access network device.
  • a communication method wherein the execution subject of the method may be a first network element, or a component applied in the first network element, such as a chip, a processor, etc.
  • the following description is made by taking the execution subject being the first network element as an example.
  • the first network element obtains a first measurement quantity to be input and obtains a first AI model.
  • the first measurement quantity is input into the first AI model to obtain a first strategy output by the first AI model.
  • the first strategy is executed.
  • an indication message is sent to a third network element, wherein the indication message is used to indicate that an abnormality occurs in the execution of the first strategy.
  • the first network element executes the first strategy determined based on the AI model, it can detect the execution status and notify the second network element. In this way, the network elements can perceive the execution status of other network elements executing the strategy based on the AI model, which can improve network performance to a certain extent. In particular, when it is determined that an abnormality occurs in the execution of the first strategy, the situation is notified to other network elements, which can facilitate other network elements to respond based on the abnormality and improve network performance.
  • the first network element is a first terminal, the second network element is a second terminal; or the third network element is a first access network device.
  • the third network element is the second access network device.
  • the first network element receives performance information from a third network element; based on the performance information of the third network element, the first strategy is modified or the first AI model is modified.
  • a communication method wherein the execution subject of the method may be a third network element, or a component applied in the third network element, such as a chip, a processor, etc.
  • the third network element receives indication information from the first network element that an abnormality occurs in the execution of a first strategy by the first network element, and the first strategy is a strategy output by the first AI model after the first network element inputs the first measurement quantity to be input into the first AI model.
  • the third network element determines whether an abnormality occurs in the strategy executed by the third network element and/or the third network element sends the performance information of the third network element to the first network element, and the performance information is used by the first network element to correct the first strategy or the first AI model.
  • Network elements can sense the execution status of other network elements' policies based on AI models, which can improve network performance to a certain extent.
  • AI models which can improve network performance to a certain extent.
  • the situation is notified to a third network element, which can facilitate the third network element to respond based on the abnormality and improve network performance.
  • the first network element is a first terminal, and the second network element is a second terminal; or, the third network element is a first access network device, and the third network element is a second access network device.
  • a communication device wherein the device has the functions of implementing any of the above aspects and any possible implementation of any of the above aspects. These functions can be implemented by hardware, or by hardware executing corresponding software implementations.
  • the hardware or software includes one or more functional modules corresponding to the above functions.
  • a communication device comprising a processor and, optionally, a memory; the processor and the memory are coupled; the memory is used to store computer programs or instructions; the processor is used to execute part or all of the computer programs or instructions in the memory, and when the part or all of the computer programs or instructions are executed, it is used to implement the functions of any of the above aspects and any possible implementation methods of any aspect.
  • the device may further include a transceiver, the transceiver being configured to send a signal processed by the processor or receive a signal input to the processor.
  • the transceiver may perform a sending action or a receiving action in any aspect and any possible implementation of any aspect.
  • the present application provides a chip system, comprising one or more processors (also referred to as processing circuits), wherein the processors are electrically coupled to a memory (also referred to as a storage medium); the memory may be located in the chip system or may not be located in the chip system; the memory is used to store computer programs or instructions; the processor is used to execute part or all of the computer programs or instructions in the memory, and when the part or all of the computer programs or instructions are executed, they are used to implement the functions of any of the above aspects and any possible implementation methods of any aspect.
  • processors also referred to as processing circuits
  • the chip system may further include an input/output interface (also referred to as a communication interface), the input/output interface being used to output a signal processed by the processor, or to receive a signal input to the processor.
  • the input/output interface may perform a sending action or a receiving action in any aspect and any possible implementation of any aspect. Specifically, the output interface performs a sending action, and the input interface performs a receiving action.
  • the chip system may be composed of a chip, or may include a chip and other discrete devices.
  • a computer-readable storage medium for storing a computer program, wherein the computer program includes instructions for implementing the functions of any aspect and any possible implementation of any aspect.
  • a computer-readable storage medium is used to store a computer program, which, when executed by a computer, can enable the computer to execute any of the above aspects and any possible implementation method of any of the aspects.
  • a computer program product comprising: a computer program code, when the computer program code is run on a computer, the computer executes a method in any one of the above aspects and any possible implementation of any one of the aspects.
  • a communication system comprising: executing the first aspect and any possible A first network element for executing the method in the implementation and a second network element for executing the method in the second aspect and any possible implementation of the second aspect.
  • it also includes a third network element for executing the method in the third aspect and any possible implementation of the third aspect.
  • a communication system comprising a first network element executing the method in the fourth aspect and any possible implementation of the fourth aspect, and a third network element executing the method in the fifth aspect and any possible implementation of the fifth aspect.
  • FIG1a is a schematic diagram of a communication system structure provided in an embodiment of the present application.
  • FIG1b is a schematic diagram of a communication system structure provided in an embodiment of the present application.
  • FIG2 is a schematic diagram of an AI model application architecture provided in an embodiment of the present application.
  • FIG3 is a schematic diagram of a communication process provided in an embodiment of the present application.
  • FIG4 is a schematic diagram of a communication process provided in an embodiment of the present application.
  • FIG5 is a structural diagram of a communication device provided in an embodiment of the present application.
  • FIG6 is a structural diagram of a communication device provided in an embodiment of the present application.
  • system architecture of the method provided by the embodiment of the present application is briefly described below. It is understandable that the system architecture described in the embodiment of the present application is to more clearly illustrate the technical solution of the embodiment of the present application, and does not constitute a limitation on the technical solution provided by the embodiment of the present application.
  • the technical solution of the embodiment of the present application can be applied to various communication systems, such as satellite communication systems and traditional mobile communication systems.
  • the satellite communication system can be integrated with the traditional mobile communication system (i.e., ground communication system).
  • Communication systems such as wireless local area network (WLAN) communication system, wireless fidelity (WiFi) system, long term evolution (LTE) system, LTE frequency division duplex (FDD) system, LTE time division duplex (TDD), fifth generation (5G) system or new radio (NR), sixth generation (6G) system, and other future communication systems, etc., also support communication systems that integrate multiple wireless technologies, for example, it can also be applied to drones, satellite communication systems, high altitude platform (HAPS) communication and other non-terrestrial networks (NTN) that integrate ground mobile communication networks.
  • HAPS high altitude platform
  • NTN non-terrestrial networks
  • Figure 1a is a schematic diagram of a 5G communication system architecture to which the present application may be applied, including terminal equipment, access network equipment, and core network equipment.
  • Network equipment can communicate and interact with core network equipment to provide communication services to terminal equipment.
  • Core network equipment is, for example, equipment in the core network (CN) of a 5G network.
  • the core network provides an interface to the data network, provides communication connection, authentication, management, policy control, and data service bearing for user equipment (UE).
  • UE user equipment
  • Terminal device can be a wireless terminal or a wired terminal, and can also be called user equipment (UE). It can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; it can also be deployed on the water (such as ships, etc.); it can also be deployed in the air (such as airplanes, balloons and satellites, etc.). Terminal device refers to a device that provides voice and/or data connectivity to users, a handheld device with wireless connection function, or other processing equipment connected to a wireless modem. It can communicate with one or more core networks via a wireless access network.
  • UE user equipment
  • the terminal device can be a mobile phone, a tablet computer (pad), a computer with wireless transceiver function, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical, a wireless terminal in smart grid, a wireless terminal in transportation safety, a wireless terminal in smart city, a wireless terminal in smart home, etc.
  • Wireless terminals can be mobile phones (or "cellular" phones) and computers with mobile terminals, for example, portable, pocket-sized, handheld, computer-built-in or vehicle-mounted mobile devices that exchange speech and/or data with a wireless access network.
  • PCS personal communication service
  • SIP session initiation protocol
  • WLL wireless local loop
  • PDA personal digital assistants
  • the terminal can also be called a system, a subscriber unit (SU), a subscriber station (SS), a mobile station (MB), a mobile station (mobile), a remote station (RS), an access point (access point, AP), remote terminal (RT), access terminal (AT), user terminal (UT), and user agent (UA).
  • SU subscriber unit
  • SS subscriber station
  • MB mobile station
  • RS remote station
  • RT access point
  • AT user terminal
  • U user agent
  • the (R)AN device in this application is a device that provides wireless communication functions for terminal devices.
  • the (R)AN device is also called an access network device.
  • the RAN device in this application includes but is not limited to: the next generation base station (g nodeB, gNB) in 5G, evolved node B (evolved node B, eNB), radio network controller (radio network controller, RNC), node B (node B, NB), base station controller (base station controller, BSC), base transceiver station (base transceiver station, BTS), home base station (for example, home evolved nodeB, or home node B, HNB), baseband unit (baseBand unit, BBU), transmission point (transmitting and receiving point, TRP), transmission point (transmitting point, TP), mobile switching center, etc.
  • g nodeB, gNB next generation base station
  • eNB evolved node B
  • RNC radio network controller
  • node B node B
  • base station controller base station controller
  • the names of devices with base station functions may be different.
  • RAN Fifth Generation
  • gNB Fifth Generation NodeB
  • eNB evolved NodeB
  • Node B Node B
  • the base station can be a centralized unit (CU) and distributed unit (DU) separated architecture.
  • RAN can be connected to the core network (for example, it can be the core network of Long Term Evolution LTE, or it can be the core network of 5G, etc.).
  • CU and DU can be understood as the division of the base station from the perspective of logical functions.
  • CU and DU can be physically separated or deployed together. Multiple DUs can share one CU.
  • One DU can also be connected to multiple CUs (not shown in Figure 1b).
  • CU and DU can be connected through an interface, such as an F1 interface.
  • CU and DU can be divided according to the protocol layer of the wireless network.
  • CU is used to perform the functions of the radio resource control (RRC) layer, the service data adaptation protocol (SDAP) layer, and the packet data convergence protocol (PDCP) layer; while DU is used to perform the functions of the radio link control (RLC) layer, the media access control (MAC) layer, the physical layer, etc.
  • RRC radio resource control
  • SDAP service data adaptation protocol
  • PDCP packet data convergence protocol
  • DU is used to perform the functions of the radio link control
  • RLC radio link control
  • MAC media access control
  • the physical layer etc.
  • CU or DU can be divided into functions with more protocol layers.
  • CU or DU can also be divided into partial processing functions with protocol layers.
  • the functions of CU or DU can also be divided according to the service type or other system requirements. For example, according to the latency, the functions that need to meet the latency requirements are set in the DU, and the functions that do not need to meet the latency requirements are set in the CU.
  • the CU can also have one or more functions of the core network.
  • One or more CUs can be set centrally or separately.
  • the CU can be set on the network side for centralized management.
  • the DU can have multiple RF functions, or the RF function can be set remotely.
  • the functions of CU can be implemented by one entity or by different entities.
  • the functions of CU can be further divided, for example, the control plane (CP) and the user plane (UP) are separated, that is, the control plane (CU-CP) of CU and the user plane (CU-UP) of CU.
  • CU-CP and CU-UP can be implemented by different functional entities and connected through the E1 interface.
  • the CU-CP and CU-UP can be coupled with DU to jointly complete the functions of the base station.
  • the control plane CU-CP of CU also includes a further divided architecture, that is, the existing CU-CP is further divided into CU-CP1 and CU-CP2.
  • CU-CP1 includes various wireless resource management functions
  • CU-CP2 only includes RRC functions and PDCP-C functions (that is, the basic functions of control plane signaling at the PDCP layer).
  • the core network equipment may include one or more of the following network elements:
  • the access management network element (also known as the access management network element, mobility management network element, access and mobility management network element) is a control plane network element provided by the operator network, which is responsible for the access control and mobility management of terminal devices accessing the operator network, such as mobile state management, allocation of user temporary identity, authentication and user functions.
  • the access management network element can be an access and mobility management function (AMF) network element.
  • AMF access and mobility management function
  • the access management network element can still be an AMF network element, or it can have other names, which are not limited in this application.
  • the session management network element is mainly responsible for session management in the mobile network, such as session establishment, modification, and release. Specific functions include allocating IP addresses to users, selecting user plane network elements that provide message forwarding functions, etc.
  • the session management network element can be a session management function (SMF) network element.
  • SMF session management function
  • the session management network element can still be an SMF network element, or it can have other names, which are not limited in this application.
  • the user plane network element is responsible for forwarding and receiving user data in the terminal device. It can receive user data from the data network and transmit it to the terminal device through the access network device; the user plane network element can also receive user data from the terminal device through the access network device and forward it to the data network.
  • the transmission resources and scheduling functions that provide services to the terminal device in the user plane network element are managed and controlled by the SMF network element.
  • the user plane network element can be a user plane function (UPF) network element.
  • UPF user plane function
  • future communication systems the user plane network element will still be It can be a UPF network element, or it can have other names, which is not limited in this application.
  • the data management network element is used to generate authentication credentials, user identification processing (such as storing and managing user permanent identities, etc.), access control and contract data management, etc.
  • the data management network element can be a unified data management (UDM) network element.
  • the unified data management can still be a UDM network element, or it can have other names, which are not limited in this application.
  • the policy control network element mainly supports providing a unified policy framework to control network behavior, provides policy rules to the control layer network functions, and is responsible for obtaining user contract information related to policy decisions.
  • the policy control network element can be a policy and charging rules function (PCRF) network element.
  • the policy control network element can be a policy control function (PCF) network element.
  • PCF policy control function
  • the policy control network element can still be a PCF network element, or it can have other names, which are not limited in this application.
  • the network storage network element can be used to provide network element discovery function and provide network element information corresponding to the network element type based on the request of other network elements.
  • NRF also provides network element management services, such as network element registration, update, deregistration, network element status subscription and push, etc.
  • the network storage network element can be a network repository function (NRF) network element.
  • NRF network repository function
  • the network storage network element can still be an NRF network element, or it can have other names, which are not limited in this application.
  • the network open network element is a control plane network element provided by the operator.
  • the network open network element opens the external interface of the operator network to the third party in a secure manner. It can be used to provide services and capabilities provided by the 3rd generation partnership project (3GPP) network function equipment for securely opening to the outside.
  • 3GPP 3rd generation partnership project
  • the network open network element can serve as a relay for the communication between the session management network element and the network element of the third party.
  • the network open network element serves as a relay, it can serve as a translator of the identification information of the contracted user and the identification information of the network element of the third party.
  • the network open network element when the network open network element sends the SUPI of the contracted user from the operator network to the third party, the SUPI can be translated into its corresponding external identity (identity, ID). Conversely, when the network open network element sends the external ID (network element ID of the third party) to the operator network, it can translate it into SUPI.
  • the network open function network element can be a network exposure function (NEF) network element. In future communication systems, the network open function network element may still be a NEF network element, or may have other names, which are not limited in this application.
  • NEF network exposure function
  • the network slice selection network element can be used to select a suitable network slice for the terminal's service.
  • the network slice selection network element can be a network slice selection function (NSSF) network element.
  • the network open function network element can still be an NSSF network element, or it can have other names, which are not limited in this application.
  • the network data analysis network element can collect data from various network functions (NF), such as policy control network elements, session management network elements, user plane network elements, access management network elements, and application function network elements (through network capability exposure function network elements), and perform analysis and prediction.
  • NF network functions
  • the network data analysis network element can be a network data analysis function (NWDAF).
  • NWDAF network data analysis function
  • the network exposure function network element can still be an NWDAF network element, or it can have other names, which are not limited in this application.
  • the unified data storage network element is responsible for storing structured data information, including contract information, policy information, and network data or business data defined in a standard format.
  • the unified data storage network element can be a unified data repository (UDR).
  • the network open function network element can still be a UDR network element, or it can have other names, which are not limited in this application.
  • the network element may also be referred to as a "device”, "entity”, etc.
  • the above network element or function may be a network element in a hardware device, a software function running on dedicated hardware, or a virtualized function instantiated on a platform (for example, a cloud platform).
  • the above network element or function may be divided into one or more services, and further, services that exist independently of the network function may also appear.
  • an instance of the above function, or an instance of a service included in the above function, or an instance of a service that exists independently of the network function may be referred to as a service instance.
  • the technical solution of the embodiments of the present application can be applied to scenarios where AI strategies are executed, and is particularly suitable for scenarios where performance degradation occurs during the execution of AI strategies, and is not limited to communication systems.
  • the model training subject obtains a trained AI model by analyzing the training data (Training data) provided by the data source (Data source) (such as model building, training approximation, reinforcement learning, etc.).
  • the model reasoning subject uses the trained AI model to make reasonable predictions (such as performance prediction, load prediction, UE trajectory prediction, etc. on the access network side) based on the inferred data provided by the data source (Data source).
  • the policy adjustment is uniformly planned by the action (Actor) entity and sent to multiple network entities for operation.
  • the relevant data of the network entities will be collected again in the data source (Data source).
  • the training process of the AI model can be executed in the OAM, in the access network device, in the gNB-CU, in the terminal, or in a separate network element entity (e.g., the RAN intelligent controller (RIC)).
  • the process of reasoning based on the AI model can be executed in the access network device, in the gNB-CU, in the terminal, or in a separate network element entity AIC.
  • the OAM and the access network device can exchange information through the current northbound interface, such as model information, data for training the model, etc.
  • the access network device and the gNB-CU can reuse the current F1, Xn, Uu and other interfaces for information exchange with other network elements.
  • the network element entity AIC can establish a communication link with other network elements (e.g., OAM, access network device, etc.), such as a wired link or a wireless link.
  • other network elements e.g., OAM, access network device, etc.
  • the CP and UP of the CU are separated, the CP is usually responsible for receiving the AI model and the subsequent AI reasoning and policy generation functions.
  • CU-CP is further divided into CU-CP1 and CU-CP2
  • CU-CP1 is usually responsible for receiving the model and subsequent AI reasoning functions, and generating specific interaction signaling, and CU-CP2 interacts with other network elements.
  • Energy Saving Strategy Its specific parameters include but are not limited to one or more of the following: which cell or cells to deactivate, which carrier or carriers to shut down (or open), which channel or channels to shut down (or open), which time slot or time slots to shut down (or open), the value of the transmission power adjustment (the adjustment can be reduced or increased), and the adjusted transmission power value.
  • Energy saving strategies can be applied to access network equipment. For example, the access network equipment collects the load, energy consumption, energy efficiency information of itself and neighboring cells, as well as the trajectory information and measurement results of UE, and predicts the trend of its own load. In combination with the purpose of the cell, KPI requirements, etc., energy saving measures are taken in a timely and appropriate manner without affecting network coverage and user access.
  • the simplest energy saving strategy includes directly deactivating the cell.
  • Other energy saving strategies include carrier shutdown, channel shutdown, time slot shutdown, transmission power adjustment, etc.
  • More complex energy saving strategies also include combining the above energy saving measures.
  • the AI model can predict the load of each cell and formulate a reasonable energy saving strategy to ensure that the energy consumption of the base station is reduced or the energy efficiency of the base station is improved without affecting user services and cell coverage.
  • the current energy-saving strategy can be modified, or it can be directly restored to normal working state, and consider re-forecasting the load or changing the AI model used for re-inference.
  • Load Balancing Strategy Its specific parameters include but are not limited to one or more of the following: which UEs among the UEs accessing to itself are selected by the access network device to switch to the adjacent access network device or adjacent cell, the switching threshold or threshold adjustment value for the current access network device (or current cell) and the adjacent access network device (or adjacent cell) to switch to each other, and the load trend of itself.
  • the load balancing strategy can be applied to access network devices. For example, the access network device collects the load, energy consumption, energy efficiency information of itself and the adjacent cells, as well as the trajectory information and measurement results of the UE, and predicts the trend of its own load.
  • some UEs are reasonably selected to switch to the adjacent cells, or receive UEs from the adjacent cells, so that the load levels between the access network devices of the entire network are close, and the situation where some access network devices are heavily loaded and affect normal services while some access network devices have idle resources is reduced.
  • the prediction accuracy is not 100%, it may lead to unreasonable UE selection or unreasonable switching target cell, switching failure or UE service being affected, or inaccurate load prediction leading to poor load balancing effect, or temporary abnormal load changes causing the original load balancing strategy to no longer apply.
  • Mobility Optimization strategy Its specific parameters include but are not limited to one or more of the following: predicted trajectory of UE, switching time node, target cell for switching, etc. Mobility optimization strategy can be applied to access network equipment. For example, the access network equipment collects historical trajectory information of UE and predicts the future trajectory of UE in combination with the measurement information of UE. Based on the predicted trajectory, it is judged in advance whether the UE will switch, and the switching configuration is sent in advance and the target cell is notified to prepare access resources, so as to reduce the delay of UE in the switching process and reduce the probability of switching and access failure. However, since the trajectory prediction accuracy is not 100%, when the predicted trajectory is wrong, it will cause UE switching failure and service interruption. In this case, you can consider retraining the model and reasoning in combination with abnormal situations, or consider replacing the model to avoid similar abnormal situations in subsequent UEs.
  • CSI-RS Feedback Enhancement strategy whose specific parameters include but are not limited to one or more of the following: Downlink channel matrix.
  • the main process is as follows: First, the access network device and the UE exchange a dictionary, which is usually a model trained in advance by the access network device according to the UE capabilities and its own requirements, and then sends an Encoder and Quantizer tool to the UE; then, the UE can compress and quantize the matrix to be fed back according to the downlink channel matrix and the existing dictionary, and pass the result to the access network device side; next, the access network device side reversely recovers the original channel matrix based on the dictionary and the data reported by the UE.
  • a dictionary which is usually a model trained in advance by the access network device according to the UE capabilities and its own requirements, and then sends an Encoder and Quantizer tool to the UE; then, the UE can compress and quantize the matrix to be fed back according to the downlink channel matrix and the existing dictionary, and pass the result to the access network device side; next, the access
  • Beam Management Enhancement strategy whose specific parameters include but are not limited to one or more of the following: Sparse scanning matrix.
  • the main process is as follows: The access network device instructs the UE to perform beam scanning according to the sparse scanning matrix, and the UE performs beam scanning according to this matrix and feeds back the scanning results. Based on the UE's sparse scanning results, the access network device infers the optimal CSI-RS beam and starts the next stage of scanning for the UE. The UE feeds back the optimal CSI-RS beam ID.
  • Positioning Accuracy Enhancements strategy whose specific parameters include but are not limited to one or more of the following: UE location. Compared with the UE location determined by the existing technology, a more precise location will be predicted, such as whether the UE is located in the city or outdoors.
  • the embodiments of the present application provide a variety of communication methods, and network elements can exchange execution status information when executing strategies. Any network element can make a response strategy based on the execution status of the other party when executing strategies. In this way, multiple network elements can execute better strategies and improve network performance.
  • Embodiment 1 is a diagrammatic representation of Embodiment 1:
  • the communication process can be applied to but not limited to any of the following communication scenarios:
  • the first network element is a first access network device
  • the second network element is a core network device
  • the third network element is a second access network device.
  • the first network element is a first access network device
  • the second network element is an OAM
  • the third network element is a second access network device.
  • the first network element is a first terminal
  • the second network element is a first access network device
  • the third network element is a second terminal.
  • the first network element is a first terminal
  • the second network element is a first access network device
  • the third network element is a second access network device.
  • Access network devices communicate with core network devices or OAM through northbound interfaces or NG interfaces. Access network devices can communicate with OAM directly or through core network devices (such as AMF). Access network devices communicate with each other through Xn interfaces. Terminals communicate with access network devices through Uu interfaces.
  • the first terminal or the second terminal can be the terminal device introduced in Figure 1a and Figure 1b;
  • the first access network device and the second access network device can be the access network device introduced in Figure 1a, or the gNB-CU, gNB-CU-CP, gNB-CU-CP1, gNB-CU-CP2, gNB-DU introduced in Figure 1b;
  • the core network device can be any network element introduced in Figure 1a and Figure 1b, for example, AMF, UDM, SMF, UPF, PCF, NEF, UDR, etc.
  • Step 301 A first network element obtains a first measurement quantity to be input and obtains a first AI model.
  • AI model is a model obtained based on artificial intelligence AI.
  • the present application does not limit the subject of model training. It can be that the first network element trains the first AI model based on AI, or other devices/network elements except the first network element train the first AI model based on AI, and send the information of the first AI model (such as model parameters, model parameters may include but are not limited to one or more of the following input layer, output layer, hidden layer, etc. The number of neurons in each layer, the connection relationship, the weight value of each layer, and other parameters involved in the model) to the first network element.
  • the second network element trains the model based on AI, and sends the information of the trained model to the first network element.
  • the second network element can also send the information of the trained model to the third network element and/or other network elements.
  • the first network element and the second network element train their respective models based on AI.
  • the first network element trains the model based on AI, and sends the trained model to the second network element.
  • the process of training the model based on AI can be referred to the introduction of Figure 2, which will not be repeated here.
  • the trained model can be applied to different purposes, such as energy saving, load balancing, mobility optimization, channel state information reference signal CSI-RS feedback enhancement, beam management enhancement, positioning accuracy enhancement, etc.
  • the first measurement amount includes but is not limited to one or more of the following: information measured by the first access network device, information measured by an adjacent (adjacent to the first access network device) access network device or an adjacent cell, information measured by a UE accessing the first access network device, and information measured by a UE accessing an adjacent access network device or an adjacent cell. It should be noted that the UE, the adjacent access network device, and the adjacent cell can send their own measurement information to the first access network device.
  • the first measurement amount includes but is not limited to one or more of the following: information measured by the access network device accessed by the first terminal, information measured by the adjacent (adjacent to the first access network device) access network device or adjacent cell of the access network device accessed by the first terminal, information measured by other UEs accessing the first access network device, and information measured by UEs accessing adjacent access network devices or adjacent cells. It should be noted that adjacent access network devices, adjacent cells, and other UEs can send their own measurement information to the first terminal.
  • the measured information includes service information, load information, etc.
  • Step 302 The first network element inputs the first measurement quantity into a first AI model to obtain a first strategy output by the first AI model.
  • the type of the first strategy includes any one of the following: energy saving strategy, load balancing strategy, mobility optimization strategy, channel state information reference signal CSI-RS feedback enhancement strategy, beam management enhancement strategy, positioning accuracy enhancement strategy.
  • energy saving strategy load balancing strategy
  • mobility optimization strategy mobility optimization strategy
  • channel state information reference signal CSI-RS feedback enhancement strategy channel state information reference signal
  • beam management enhancement strategy positioning accuracy enhancement strategy.
  • the specific content (or purpose or principle) of each type of strategy can be introduced in the previous article and will not be repeated.
  • Step 303 The first network element executes the first strategy.
  • the first network element can execute one or more strategies at a certain time, and the first strategy is one of the one or more strategies.
  • the first network element can send execution status information when executing the strategy to the second network element.
  • This application takes the first strategy as an example for introduction.
  • Step 304 The first network element sends execution status information when executing the first policy to the second network element.
  • the second network element receives, from the first network element, execution status information when the first network element executes the first strategy.
  • the execution status information is used to indicate whether an exception occurs or no exception occurs when the first network element executes the first policy.
  • the first network element informs the second network element of its execution status when executing the first strategy, so that the second network element can perceive the execution status of the first network element when executing the strategy.
  • Any network element can self-check or adjust its own strategy based on the execution status of the other party when executing the strategy. In this way, multiple network elements can execute better strategies and improve network performance.
  • the first network element may periodically determine the execution status of the first strategy (i.e., whether it is normal), and the period may be 1 minute, 5 minutes, or 10 minutes, etc. Alternatively, the first network element may determine the execution status of the first strategy (i.e., whether it is normal) based on an event trigger (e.g., a certain measurement quantity exceeds a set threshold).
  • an event trigger e.g., a certain measurement quantity exceeds a set threshold.
  • the first network element After determining the execution status, the first network element can send the execution status information to the second network element. Alternatively, if the first network element periodically sends the execution status information, after determining the execution status, the first network element can send the execution status information to the second network element when the next most recent reporting cycle time arrives.
  • the first network element is pre-configured with conditions (or criteria) for determining whether each type of policy is executed abnormally.
  • the first network element can determine whether each type of policy is executed abnormally based on the conditions (or criteria).
  • the following situations are generally applicable to the scenario where the first network element is a first access network device.
  • the load of the adjacent cell or adjacent access network device is greater than or equal to the set load threshold; for example, the set threshold is 80% or 85% of the maximum load.
  • the adjacent cell or adjacent access network device can send information that the load is too high to the first access network device.
  • the first access network device receives the information that the load is too high from the adjacent access network device or the adjacent cell, and the first access network needs to modify the current energy-saving strategy to migrate the load from the adjacent access network device or the adjacent cell to the first access network device, thereby reducing the load level of the adjacent access network device or the adjacent cell.
  • the duration of the terminal device accessing the first access network device is greater than or equal to the set duration threshold.
  • terminal devices greater than or equal to the set number threshold fail to access the first access network device.
  • the set number threshold can be a specific value, such as 100, 1000, etc.
  • the set number threshold can also be a ratio value, such as a certain ratio of the maximum number of UEs that the first access network device can access.
  • the UE's service in the cell of the first access network device cannot be carried out normally, for example, the uplink rate of the terminal device accessing the first access network device is less than or equal to the set first rate threshold; for example, the downlink rate of the terminal device accessing the first access network device is less than or equal to the set second rate threshold; the rate threshold can be set for each service.
  • the AI-based energy-saving strategy has a poor effect. For example, after executing the AI-based energy-saving strategy, the energy consumption of the first access network device is not reduced to the expected level, for example, it is expected to be less than 50% of the non-energy-saving state; or the energy efficiency (system load/overall energy consumption) does not meet expectations, for example, it is expected to be more than 150% of the non-energy-saving state; or it is not as effective as the traditional energy-saving strategy.
  • the first access network device exits the AI-based energy-saving strategy at a certain period (for example, half an hour) and executes the traditional energy-saving strategy. It is found that the energy consumption or energy efficiency of the traditional energy-saving strategy is low.
  • the energy efficiency of the first access network device is less than or equal to the set efficiency threshold.
  • the first access network device The energy efficiency of executing the first strategy is less than or equal to the energy efficiency of the first access network device executing the second strategy or the energy efficiency of not executing the energy-saving strategy, wherein the second strategy is an energy-saving strategy based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional energy-saving strategy, rather than an energy-saving strategy based on AI.
  • the load balancing policy is executed abnormally: the following situations are generally applicable to the scenario where the first network element is a first access network device.
  • the actual load of the first access network device is too far away from the predicted load.
  • the load difference between the actual load of the first access network device and the load predicted by the first access network device based on the first strategy is greater than or equal to the set first load difference threshold.
  • the difference between the actual load of the adjacent access network device (or adjacent cell) and the predicted load is too large, for example, the load difference between the actual load of the adjacent cell or adjacent access network device and the load predicted by the adjacent cell or adjacent access network device is greater than or equal to the set second load difference threshold. It is understandable that the adjacent access network device or adjacent cell can send information that the difference between the actual load and the predicted load is too large to the first access network device.
  • the load difference between the actual load of the first access network device and the actual load of the adjacent access network device is greater than or equal to the set third load difference threshold.
  • the load ratio between the actual load of the first access network device and the actual load of the adjacent access network device continues to exceed the set ratio (for example, 3:1). It is understandable that the adjacent access network device or the adjacent cell can send the actual load to the first access network device.
  • the current AI-based load balancing strategy has poor effect or is not as good as the traditional load balancing strategy.
  • the traditional load balancing strategy can be returned periodically or triggered by an event (for example, when the load difference between the first access network device and the adjacent access network device (or adjacent cell) is greater than a certain threshold). If the load balancing effect of the traditional load balancing strategy is better than the AI-based load balancing strategy (for example, the load ratio of the first access network device to the connected access network device (or adjacent cell) is closer to 1:1), it can be considered that the AI-based load balancing strategy is abnormal.
  • the first load difference is greater than the second load difference, wherein the first load difference is the load difference between the actual load of the first access network device and the actual load of the adjacent access network device when the first access network device executes the first strategy, and the second load difference is the load difference between the first access network device and the adjacent access network device when the first access network device executes the second strategy or does not execute the load balancing strategy.
  • the second strategy is a load balancing strategy based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional load balancing strategy, rather than a load balancing strategy based on AI.
  • the UE has problems such as RLF or potential failure (SHR).
  • RLF potential failure
  • one of the main processes is to switch the UE from the heavily loaded access network device to the lightly loaded access network device, but at the same time, it is also necessary to ensure that the UE's service is not affected and no link problems such as RLF occur.
  • the AI-based load balancing strategy if more UE dropped calls or potential failures occur compared to before execution, or the UE's service capabilities (such as throughput) decrease, it can be considered that the AI-based load balancing strategy is abnormal.
  • the first throughput is less than the second throughput, wherein the first throughput is the throughput of the terminal device accessing the first access network device when the first access network device executes the first strategy, and the second throughput is the throughput of the terminal device accessing the first access network device when the first access network device executes the second strategy or does not execute the load balancing strategy, and the second strategy is a load balancing strategy based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional load balancing strategy, rather than a load balancing strategy based on AI.
  • the first number is greater than the second number, wherein the first number is the first number of terminal devices that experience radio link failure RLF or potential radio link failure or dropped calls during the terminal device switching process when the first access network device executes the first strategy, and the second number is the number of terminal devices that experience radio link failure RLF or potential radio link failure or dropped calls during the terminal device switching process when the first access network device executes the second strategy or does not execute the load balancing strategy.
  • the second strategy is a load balancing strategy obtained based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional load balancing strategy, rather than a load balancing strategy obtained based on AI.
  • the following situations are generally applicable to the scenario where the first network element is a first access network device.
  • the mobility report such as SON received by the first access network device shows that the call drop rate has not decreased, the switching/access delay has not decreased, the access success rate has not increased, etc.
  • the report reported by the UE shows that the call drop rate has not decreased to the preset value (for example, it has decreased to 50% of the call drop rate when the mobility optimization strategy is not executed), or the switching/access delay has not decreased (for example, it has decreased to 50% of the switching/access delay when the mobility optimization strategy is not executed), or the access success rate has not increased to the predicted value (for example, 150% when the mobility optimization strategy is not executed), then it can be considered that the current AI-based mobility optimization strategy is not implemented.
  • the AI's mobility optimization strategy is abnormal.
  • the call drop rate of the terminal device accessing the first access network device is greater than or equal to the set call drop rate threshold, for example, the threshold is 50% of the call drop rate when the mobility optimization strategy is not implemented.
  • the delay of the terminal device switching from other access network devices to the first access network device is greater than or equal to the set delay threshold, for example, the threshold is 50% of the switching delay when the mobility optimization strategy is not implemented.
  • the delay of the terminal device accessing the first access network device is greater than or equal to the set delay threshold, for example, the threshold is 50% of the access delay when the mobility optimization strategy is not implemented.
  • the access success rate of the terminal device accessing the first access network device is less than or equal to the set threshold, for example, the threshold is 150% when the mobility optimization strategy is not implemented.
  • the traditional mobility optimization strategy it is not as efficient as the traditional mobility optimization strategy. For example, in the process of executing the AI-based mobility optimization strategy, you can return to the traditional mobility optimization strategy periodically or on an event trigger (for example, when the call drop rate is higher than a certain threshold). If the switching effect of the traditional mobility optimization strategy is better than the switching effect of the AI-based mobility optimization strategy (for example, the call drop rate is lower), it can be considered that the AI-based mobility optimization strategy is abnormal.
  • the first call drop rate is greater than or equal to the second call drop rate, wherein the first call drop rate is the call drop rate of the terminal device accessing the first access network device when the first access network device executes the first strategy, and the second call drop rate is the call drop rate of the terminal device accessing the first access network device when the first access network device executes the second strategy or does not execute the mobility optimization strategy, and the second strategy is a mobility optimization strategy obtained based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional mobility optimization strategy, rather than a mobility optimization strategy obtained based on AI.
  • the first switching delay is greater than or equal to the second switching delay, wherein the first switching delay is the delay for the terminal device to switch from other access network devices to the first access network device when the first access network device executes the first strategy, and the second switching delay is the delay for the terminal device to switch from other access network devices to the first access network device when the first access network device executes the second strategy or does not execute the mobility optimization strategy, and the second strategy is a mobility optimization strategy obtained based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional mobility optimization strategy, rather than a mobility optimization strategy obtained based on AI.
  • the first access delay is greater than or equal to the second access delay, wherein the first access delay is the delay for the terminal device to access the first access network device when the first access network device executes the first strategy, and the second access delay is the delay for the terminal device to access the first access network device when the first access network device executes the second strategy or does not execute the mobility optimization strategy.
  • the second strategy is a mobility optimization strategy obtained based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional mobility optimization strategy, rather than a mobility optimization strategy obtained based on AI.
  • the first access success rate is less than or equal to the second access success rate, wherein the first access success rate is the access success rate of the terminal device accessing the first access network device when the first access network device executes the first strategy, and the second access success rate is the access success rate of the terminal device accessing the first access network device when the first access network device executes the second strategy or does not execute the mobility optimization strategy.
  • the second strategy is a mobility optimization strategy obtained based on non-AI, and the second strategy is of the same type as the first strategy; the second strategy can be understood as a traditional mobility optimization strategy, rather than a mobility optimization strategy obtained based on AI.
  • the predicted movement trajectory of the terminal device is different from the actual movement trajectory of the terminal device.
  • the trajectory of the UE is predicted to be from cell 1 to cell 2, but according to the measurement results reported by the UE, the UE actually moves from cell 1 to cell 3, and the trajectory prediction of the UE is inaccurate.
  • the following situations are generally applicable to the scenario where the first network element is the first terminal or the first access network device.
  • the UE downlink throughput does not meet expectations, for example, it is expected to reach 120% of the traditional CSI-RS feedback enhancement strategy, or the UE downlink throughput decreases, for example, it decreases to 80% of the traditional CSI-RS feedback enhancement strategy.
  • the throughput of the terminal device accessing the first access network device is less than or equal to the set throughput threshold.
  • the UE may send the throughput to the first access network device so that the access network device can make a judgment.
  • the following situations are generally applicable to the scenario where the first network element is the first terminal or the first access network device.
  • the UE beam scanning duration is not reduced to the expected value, for example, 80% of the expected traditional beam scanning scheme.
  • the beam scanning duration value of the terminal device connected to the first access network device is greater than or equal to the set scanning duration threshold.
  • the UE beam scanning duration does not exceed the traditional beam scanning scheme.
  • the first duration is greater than or equal to the second duration, wherein the first duration is the beam scanning duration of the terminal device accessing the first access network device when the first access network device executes the first strategy, and the second duration is the beam scanning duration of the terminal device accessing the first access network device when the first access network device executes the second strategy or does not execute the beam management enhancement strategy, and the second strategy is a beam management enhancement strategy obtained based on non-AI.
  • the access success rate of the UE is reduced, for example, the access success rate of the terminal device accessing the first access network device is less than or equal to the set success rate threshold; for example, the threshold is 80% of the traditional beam scanning scheme.
  • the first access success rate is less than or equal to the second access success rate, wherein the first access success rate is the access success rate of the terminal device accessing the first access network device when the first access network device executes the first strategy, and the second access success rate is the access success rate of the terminal device accessing the first access network device when the first access network device executes the second strategy or does not execute the beam management enhancement strategy, and the second strategy is a beam management enhancement strategy obtained based on non-AI;
  • the throughput of the UE is reduced, for example, because the target beam of the sparse scanning is inappropriate, the UE does not select the optimal beam, and the throughput is reduced compared with the traditional beam scanning scheme.
  • the throughput of the terminal device accessing the first access network device is less than or equal to the set throughput threshold.
  • the first throughput is less than the second throughput, wherein the first throughput is the throughput of the terminal device accessing the first access network device when the first access network device executes the first strategy, and the second throughput is the throughput of the terminal device accessing the first access network device when the first access network device executes the second strategy or does not execute the beam management enhancement strategy, and the second strategy is a beam management enhancement strategy obtained based on non-AI.
  • the UE may send one or more of the beam scanning duration value, access success rate and throughput to the first access network device so that the access network device can make a judgment.
  • the positioning accuracy enhancement strategy is executed abnormally:
  • the following situations are generally applicable to the scenario where the first network element is the first terminal or the first access network device.
  • a certain proportion (for example, more than 80%) of UEs have incorrect LOS/NLOS judgment results.
  • the number of errors in the line-of-sight transmission los or non-line-of-sight transmission Nlos performed by the first access network device to the terminal device is greater than a set threshold.
  • the above describes the conditions for determining whether an abnormality occurs in each type of policy. These conditions are only examples and should not be used to limit the determination of whether an abnormality occurs in the policy.
  • the conditions for determining whether an abnormality occurs can be used interchangeably between the various types of policies in the present application.
  • the network element that can directly measure these parameters can measure them and send them to the first network element.
  • the first network element determines that the execution of the first policy is normal, it sends an indication information that the first policy is executed normally to the second network element (that is, the execution status information is an indication information that the first policy is executed normally).
  • the indication information that the first policy is executed normally includes but is not limited to one or more of the following: an indication of normal execution, a measurement parameter for judging whether the policy is normal, a configuration parameter for judging whether the policy is normal, the effective time of the configuration parameter, the effective time of the configuration parameter, an identifier of the first AI model (for identifying the first AI model, such as a number or index), a parameter of the first AI model, an identifier of the first policy (for identifying the first policy, such as a number or index), the first policy, and the effective time of the first policy.
  • the value of 1 bit or more bits is used to indicate normal execution or abnormal execution. For example, when the value of 1 bit is 0, it indicates normal execution, and when the value of 1 bit is 1, it indicates abnormal execution.
  • the measurement parameters used to determine whether the strategy is normal can be understood as the performance parameters of each network element (the first network element and/or other network elements communicating with the first network element) when the first network element executes the first strategy, including: one or more of the first measurement quantities obtained in step 301, and/or, one or more parameters (excluding the set threshold) involved in the above-mentioned determination of the situation occurring when the execution is abnormal (or the conditions satisfied when the execution is abnormal).
  • the performance parameters related to normal execution include but are not limited to one or more of the following: current load of each cell, predicted load of each cell, switching success rate, access success rate, etc.
  • the configuration parameters used to determine whether the policy is normal can be understood as various thresholds/threshold values set for determining whether it is abnormal, including but not limited to one or more set thresholds in the conditions for determining whether the policy is executed abnormally as described above.
  • the configuration parameters used to determine whether the policy is normal include but are not limited to one or more of the following: setting a load threshold, setting a duration threshold, setting a value threshold, setting a first rate threshold, setting a second rate threshold, setting an efficiency threshold, setting a first load difference threshold, setting a second load difference threshold, setting a third load difference threshold, setting a throughput threshold, setting a scan duration threshold, setting a success rate threshold, and the effective time of various thresholds/threshold values.
  • the first network element when the first network element determines that an abnormality occurs in executing the first strategy, the first network element sends a signal to execute the first strategy to the second network element.
  • Indication information of an abnormality i.e., the execution status information is indication information of an abnormality in the execution of the first policy.
  • the indication information of an abnormality in the execution of the first policy includes, but is not limited to, one or more of the following: an indication of an abnormality in the execution, a time of the abnormality in the execution, a reason for the abnormality in the execution, a measurement parameter for determining an abnormality in the policy, a configuration parameter for determining an abnormality in the policy, an effective time of the configuration parameter, a correction method that the first network element expects to adopt for the execution abnormality, information required for the correction method that the first network element expects to adopt for the execution abnormality, an identifier of the first AI model (for identifying the first AI model, such as a number or index), a parameter of the first AI model, an identifier of the first policy (for identifying the first policy, such as a number or index), the first policy, and the effective time of the first policy.
  • an indication of an abnormality in the execution includes, but is not limited to, one or more of the following: an indication of an abnormality in the execution, a time of the abnormality in the
  • the abnormal execution time is, for example, 19:00 or 7:00 pm.
  • the reason for the execution exception may be a certain type of policy exception, for example, an energy saving policy exception, a load balancing policy exception, a mobility optimization policy, a channel state information reference signal CSI-RS feedback enhancement policy exception, a beam management enhancement policy exception, or a positioning accuracy enhancement policy exception.
  • the reason for the execution exception can be refined to a specific reason, and the specific reason can be understood as the situation that occurs when determining the execution exception (or the condition satisfied when the execution exception) introduced above.
  • the measurement parameters used to determine whether the policy is abnormal can be understood as the performance parameters of each network element (the first network element and/or other network elements communicating with the first network element) when the first network element executes the first policy, including: one or more of the first measurement quantities obtained in step 301, and/or, one or more parameters (excluding the set threshold value) involved in the above-mentioned determination of the situation occurring when the execution is abnormal (or the conditions satisfied when the execution is abnormal).
  • the configuration parameters used to determine whether the policy is normal can be referred to in the above-mentioned description, and no further description will be given.
  • the configuration parameters used to determine whether a policy is abnormal can be understood as various thresholds/threshold values set for determining whether it is abnormal, including but not limited to one or more set thresholds in the conditions for determining whether the policy is executed abnormally as described above. You can refer to the configuration parameters used to determine whether the policy is normal as described above, and will not repeat them.
  • the correction method that the first network element expects to adopt for the execution abnormality includes but is not limited to any of the following:
  • Method 1 Execute the second strategy, the second strategy is a strategy based on non-AI, and the second strategy is of the same type as the first strategy. It can be understood that the desired correction method is to execute the traditional strategy (the second strategy is the traditional strategy) rather than the strategy based on AI.
  • Method 2 Execute the third strategy, the third strategy is a strategy determined based on the second AI model, and the third strategy is of the same type as the first strategy.
  • the desired correction method is to use a new AI model to determine a new AI strategy (the third strategy is the new AI strategy), and the third strategy can be obtained based on the following method: based on the new second AI model, the second measurement quantity to be input is input into the new second AI model to obtain the third strategy output by the second AI model.
  • the first network element can also report information about the second AI model to the second network element, such as the number or index of the second AI model, or the parameters of the second AI model.
  • the AI model can be obtained based on the correction of the first AI model, or it can be a newly trained AI model.
  • Method 3 Execute the fourth strategy, the fourth strategy is the strategy executed before the first strategy is executed, and the fourth strategy is of the same type as the first strategy. It can be understood that the desired correction method is to roll back to the previously executed strategy (the fourth strategy is the previously executed strategy).
  • Method 4 Do not execute any strategy of the same type as the first strategy. It can be understood that the desired correction method is to exit the first strategy.
  • the information required by the first network element for the correction method expected to be adopted for the execution exception may be information for correcting the first AI model, or information for correcting the first strategy. It can be understood that the corrected first AI model is the second AI model, which can be used to determine the third strategy; the corrected first strategy is the third strategy (new AI strategy).
  • the information required by the first network element for the correction method expected to be adopted for the execution exception includes but is not limited to one or more of the following: the identifier of the new AI model (such as an index or number), the parameters of the new AI model, the identifier of the new AI strategy (i.e., the third strategy) (such as the number or index of the strategy), the new AI strategy, the performance information of other network elements, the historical trajectory of UE, the load information of neighboring cells, the network configuration (such as the network energy saving configuration), etc.
  • step 305 the second network element sends the first information to the first network element.
  • the first network element receives the first information from the second network element.
  • Step 305 is an optional step and may not be performed, or may be combined with the subsequent step 309 for execution.
  • the first information may indicate one or more of the following: allowing the first network element to correct the execution exception, not allowing the first network element to correct the execution exception, the correction method allowed for the execution exception, and information that needs to be sent again by the first network element (for example, information about the first measurement quantity and model-related information. In this case, the first network element can send the corresponding information to the second network element again).
  • the value of 1 bit or more bits can be used to indicate whether the modification is allowed or not. For example, when the value of 1 bit is 0, it indicates that the modification is allowed, and when the value of 1 bit is 1, it indicates that the modification is not allowed.
  • the second network element allows the correction method to be adopted for the execution exception. It may be the correction method that the first network element expects to use, or it may not be the correction method that the first network element expects to use.
  • the second network element may determine the correction method allowed to be used based on the correction method that the first network element expects to use and/or the performance parameters of other network elements.
  • the permitted amendment methods include but are not limited to any of the following:
  • Method 1 Execute the second strategy, the second strategy is a strategy based on non-AI, and the second strategy is of the same type as the first strategy. It can be understood that the allowed correction method is to execute the traditional strategy (the second strategy is the traditional strategy) instead of the strategy based on AI.
  • the first network element corrects the anomaly caused by the first policy based on the correction method, including: the first network element skips the first policy and executes the second policy.
  • Method 2 Execute the third strategy, the third strategy is a strategy determined based on the second AI model, and the third strategy is of the same type as the first strategy.
  • the correction method allowed is to use a new AI model (the second model is the new AI model) to determine a new AI strategy (the third strategy is the new AI strategy), and the third strategy can be obtained based on the following method: input the second measurement quantity to be input into the second AI model to obtain the third strategy output by the second AI model.
  • the second AI model can be an AI model obtained by correction (training) based on the first AI model, or it can be a retrained AI model.
  • the first network element corrects the anomaly caused by the first strategy based on the correction method, including: the first network element skips the first strategy and executes the third strategy.
  • Mode 3 Execute the fourth strategy, the fourth strategy is the strategy executed before the first strategy is executed, and the fourth strategy is of the same type as the first strategy. It can be understood that the allowed correction method is to roll back to the previously executed strategy (the fourth strategy is the previously executed strategy).
  • the first network element corrects the anomaly caused by the first strategy based on the correction method, including: the first network element skips the first strategy and executes the fourth strategy.
  • Method 4 Do not execute any strategy of the same type as the first strategy. It can be understood that the allowed correction method is to exit the first strategy.
  • the first network element corrects the anomaly caused by the first strategy based on the correction method, including: the first network element skips the first strategy.
  • the allowed correction methods can be indicated by the value of 2 bits or even more bits.
  • the first information includes but is not limited to at least one of the following: an index of the third strategy, the third strategy, an index of the second AI model, a parameter of the second AI model, or performance information of the third network element; wherein the performance information of the third network element is used to determine the third strategy.
  • the first network element corrects the anomaly caused by the first policy based on the correction method, including: when the first information includes the identifier of the third policy, executing the third policy indicated by the identifier of the third policy to correct the anomaly caused by the first policy.
  • the first network element corrects the anomaly generated by the first policy based on the correction method, including: when the first information includes the third policy, executing the third policy to correct the anomaly generated by the first policy.
  • the first network element corrects the anomaly generated by the first strategy based on the correction method, including: when the first information includes the identifier of the second AI model, inputting the second measurement amount into the second AI model, obtaining the third strategy, and executing the third strategy to correct the anomaly generated by the first strategy.
  • the second measurement amount may be the same as or different from the first measurement amount. The difference may refer to different measurement information included in the measurement amount, for example, the first measurement amount includes measurement information of the first access network device and measurement information of the UE, and the second measurement amount includes measurement information of the UE, but does not include measurement information of the first access network device.
  • the difference may refer to different values of the measurement amount, for example, for load information, the load information corresponding to the first measurement amount is 80%, and the load information corresponding to the second measurement amount is 85%.
  • the measurement time of the second measurement amount is no earlier than the measurement time of the first measurement amount.
  • the first network element corrects the anomaly generated by the first strategy based on the correction method, including: when the first information includes parameters of the second AI model, inputting the second measurement quantity into the second AI model corresponding to the parameters, obtaining the third strategy, and executing the third strategy to correct the anomaly generated by the first strategy.
  • the first network element corrects the abnormality generated by the first strategy based on the correction method, including: when the first information includes performance information of the third network element, correcting the first AI model based on the performance information to obtain a second AI model; inputting a third measurement amount into the second AI model to obtain the third strategy; and executing the third strategy to correct the abnormality generated by the first strategy, wherein the third measurement amount is the same as the first measurement amount or may be different, wherein the difference may refer to different measurement information included in the measurement amount and/or different values of the measurement amount, and the difference between the first measurement amount and the second measurement amount may be referred to.
  • the principle is similar to that of the introduction, and will not be repeated. Generally, the measurement time of the third measurement quantity is not earlier than the measurement time of the first measurement quantity.
  • a possible implementation method is to correct the first AI model based on the performance information, including: using the performance information as training data, and re-training the first AI model to obtain a second AI model.
  • the first network element determines that an abnormality occurs in the execution of the first strategy, it can wait for the second network element to notify the correction method allowed to be adopted; after receiving the correction method allowed to be adopted from the second network element, it can make adjustments based on the correction method allowed to be adopted.
  • the first network element determines that an abnormality occurs in the execution of the first strategy, it does not need to wait for the second network element to notify the correction method allowed to be adopted, but decides on the correction method to be adopted by itself, and makes corrections based on the correction method decided by itself.
  • the first network element when the first network element determines that an abnormality occurs in the execution of the first strategy, the first network element can first decide on the correction method to be adopted by itself, and make corrections based on the correction method decided by itself, and after receiving the correction method allowed to be adopted from the second network element, it can make corrections based on the correction method allowed to be adopted.
  • step 305 and step 306 are not limited.
  • Step 306 When the second network element determines, based on the execution status information, that an abnormality occurs in the execution of the first policy by the first network element, the second network element sends second information to a third network element.
  • the third network element receives the second information from the second network element.
  • execution status information will only be sent when the execution strategy is abnormal, and the execution status information is indication information that an abnormality has occurred in the execution of the first strategy, then after the second network element receives the indication information that an abnormality has occurred in the execution of the first strategy, it can send the second information to the third network element, and the process of determining whether an abnormality has occurred in the execution of the first strategy by the first network element based on the execution status information can be omitted.
  • the second information is used to indicate one or more of the following:
  • An exception occurs when the first strategy is executed, an exception occurs when the first type of strategy output by the first AI model is executed (the type of the first strategy is the first type), identification information of the first network element (for example, a cell ID), one or more information sent by the first network element to the second network element, one or more information sent by the second network element to the first network element, self-inspection of the currently executed strategy by the third network element, a policy guidance plan for the third network element, and information that needs to be sent by the third network element.
  • the policy guidance scheme of the third network element is similar to the allowed correction method sent by the second network element to the first network element, for example, including any of the following: executing a non-AI-based strategy (which can be understood as a traditional strategy), executing a new AI strategy, executing a previous strategy, and exiting the current strategy.
  • a non-AI-based strategy which can be understood as a traditional strategy
  • executing a new AI strategy executing a previous strategy
  • exiting the current strategy exiting the current strategy.
  • the information required to be sent by the third network element can be understood as the information requested by the second network element, including but not limited to: performance information of the third network element, information requested by the first network element to the second network element.
  • the information requested by the first network element to the second network element includes but is not limited to the information required by the correction method that the first network element expects to adopt for the execution anomaly.
  • the third network element is any network element managed or served by the second network element; in one example, the third network element is executing the policy output by the first AI model (i.e., the first network element and the second network element are network elements using the same model); in one example, the third network element is executing the first policy; in one example, the third network element is executing a fifth policy, the fifth policy is of the same type as the first policy, and the fifth policy can be an AI-based policy or a non-AI-based policy (i.e., a traditional policy). In one example, the third network element is a network element in which the second network element is associated with the first network element and there is a potential AI policy execution anomaly.
  • step 307 and step 308 is not limited.
  • step 307 after receiving the second information from the second network element, the third network element determines whether an abnormality occurs in the policy executed by itself.
  • the second information can be understood as warning information.
  • the third network element performs self-checking according to the warning information of the second network element to eliminate risks.
  • step 308 after receiving the second information from the second network element, the third network element sends the information requested by the second network element (ie, information that needs to be sent by the third network element) to the second network element.
  • the second network element receives the information requested by the second network element from the third network element (ie, the information that needs to be sent by the third network element).
  • step 309 the second network element sends information required by the first network element to the first network element.
  • the first network element receives the information required by the first network element from the second network element.
  • the second network element may send information required by the first network element to the first network element based on the indication information of the abnormal execution of the first policy in step 304 and/or the information requested by the second network element (i.e., information required to be sent by the third network element) sent by the third network element to the second network element in step 308.
  • the information required by the first network element may refer to the content of the first information in step 305.
  • the second network element may determine (or redetermine) whether to allow the first network element to correct the execution exception or not to allow the first network element to correct the execution exception based on the information of the second network element request sent by the third network element (i.e., the information that needs to be sent by the third network element).
  • the first network element is instructed to allow the first network element to make corrections to the execution anomaly, or not allow the first network element to make corrections to the execution anomaly, or the correction methods allowed for the execution anomaly.
  • the description of the first information in step 305 above may be referred to, and will not be repeated.
  • Embodiment 2 The difference from Embodiment 1 includes: direct communication between the first network element and the third network element without the participation of the second network element.
  • the communication process can be applied to but not limited to any of the following communication scenarios:
  • the first network element is the first terminal
  • the third network element is the second terminal.
  • the first network element is a first access network device
  • the third network element is a second access network device.
  • Access network devices communicate with each other through Xn interfaces, and terminals communicate with each other through Sidelink interfaces, such as PC5 interfaces.
  • the first terminal or the second terminal can be the terminal device introduced in Figure 1a and Figure 1b; the first access network device and the second access network device can be the access network device introduced in Figure 1a, or the gNB-CU, gNB-CU-CP, gNB-CU-CP1, gNB-CU-CP2, gNB-DU introduced in Figure 1b.
  • Step 401 Obtain a first measurement quantity to be input, and obtain a first AI model.
  • the AI model is a model obtained based on artificial intelligence AI.
  • Step 402 Input the first measurement quantity to be input into a first AI model to obtain a first strategy output by the first AI model.
  • Step 403 Execute the first strategy.
  • Step 404 The first network element sends execution status information when executing the first policy to the third network element.
  • the third network element receives, from the first network element, execution status information when the first network element executes the first strategy.
  • the execution status information is used to indicate whether an exception occurs or no exception occurs when the first network element executes the first policy.
  • step 401 to step 404 may refer to the process from step 301 to step 304 and will not be repeated here.
  • step 404 and step 304 include: the execution status information sent by the first network element to the third network element may include one or more items of the execution status information sent by the first network element to the second network element in step 304, and/or one or more items of the second information sent by the second network element to the third network element in step 306.
  • the execution status information sent by the first network element to the third network element includes but is not limited to one or more of the following:
  • step 405 after receiving the execution status information from the first network element, the third network element determines whether an abnormality occurs in the policy executed by itself.
  • the third network element determines whether an abnormality occurs in the policy executed by itself.
  • step 406 after receiving the execution status information from the first network element, the third network element sends to the first network element the information requested by the first network element in the execution status information (ie, the information that needs to be sent by the third network element).
  • the third network element determines that the execution status information is indicative information of abnormal execution of the first policy
  • the third network element sends information requested by the first network element in the execution status information (ie, information that needs to be sent by the third network element) to the first network element.
  • the information that needs to be sent by the third network element includes but is not limited to one or more of the following:
  • the information required for the correction method that the first network element expects to adopt for the execution anomaly includes, but is not limited to, one or more of the following: an identifier of a new AI model (such as an index or number), parameters of a new AI model, an identifier of a new AI strategy (i.e., a third strategy) (such as a strategy number or index), a new AI strategy, performance information of other network elements, UE historical trajectory, neighboring cell load information, network configuration (such as network energy-saving configuration), etc.
  • the first network element may also modify the first strategy based on the information of the first network element request sent by the third network element.
  • the first AI model is being corrected or modified.
  • a second strategy is obtained.
  • the first network element can skip the first strategy and execute the second strategy.
  • the first network element modifies the first AI model to obtain a second AI model; the measurement amount is input into the second AI model to obtain the third strategy; the first network element can skip the first strategy and execute the third strategy.
  • One way to modify the first AI model based on performance information is to use the performance information as training data and retrain the first AI to obtain the second AI model.
  • the third network element or the first network element may also send the information interacted between the first network element and the third network element to other network elements (eg, core network equipment, OAM, etc.).
  • other network elements eg, core network equipment, OAM, etc.
  • Embodiment 3 is a diagrammatic representation of Embodiment 3
  • the first network element When the first network element communicates with the second network element, it may also communicate with the third network element. That is, for the first network element, the process of FIG. 3 and the process of FIG. 4 may be executed simultaneously.
  • step 304 when the first network element executes step 304 and step 404 simultaneously, step 304 only serves as a notification, and the second network element does not need to provide feedback to the first network element, for example, there is no need to execute step 305 or step 309.
  • step 404 when the first network element executes step 304 and step 404 simultaneously, step 404 only serves as a notification, and the third network element does not need to provide feedback to the first network element, for example, there is no need to execute step 406.
  • the first network element when the first network element executes step 304 and step 404 simultaneously, the first network element may make a decision based on the feedback from the second network element and the feedback from the third network element.
  • the above describes the method of the embodiment of the present application, and the following describes the device in the embodiment of the present application.
  • the method and the device are based on the same technical concept. Since the principles of the method and the device to solve the problem are similar, the implementation of the device and the method can refer to each other, and the repeated parts will not be repeated.
  • the embodiment of the present application can divide the functional modules of the device according to the above method example. For example, each function can be divided into each functional module, or two or more functions can be integrated into one module. These modules can be implemented in the form of hardware or in the form of software functional modules. It should be noted that the division of modules in the embodiment of the present application is schematic and is only a logical function division. There may be other division methods in the specific implementation.
  • a schematic diagram of the structure of a communication device 500 is provided, and the communication device 500 may include: a processing module 510, and optionally, a receiving module 520a, a sending module 520b, and a storage module 530.
  • the processing module 510 may be connected to the storage module 530, the receiving module 520a, and the sending module 520b, respectively, and the storage module 530 may also be connected to the receiving module 520a and the sending module 520b.
  • the above-mentioned receiving module 520a and sending module 520b can also be integrated together and defined as a transceiver module.
  • the communication device 500 may be a first network element, or a chip or a functional unit applied in the first network element.
  • the communication device 500 has any function of the first network element in the above method, for example, the communication device 500 can execute each step performed by the first network element in the above method of FIG. 3 and FIG. 4.
  • the receiving module 520a may execute the receiving action performed by the first network element in the above method embodiment.
  • the sending module 520b can execute the sending action performed by the first network element in the above method embodiment.
  • the processing module 510 may execute other actions except the sending action and the receiving action among the actions executed by the first network element in the above method embodiment.
  • the processing module 510 is used to obtain a first measurement quantity to be input and a first artificial intelligence AI model; input the first measurement quantity into the first AI model to obtain a first strategy output by the first AI model; execute the first strategy; when it is determined that an abnormality occurs in the execution of the first strategy, the sending module 520b is used to send an indication information of an abnormality in the execution of the first strategy to the second network element.
  • the type of the first strategy includes any one of the following: energy saving strategy, load balancing strategy, mobility optimization strategy, channel state information reference signal CSI-RS feedback enhancement strategy, beam management enhancement strategy, and positioning accuracy enhancement strategy.
  • the indication information of an exception in the execution of the first policy includes one or more of the following: an indication of the execution exception, the time of the execution exception, the cause of the execution exception, a measurement parameter used to determine whether a policy exception occurs, a configuration parameter used to determine whether a policy exception occurs, the effective time of the configuration parameter, the correction method that the first network element expects to adopt for the execution exception, the information required for the correction method that the first network element expects to adopt for the execution exception, the identifier of the first AI model, the parameters of the first AI model, the identifier of the first policy, the first policy or the effective time of the first policy.
  • the receiving module 520a is used to receive first information from the second network element, and the first information is used Indicates the correction method allowed for the execution exception; the processing module 510 is used to correct the exception generated by the first strategy based on the correction method.
  • the allowed correction methods include any one of the following: executing a second strategy, the second strategy is a strategy based on non-AI, and the second strategy is of the same type as the first strategy; executing a third strategy, the third strategy is a strategy determined based on a second AI model, and the third strategy is of the same type as the first strategy; executing a fourth strategy, the fourth strategy is a strategy executed before executing the first strategy, and the fourth strategy is of the same type as the first strategy.
  • the first information includes at least one of the following: an identifier of the third strategy, the third strategy, an identifier of the second AI model, parameters of the second AI model, or performance information of the third network element; wherein the performance information of the third network element is used to determine the third strategy.
  • the processing module 510 is specifically used to, when the first information includes an identifier of the third strategy, execute the third strategy indicated by the identifier of the third strategy to correct the abnormality caused by the first strategy; or, when the first information includes the third strategy, execute the third strategy to correct the abnormality caused by the first strategy; or, when the first information includes the identifier of the second AI model, input the second measurement amount into the second AI model to obtain the third strategy, and execute the third strategy to correct the abnormality caused by the first strategy, wherein the second measurement amount is the same as the first measurement amount, or the measurement time of the second measurement amount is not earlier than the measurement time of the first measurement amount; or, when the first information includes the second A I model, inputting the second measurement amount into the second AI model corresponding to the parameter to obtain the third strategy, and executing the third strategy to correct the abnormality generated by the first strategy, wherein the second measurement amount is the same as the first measurement amount, or the measurement time of the second measurement amount is not earlier than the measurement time of the first measurement amount; or
  • the apparatus is a first access network device, and the second network element is a core network element, or an operation, administration and maintenance (OAM) network element; or, the apparatus is a first terminal device, and the second network element is an access network device; or, the apparatus is a first access network device, and the second network element is a second access network device; or, the apparatus is a first terminal device, and the second network element is a second terminal device.
  • OAM operation, administration and maintenance
  • the sending module 520b is used to send an indication message of an abnormality in the execution of the first strategy to a third network element; the receiving module 520a is used to receive performance information from the third network element; and the processing module 510 is used to modify the first strategy or modify the first AI model based on the performance information of the third network element.
  • the storage module 530 may store computer execution instructions of the method executed by the first network element, so that the processing module 510, the receiving module 520a and the sending module 520b execute the method executed by the first network element in the above example.
  • the communication device 500 may be a second network element, or a chip or a functional unit applied in the second network element.
  • the communication device 500 has any function of the second network element in the above method, for example, the communication device 500 can execute each step performed by the second network element in the method of FIG. 3 above.
  • the receiving module 520a may execute the receiving action performed by the second network element in the above method embodiment.
  • the sending module 520b may execute the sending action performed by the second network element in the above method embodiment.
  • the processing module 510 may execute other actions except the sending action and the receiving action among the actions executed by the second network element in the above method embodiment.
  • the receiving module 520a is used to receive indication information from a first network element that an exception occurs when the first network element executes a first strategy, where the first strategy is a strategy output by the first AI model after the first network element inputs a first measurement quantity to be input into the first AI model;
  • the sending module 520b is used to send second information to a third network element, where the second information is used to indicate that an exception occurs when the first strategy is executed or an exception occurs when a first type of strategy output by the first AI model is executed, and the type of the first strategy is the first type.
  • the third network element satisfies any of the following conditions: the third network element is executing the policy output by the first AI model; or, the third network element is executing the first policy; or, the third network element is executing a fifth policy, and the fifth policy is of the same type as the first policy.
  • the processing module 510 is used to determine the correction method allowed for the execution exception; the sending module 520b is also used to send first information to the first network element, where the first information is used to indicate a correction method that is allowed to be used for the execution exception.
  • the allowed correction methods include: executing a second strategy, where the second strategy is a strategy based on non-AI, and the second strategy is of the same type as the first strategy; or, executing a third strategy, where the third strategy is a strategy determined based on a second AI model, and the third strategy is of the same type as the first strategy; or, executing a fourth strategy, where the fourth strategy is a strategy executed before executing the first strategy, and the fourth strategy is of the same type as the first strategy; and not executing any strategy.
  • the first information includes at least one of the following: an identifier of the third strategy, the third strategy, an identifier of the second AI model, parameters of the second AI model, or performance information of the third network element; wherein the performance information of the third network element is used to determine the third strategy.
  • the receiving module 520a is used to receive performance information from the third network element; the processing module 510 is used to determine the performance information of the third network element as information for correcting the first AI model; or, determine the performance information of the third network element as information for correcting the first strategy; or, determine the information of the second AI model based on the performance information and AI of the third network element; or, determine the information of the third strategy based on the performance information and AI of the third network element.
  • the receiving module 520a is further used to receive performance information of a third network element.
  • the storage module 530 may store computer execution instructions of the method executed by the second network element, so that the processing module 510, the receiving module 520a and the sending module 520b execute the method executed by the second network element in the above example.
  • the communication device 500 may be a third network element, or a chip or functional unit applied to the third network element.
  • the communication device 500 has any function of the third network element in the above method, for example, the communication device 500 can execute each step performed by the third network element in the above method of FIG. 3 and FIG. 4.
  • the receiving module 520a may execute the receiving action performed by the third network element in the above method embodiment.
  • the sending module 520b may execute the sending action performed by the third network element in the above method embodiment.
  • the processing module 510 may execute other actions except the sending action and the receiving action among the actions executed by the third network element in the above method embodiment.
  • the receiving module 520a is used to receive second information from a second network element, the second information being used to indicate that an exception occurs when a first strategy is executed or an exception occurs when a first type of strategy output by a first AI model is executed, and the first strategy belongs to the first type;
  • the sending module 520b is used to send performance information of the third network element to the second network element;
  • the processing module 510 is used to determine whether an exception occurs in the strategy executed by the third network element.
  • the receiving module 520a is used to receive indication information from a first network element that an abnormality occurs in the execution of a first strategy by the first network element, where the first strategy is a strategy output by the first AI model after the first network element inputs a first measurement quantity to be input into the first AI model; the sending module 520b is used to send the performance information of the third network element to the first network element, and the processing module 510 is used to determine whether an abnormality occurs in the strategy executed by the third network element.
  • the storage module 530 may store computer execution instructions of a method executed by a third network element, so that the processing module 510, the receiving module 520a, and the sending module 520b execute the method executed by the third network element in the above example.
  • the storage module may include one or more memories, and the memory may be a device in one or more devices or circuits for storing programs or data.
  • the storage module may be a register, a cache, or a RAM, etc., and the storage module may be integrated with the processing module.
  • the storage module may be a ROM or other types of static storage devices that can store static information and instructions, and the storage module may be independent of the processing module.
  • the transceiver module may be an input or output interface, a pin or a circuit, etc.
  • the device can be implemented by a general bus architecture.
  • FIG6 a schematic block diagram of a communication device 600 is provided.
  • the communication device 600 may include: a processor 610, and optionally, a transceiver 620 and a memory 630.
  • the transceiver 620 may be used to receive a program or instruction and transmit it to the processor 610, or the transceiver 620 may be used for the communication device 600 to communicate and interact with other communication devices, such as interactive control signaling and/or business data.
  • the transceiver 620 may be a code and/or data reading and writing transceiver, or the transceiver 620 may be a signal transmission transceiver between a processor and a transceiver.
  • the processor 610 and the memory 630 are electrically coupled.
  • the communication device 600 may be a first network element, or a chip used in the first network element. It should be understood that the device has any function of the first network element in the above method.
  • the communication device 600 can execute the method of FIG. 3 and FIG. 4. The steps performed by the first network element.
  • the memory 630 is used to store a computer program; the processor 610 can be used to call the computer program or instruction stored in the memory 630 to perform the method performed by the first network element in the above example, or to perform the method performed by the first network element in the above example through the transceiver 620.
  • the communication device 600 can be a second network element, or a chip used in the second network element. It should be understood that the device has any function of the second network element in the above method.
  • the communication device 600 can execute the various steps performed by the second network element in the method of Figure 3.
  • the memory 630 is used to store computer programs; the processor 610 can be used to call the computer program or instructions stored in the memory 630 to execute the method executed by the second network element in the above example, or execute the method executed by the second network element in the above example through the transceiver 620.
  • the communication device 600 can be a third network element, or a chip used in a third network element. It should be understood that the device has any function of the third network element in the above method.
  • the communication device 600 can execute the various steps performed by the third network element in the methods of Figures 3 and 4.
  • the memory 630 is used to store computer programs; the processor 610 can be used to call the computer program or instructions stored in the memory 630 to execute the method executed by the third network element in the above example, or execute the method executed by the third network element in the above example through the transceiver 620.
  • the processing module 510 in FIG. 5 may be implemented by the processor 610 .
  • the receiving module 520a and the sending module 520b in Fig. 5 may be implemented by the transceiver 620.
  • the transceiver 620 is divided into a receiver and a transmitter, the receiver performs the function of the receiving module, and the transmitter performs the function of the sending module.
  • the storage module 530 in FIG. 5 may be implemented by the memory 630 .
  • the device may be implemented by a general-purpose processor (a general-purpose processor may also be referred to as a chip or a chip system).
  • a general-purpose processor may also be referred to as a chip or a chip system.
  • a general processor implemented in a device applied to a first network element or a device applied to a second network element or a third network element includes: a processing circuit (a processing circuit may also be referred to as a processor); optionally, it also includes: an input/output interface connected to and communicating with the internal part of the processing circuit, and a storage medium (a storage medium may also be referred to as a memory), wherein the storage medium is used to store instructions executed by the processing circuit to execute the method executed by the first network element or the second network element or the third network element in the above example.
  • the processing module 510 in FIG. 5 may be implemented by a processing circuit.
  • the receiving module 520a and the sending module 520b in Figure 5 can be implemented by an input-output interface.
  • the input-output interface is divided into an input interface and an output interface, the input interface performs the function of the receiving module, and the output interface performs the function of the sending module.
  • the storage module 530 in FIG. 5 may be implemented by a storage medium.
  • the device of the embodiment of the present application can also be implemented using the following: one or more FPGAs (field programmable gate arrays), PLDs (programmable logic devices), controllers, state machines, gate logic, discrete hardware components, any other suitable circuits, or any combination of circuits capable of performing the various functions described throughout the present application.
  • FPGAs field programmable gate arrays
  • PLDs programmable logic devices
  • controllers state machines, gate logic, discrete hardware components, any other suitable circuits, or any combination of circuits capable of performing the various functions described throughout the present application.
  • the embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a computer, enables the computer to perform the above communication method.
  • the computer program includes instructions for implementing the above communication method.
  • the embodiment of the present application also provides a computer program product, including: computer program code, when the computer program code is executed on a computer, the computer can execute the communication method provided above.
  • An embodiment of the present application also provides a communication system, which includes: one or more of a first network element, a second network element, and a third network element that execute the above-mentioned communication method.
  • the processor mentioned in the embodiment of the present application may be a central processing unit (CPU), a baseband processor, the baseband processor and the CPU may be integrated together or separated, or may be a network processor (NP) or a combination of a CPU and a NP.
  • the processor may further include a hardware chip or other general-purpose processor.
  • the above-mentioned hardware chip may be an application-specific integrated circuit (ASIC), a programmable logic device (PLD) or a combination thereof.
  • the above-mentioned PLD may be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL) and other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. or any combination thereof.
  • the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
  • the memory mentioned in the embodiments of the present application may be a volatile memory or a non-volatile memory, or may include volatile and non-volatile memory. Both memory.
  • non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM) or flash memory.
  • Volatile memory can be random access memory (RAM), which is used as an external cache.
  • RAM random access memory
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • SDRAM synchronous dynamic random access memory
  • DDR SDRAM double data rate synchronous dynamic random access memory
  • ESDRAM enhanced synchronous dynamic random access memory
  • SLDRAM synchronous connection dynamic random access memory
  • DR RAM direct memory bus random access memory
  • the transceiver mentioned in the embodiments of the present application may include a separate transmitter and/or a separate receiver, or a transmitter and a receiver integrated into one.
  • the transceiver may operate under the instruction of a corresponding processor.
  • the transmitter may correspond to a transmitter in a physical device
  • the receiver may correspond to a receiver in a physical device.
  • the disclosed systems, devices and methods can be implemented in other ways.
  • the device embodiments described above are only schematic.
  • the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
  • the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, or it can be an electrical, mechanical or other form of connection.
  • the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present application.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.
  • the above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.
  • the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium.
  • the technical solution of the present application is essentially or partly contributed to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program code.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

本申请涉及通信技术领域，提供一种通信方法及装置，用以提高网络性能。第一网元获取待输入的第一测量量、以及获取第一AI模型。然后，第一网元将所述待输入的第一测量量输入至所述第一AI模型中，得到所述第一AI模型输出的第一策略。接下来，第一网元执行所述第一策略。再接下来，第一网元在确定执行所述第一策略出现异常时，向第二网元发送执行所述第一策略出现异常的指示信息。网元之间可以相互感知其它网元执行基于AI模型得到的策略的执行状态，可以在一定程度上提高网络性能。尤其是当确定执行第一策略出现异常时，将该情况通知给其它网元，可以便于其它网元基于该异常做出应对，提高网络性能。

Description

一种通信方法及装置
相关申请的交叉引用
本申请要求在2022年09月27日提交中国专利局、申请号为202211186062.7、申请名称为“一种通信方法及装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及通信技术领域，尤其涉及一种通信方法及装置。
背景技术
人工智能(artificial intelligence，AI)，是一种通过模拟人脑进行复杂计算的技术。将AI运用到无线通信中，通过智能收集和分析数据，可以提升网络性能和用户体验。AI模型的训练过程可以在核心网设备或接入网设备上执行，接入网设备可以采用训练好的AI模型得到节能策略、负载均衡策略等，并执行对应的策略。目前，各个接入网设备之间独立地基于AI模型确定策略并执行，限制网络性能的提升。
发明内容
本申请实施例提供一种通信方法及装置，用以提高网络性能。
第一方面，提供了一种通信方法，该方法的执行主体可以是第一网元，也可以是应用于第一网元中的部件，例如芯片、处理器等。下面以执行主体是第一网元为例进行描述。首先，第一网元获取待输入的第一测量量、以及获取第一AI模型。然后，第一网元将所述第一测量量输入至所述第一AI模型中，得到所述第一AI模型输出的第一策略。接下来，第一网元执行所述第一策略。再接下来，第一网元在确定执行所述第一策略出现异常时，向第二网元发送指示信息，所述指示信息用于指示执行所述第一策略出现异常。
第一网元在执行基于第一AI模型确定的第一策略时，可以检测执行状态，并通知给第二网元，这样网元之间可以相互感知其它网元执行基于AI模型得到的策略的执行状态，可以在一定程度上提高网络性能。尤其是当确定执行第一策略出现异常时，将该情况通知给其它网元，可以便于其它网元基于该异常做出应对，提高网络性能。
其中，所述第一网元为第一接入网设备，所述第二网元为核心网网元、或操作、管理与维护OAM网元；或者，所述第一网元为第一终端设备，所述第二网元为接入网设备；或者，第一网元为第一接入网设备，第二网元为第二接入网设备；或者，第一网元为第一终端设备，第二网元为第二终端设备。
在一种可能的实现中，所述第一策略的类型包括以下的任一项：节能策略、负载均衡策略、移动性优化策略、信道状态信息参考信号CSI-RS反馈增强策略、波束管理增强策略、定位精度增强策略。
在一种可能的实现中，所述执行所述第一策略出现异常的指示信息包括以下的一项或多项：执行异常的指示、执行异常的时间、执行异常的原因、用于判断策略出现异常的测量参数、用于判断策略出现异常的配置参数、配置参数的生效时间、所述第一网元针对执行异常期望采用的修正方式、所述第一网元针对执行异常期望采用的修正方式所需的信息、所述第一AI模型的标识、所述第一AI模型的参数、所述第一策略的标识、所述第一策略或者所述第一策略的生效时间。通过一个或多个参数信息，可以使第二网元更加清楚地知道异常的相关情况，以便做出更合理的应对措施。
在一种可能的实现中，第一网元接收来自所述第二网元的第一信息，所述第一信息用于指示针对执行异常允许采用的修正方式；第一网元基于所述修正方式对所述第一策略产生的异常进行修正。第二网元可以作为管理者或决策者，来向第一网元通知其允许采用的修正方式，第一网元针对异常进行合理修正，进一步提高网络性能。
在一种可能的实现中，所述允许采用的修正方式包括：执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同；可以理解为允许采用的修正方式为执行传统策略(第二策略即为传统策略)，而非基于AI得到的策略。则第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元跳过所述第一策略，执行所述第二策略。
在一种可能的实现中，所述允许采用的修正方式包括：执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同；可以理解为允许采用的修正方式为采用新的AI模型(第二模型即为新的AI模型)确定新的AI策略(第三策略即为新的AI策略)，第三策略可以基于以下方式得到：将待输入的第二测量量输入至第二AI模型中，得到第二AI模型输出的第三策略。第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元跳过所述第一策略，执行所述第三策略。
在一种可能的实现中，所述允许采用的修正方式包括：执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同；可以理解为允许采用的修正方式为回退到之前执行的策略(第四策略即为之前执行的策略)。则第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元跳过所述第一策略，执行所述第四策略。
在一种可能的实现方式中，所述允许采用的修正方式包括：不执行与所述第一策略类型相同的策略。可以理解为允许采用的修正方式为退出第一策略。则第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元退出所述第一策略。
在一种可能的实现中，在所述允许采用的修正方式为执行第三策略时，所述第一信息包括以下至少一项：所述第三策略的标识、所述第三策略、所述第二AI模型的标识、所述第二AI模型的参数或者所述第三网元的性能信息；其中，所述第三网元的性能信息用于确定所述第三策略。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第三策略的标识时，执行所述第三策略的标识指示的所述第三策略，以对所述第一策略产生的异常进行修正。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第三策略时，执行所述第三策略，以对所述第一策略产生的异常进行修正。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第二AI模型的标识时，第一获取基于所述标识获取所述第二AI模型，将第二测量量输入所述第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第二测量量与所述第一测量量相同，或所述第二测量量的测量时间不早于所述第一测量量的测量时间。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第二AI模型的参数时，将第二测量量输入所述参数对应的第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第二测量量与所述第一测量量相同，或所述第二测量量的测量时间不早于所述第一测量量的测量时间。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第三网元的性能信息时，基于所述性能信息修正所述第一AI模型，得到第二AI模型；将第三测量量输入所述第二AI模型，得到所述第三策略；并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第三测量量与所述第一测量量相同，或所述第三测量量的测量时间不早于所述第一测量量的测量时间。
第二方面，提供了一种通信方法，该方法的执行主体可以是第二网元，也可以是应用于第二网元中的部件，例如芯片、处理器等。下面以执行主体是第二网元为例进行描述。首先，第二网元接收来自第一网元的指示信息，所述指示信息用于指示所述第一网元执行第一策略出现异常，所述第一策略为第一网元将待输入的第一测量量输入至第一AI模型后所述第一AI模型输出的策略；第二网元向第三网元发送第二信息，所述第二信息用于指示所述第一策略被执行时出现异常或第一AI模型输出的第一类型的策略被执行时出现异常，所述第一策略的类型为所述第一类型。
网元之间可以相互感知其它网元执行基于AI模型得到的策略的执行状态，可以在一定程度上提高网络性能。具体的，第一网元确定执行第一策略出现异常时，将该情况通知给第二网元，第二网元基于该异常做出应对，即向第三网元通知第一策略被执行时出现异常或基于AI得到的第一类型的策略被执行时出现异常，以便第三网元做出对应，提高网络性能。
其中，所述第一网元为第一接入网设备，所述第二网元为核心网网元、或操作、管理与维护OAM网元，所述第三网元为第二接入网设备；或者，所述第一网元为第一终端设备，所述第二网元为接入网 设备，所述第三网元为第二终端设备；所述第一网元为终端设备，所述第二网元为第一接入网设备，所述第三网元为第二接入网设备。
在一种可能的实现中，所述第一策略的类型包括以下的任一项：节能策略、负载均衡策略、移动性优化策略、信道状态信息参考信号CSI-RS反馈增强策略、波束管理增强策略、定位精度增强策略。
在一种可能的实现中，所述执行所述第一策略出现异常的指示信息包括以下的一项或多项：执行异常的指示、执行异常的时间、执行异常的原因、用于判断策略出现异常的测量参数、用于判断策略出现异常的配置参数、配置参数的生效时间、所述第一网元针对执行异常期望采用的修正方式、所述第一网元针对执行异常期望采用的修正方式所需的信息、所述第一AI模型的标识、所述第一AI模型的参数、所述第一策略的标识、所述第一策略的生效时间。
在一种可能的实现中，所述第三网元满足以下任一条件：所述第三网元正在执行所述第一AI模型输出的策略；或者，所述第三网元正在执行所述第一策略；或者，所述第三网元正在执行第五策略，所述第五策略与所述第一策略的类型相同。
在一种可能的实现中，第二网元确定针对执行异常允许采用的修正方式；并向所述第一网元发送第一信息，所述第一信息用于指示针对执行异常允许采用的修正方式。
在一种可能的实现中，所述允许采用的修正方式包括：执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同；或者，执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同；或者，执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同。
在一种可能的实现中，在所述允许采用的修正方式为执行第三策略时，所述第一信息包括以下至少一项：所述第三策略的标识、所述第三策略、所述第二AI模型的标识、所述第二AI模型的参数或者所述第三网元的性能信息；其中，所述第三网元的性能信息用于确定所述第三策略。
在一种可能的实现中，所述第三网元的性能信息用于确定所述第三策略，包括：所述第三网元的性能信息用于修正所述第一AI模型，修正后的第一AI模型用于确定所述第三策略。
在一种可能的实现中，第二网元接收来自所述第三网元的性能信息。
第三方面，提供了一种通信方法，该方法的执行主体可以是第三网元，也可以是应用于第三网元中的部件，例如芯片、处理器等。下面以执行主体是第三网元为例进行描述。首先，第三网元接收来自第二网元的第二信息，所述第二信息用于指示第一策略被执行时出现异常或第一AI模型输出的第一类型的策略被执行时出现异常，所述第一策略属于所述第一类型；其中，所述第一策略基于第一AI模型得到；向第二网元发送所述第三网元的性能信息，和/或，确定所述第三网元执行的策略是否出现异常；其中，所述性能信息用于所述第一网元对所述第一策略进行修正或对所述第一AI模型进行修正。
网元之间可以相互感知其它网元执行基于AI模型得到的策略的执行状态，可以在一定程度上提高网络性能。尤其是当策略出现异常时，将该情况通知给第三网元，可以便于第三网元基于该异常做出应对，提高网络性能。
其中，所述第二网元为核心网网元、或操作、管理与维护OAM网元，所述第三网元为第二接入网设备；或者，所述第二网元为接入网设备，所述第三网元为第二终端设备；所述第二网元为第一接入网设备，所述第三网元为第二接入网设备。
第四方面，提供了一种通信方法，该方法的执行主体可以是第一网元，也可以是应用于第一网元中的部件，例如芯片、处理器等。下面以执行主体是第一网元为例进行描述。首先，第一网元获取待输入的第一测量量及获取第一AI模型。然后，将所述第一测量量输入至第一AI模型中，得到所述第一AI模型输出的第一策略。然后执行所述第一策略。接下来，在确定执行所述第一策略出现异常时，向第三网元发送指示信息，所述指示信息用于指示执行所述第一策略出现异常。
第一网元在执行基于AI模型确定的第一策略时，可以检测执行状态，并通知给第二网元，这样网元之间可以相互感知其它网元执行基于AI模型得到的策略的执行状态，可以在一定程度上提高网络性能。尤其是当确定执行第一策略出现异常时，将该情况通知给其它网元，可以便于其它网元基于该异常做出应对，提高网络性能。
其中，所述第一网元为第一终端，所述第二网元为第二终端；或者，所述第三网元为第一接入网设 备，第三网元为第二接入网设备。
在一种可能的实现中，第一网元接收来自第三网元的性能信息；基于所述第三网元的性能信息，对所述第一策略进行修正或对所述第一AI模型进行修正。
第五方面，提供了一种通信方法，该方法的执行主体可以是第三网元，也可以是应用于第三网元中的部件，例如芯片、处理器等。下面以执行主体是第三网元为例进行描述。首先，第三网元接收来自第一网元的所述第一网元执行第一策略出现异常的指示信息，所述第一策略为第一网元将待输入的第一测量量输入至第一AI模型后所述第一AI模型输出的策略。然后，第三网元确定所述第三网元执行的策略是否出现异常和/或第三网元向所述第一网元发送所述第三网元的性能信息，所述性能信息用于所述第一网元对所述第一策略进行修正或对所述第一AI模型进行修正。
网元之间可以相互感知其它网元执行基于AI模型得到的策略的执行状态，可以在一定程度上提高网络性能。尤其是当策略出现异常时，将该情况通知给第三网元，可以便于第三网元基于该异常做出应对，提高网络性能。
其中，所述第一网元为第一终端，所述第二网元为第二终端；或者，所述第三网元为第一接入网设备，第三网元为第二接入网设备。
第六方面，提供了一种通信装置，所述装置具有实现上述任一方面及任一方面的任一可能的实现中的功能。这些功能可以通过硬件实现，也可以通过硬件执行相应的软件实现。所述硬件或软件包括一个或多个与上述功能相对应的功能模块。
第七方面，提供了一种通信装置，包括处理器，可选的，还包括存储器；所述处理器和所述存储器耦合；所述存储器，用于存储计算机程序或指令；所述处理器，用于执行所述存储器中的部分或者全部计算机程序或指令，当所述部分或者全部计算机程序或指令被执行时，用于实现上述任一方面及任一方面的任一可能的实现的方法中的功能。
在一种可能的实现中，所述装置还可以包括收发器，所述收发器，用于发送所述处理器处理后的信号，或者接收输入给所述处理器的信号。所述收发器可以执行任一方面及任一方面的任一可能的实现中的发送动作或接收动作。
第八方面，本申请提供了一种芯片系统，该芯片系统包括一个或多个处理器(也可以称为处理电路)，所述处理器与存储器(也可以称为存储介质)之间电耦合；所述存储器可以位于所述芯片系统中，也可以不位于所述芯片系统中；所述存储器，用于存储计算机程序或指令；所述处理器，用于执行所述存储器中的部分或者全部计算机程序或指令，当所述部分或者全部计算机程序或指令被执行时，用于实现上述任一方面及任一方面的任一可能的实现的方法中的功能。
在一种可能的实现中，所述芯片系统还可以包括输入输出接口(也可以称为通信接口)，所述输入输出接口，用于输出所述处理器处理后的信号，或者接收输入给所述处理器的信号。所述输入输出接口可以执行任一方面及任一方面的任一可能的实现中的发送动作或接收动作。具体的，输出接口执行发送动作，输入接口执行接收动作。
在一种可能的实现中，该芯片系统，可以由芯片构成，也可以包括芯片和其他分立器件。
第九方面，提供了一种计算机可读存储介质，用于存储计算机程序，所述计算机程序包括用于实现任一方面及任一方面的任一可能的实现中的功能的指令。
或者，一种计算机可读存储介质，用于存储计算机程序，所述计算机程序被计算机执行时，可以使得所述计算机执行上述任一方面及任一方面的任一可能的实现的方法。
第十方面，提供了一种计算机程序产品，所述计算机程序产品包括：计算机程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行上述任一方面及任一方面的任一可能的实现中的方法。
第十一方面，提供了一种通信系统，所述通信系统包括执行上述第一方面及第一方面的任一可能的 实现中的方法的第一网元和执行上述第二方面及第二方面的任一可能的实现中的方法的第二网元。可选的，还包括执行上述第三方面及第三方面的任一可能的实现中的方法的第三网元。
第十二方面，提供了一种通信系统，所述通信系统包括执行上述第四方面及第四方面的任一可能的实现中的方法的第一网元和执行上述第五方面及第五方面的任一可能的实现中的方法的第三网元。
上述第六方面至第十二方面的技术效果可以参照第一方面至第五方面中的描述，重复之处不再赘述。
附图说明
图1a为本申请实施例中提供的一种通信系统结构示意图；
图1b为本申请实施例中提供的一种通信系统结构示意图；
图2为本申请实施例中提供的一种AI模型应用架构示意图；
图3为本申请实施例中提供的一种通信流程示意图；
图4为本申请实施例中提供的一种通信流程示意图；
图5为本申请实施例中提供的一种通信装置结构图；
图6为本申请实施例中提供的一种通信装置结构图。
具体实施方式
为便于理解本申请实施例的技术方案，下面将对本申请实施例提供的方法的系统架构进行简要说明。可理解的，本申请实施例描述的系统架构是为了更加清楚的说明本申请实施例的技术方案，并不构成对于本申请实施例提供的技术方案的限定。
本申请实施例的技术方案可以应用于各种通信系统，例如：卫星通信系统、传统的移动通信系统。其中，所述卫星通信系统可以与传统的移动通信系统(即地面通信系统)相融合。通信系统例如：无线局域网(wireless local area network，WLAN)通信系统，无线保真(wireless fidelity，WiFi)系统，长期演进(long term evolution，LTE)系统、LTE频分双工(frequency division duplex，FDD)系统、LTE时分双工(time division duplex，TDD)、第五代(5th generation，5G)系统或新无线(new radio，NR)，第六代(6th generation，6G)系统，以及其他未来的通信系统等，还支持多种无线技术融合的通信系统，例如，还可以应用于无人机、卫星通信系统、高空平台(high altitude platform station，HAPS)通信等非地面网络(non-terrestrial network，NTN)融合地面移动通信网络的系统。
例如，图1a为本申请可以适用的一种5G的通信系统架构示意图，包括终端设备、接入网设备、核心网设备。网络设备可以和核心网设备进行通信交互，向终端设备提供通信服务。核心网设备例如为5G网络核心网(core network，CN)中的设备。核心网作为承载网络提供到数据网络的接口，为用户设备(user equipment，UE)提供通信连接、认证、管理、策略控制以及对数据业务完成承载等。
终端设备(terminal device)，可以是无线终端，也可以是有线终端，也可以称为用户设备(user equipment，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(如轮船等)；还可以部署在空中(例如飞机、气球和卫星上等)。终端设备是指向用户提供语音和/或数据连通性的设备，具有无线连接功能的手持式设备、或连接到无线调制解调器的其他处理设备。可以经无线接入网与一个或多个核心网进行通信。所述终端设备可以是手机(mobile phone)、平板电脑(pad)、带无线收发功能的电脑、虚拟现实(virtual reality，VR)终端、增强现实(augmented reality，AR)终端、工业控制(industrial control)中的无线终端、无人驾驶(self driving)中的无线终端、远程医疗(remote medical)中的无线终端、智能电网(smart grid)中的无线终端、运输安全(transportation safety)中的无线终端、智慧城市(smart city)中的无线终端、智慧家庭(smart home)中的无线终端等。无线终端可以如移动电话(或称为“蜂窝”电话)和具有移动终端的计算机，例如，可以是便携式、袖珍式、手持式、计算机内置的或者车载的移动装置，它们与无线接入网交换语言和/或数据。例如，个人通信业务(personal communication service，PCS)电话、无绳电话、会话发起协议(session initiation protocol，SIP)话机、无线本地环路(wireless local loop，WLL)站、个人数字助理(personal digital assistant，PDA)等设备。终端也可以称为系统、订户单元(subscriber unit，SU)、订户站(subscriber station，SS)，移动站(mobile station，MB)、移动台(mobile)、远程站(remote station，RS)、接入点(access point， AP)、远程终端(remote terminal，RT)、接入终端(access terminal，AT)、用户终端(user terminal，UT)、用户代理(user agent，UA)。
本申请中的(R)AN设备，是一种为终端设备提供无线通信功能的设备，(R)AN设备也称为接入网设备。本申请中的RAN设备包括但不限于：5G中的下一代基站(g nodeB，gNB)、演进型节点B(evolved node B，eNB)、无线网络控制器(radio network controller，RNC)、节点B(node B，NB)、基站控制器(base station controller，BSC)、基站收发台(base transceiver station，BTS)、家庭基站(例如，home evolved nodeB，或home node B，HNB)、基带单元(baseBand unit，BBU)、传输点(transmitting and receiving point，TRP)、发射点(transmitting point，TP)、移动交换中心等。在采用不同的无线接入技术的系统中，具备基站功能的设备的名称可能会有所不同，例如，在第五代(5th generation，5G)系统中，称为RAN或者gNB(5G NodeB)；在LTE系统中，称为演进的节点B(evolved NodeB，eNB或者eNodeB)；在第三代(3rd generation，3G)系统中，称为节点B(Node B)等。
如图1b所示，基站可以是集中式单元(centralized unit，CU)和分布式单元(distributed unit，DU)分离架构。RAN可以与核心网相连(例如可以是长期演进LTE的核心网，也可以是5G的核心网等)。CU和DU可以理解为是对基站从逻辑功能角度的划分。CU和DU在物理上可以是分离的也可以部署在一起。多个DU可以共用一个CU。一个DU也可以连接多个CU(图1b中未示出)。CU和DU之间可以通过接口相连，例如可以是F1接口。CU和DU可以根据无线网络的协议层划分。例如其中一种可能的划分方式是：CU用于执行无线资源控制(radio resource control，RRC)层、业务数据适配协议(service data adaptation protocol，SDAP)层以及分组数据汇聚层协议(packet data convergence protocol，PDCP)层的功能；而DU用于执行无线链路控制(radio link control，RLC)层，媒体接入控制(media access control，MAC)层，物理(physical)层等的功能。可以理解对CU和DU处理功能按照这种协议层的划分仅仅是一种举例，也可以按照其他的方式进行划分。例如可以将CU或者DU划分为具有更多协议层的功能。例如，CU或DU还可以划分为具有协议层的部分处理功能。在一设计中，将RLC层的部分功能和RLC层以上的协议层的功能设置在CU，将RLC层的剩余功能和RLC层以下的协议层的功能设置在DU。在另一种设计中，还可以按照业务类型或者其他系统需求对CU或者DU的功能进行划分。例如按时延划分，将处理时间需要满足时延要求的功能设置在DU，不需要满足该时延要求的功能设置在CU。在另一种设计中，CU也可以具有核心网的一个或多个功能。一个或者多个CU可以集中设置，也分离设置。例如CU可以设置在网络侧方便集中管理。DU可以具有多个射频功能，也可以将射频功能拉远设置。
CU的功能可以由一个实体来实现也可以由不同的实体实现。例如，可以对CU的功能进行进一步切分，例如，将控制面(control plane，CP)和用户面(user plane，UP)分离，即CU的控制面(CU-CP)和CU用户面(CU-UP)。例如，CU-CP和CU-UP可以由不同的功能实体来实现，并通过E1接口相连，所述CU-CP和CU-UP可以与DU相耦合，共同完成基站的功能。CU的控制面CU-CP还包括一种进一步切分的架构，即把现有的CU-CP进一步切分为CU-CP1和CU-CP2。其中CU-CP1包括各种无线资源管理功能，CU-CP2仅包括RRC功能和PDCP-C功能(即控制面信令在PDCP层的基本功能)。
核心网设备可包括以下网元中的一个或多个：
接入管理网元(也可以称为接入管理网元、移动性管理网元、接入与移动性管理网元)，是由运营商网络提供的控制面网元，负责终端设备接入运营商网络的接入控制和移动性管理，例如包括移动状态管理，分配用户临时身份标识，认证和用户等功能。在5G通信系统中，该接入管理网元可以是接入与移动性管理功能(access and mobility management function，AMF)网元。在未来通信系统中，接入管理网元仍可以是AMF网元，或者，还可以有其它的名称，本申请不做限定。
会话管理网元，主要负责移动网络中的会话管理，如会话建立、修改、释放。具体功能如为用户分配IP地址、选择提供报文转发功能的用户面网元等。在5G通信系统中，该会话管理网元可以是会话管理功能(session management function，SMF)网元。在未来通信系统中，会话管理网元仍可以是SMF网元，或者，还可以有其它的名称，本申请不做限定。
用户面网元，负责终端设备中用户数据的转发和接收。可以从数据网络接收用户数据，通过接入网设备传输给终端设备；用户面网元还可以通过接入网设备从终端设备接收用户数据，转发到数据网络。用户面网元中为终端设备提供服务的传输资源和调度功能由SMF网元管理控制的。在5G通信系统中，该用户面网元可以是用户面功能(user plane function，UPF)网元。在未来通信系统中，用户面网元仍 可以是UPF网元，或者，还可以有其它的名称，本申请不做限定。
数据管理网元，用于生成认证信任状，用户标识处理(如存储和管理用户永久身份等)，接入控制和签约数据管理等。在5G通信系统中，该数据管理网元可以是统一数据管理(unified data management，UDM)网元。在未来通信系统中，统一数据管理仍可以是UDM网元，或者，还可以有其它的名称，本申请不做限定。
策略控制网元，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。在4G通信系统中，该策略控制网元可以是策略和计费规则功能(policy and charging rules function，PCRF)网元。在5G通信系统中，该策略控制网元可以是策略控制功能(policy control function，PCF)网元。在未来通信系统中，策略控制网元仍可以是PCF网元，或者，还可以有其它的名称，本申请不做限定。
网络存储网元，可用于提供网元发现功能，基于其他网元的请求，提供网元类型对应的网元信息。NRF还提供网元管理服务，如网元注册、更新、去注册以及网元状态订阅和推送等。在5G通信系统中，该网络存储网元可以是网络注册功能(network repository function，NRF)网元。在未来通信系统中，网络存储网元仍可以是NRF网元，或者，还可以有其它的名称，本申请不做限定。
网络开放网元，是由运营商提供控制面网元，网络开放网元以安全的方式对第三方开放运营商网络的对外接口，可用于提供用于安全地向外部开放由第三代合作伙伴计划(3rd generation partnership project，3GPP)网络功能设备提供的业务和能力等。例如，在会话管理网元需要与第三方的网元通信时，网络开放网元可作为会话管理网元与第三方的网元通信的中继。网络开放网元作为中继时，可作为签约用户的标识信息的翻译，以及第三方的网元的标识信息的翻译。比如，网络开放网元将签约用户的SUPI从运营商网络发送到第三方时，可以将SUPI翻译成其对应的外部身份标识(identity，ID)。反之，网络开放网元将外部ID(第三方的网元ID)发送到运营商网络时，可将其翻译成SUPI。在5G通信系统中，网络开放功能网元可以是网络开放功能(network exposure function，NEF)网元。在未来通信系统中，网络开放功能网元仍可以是NEF网元，或者，还可以有其它的名称，本申请不做限定。
网络切片选择网元，可用于为终端的业务选择合适的网络切片。在5G通信系统中，网络切片选择网元可以是网络切片选择功能(network slice selection function，NSSF)网元。在未来通信系统中，网络开放功能网元仍可以是NSSF网元，或者，还可以有其它的名称，本申请不做限定。
网络数据分析网元，可以从各个网络功能(network function，NF)，例如策略控制网元、会话管理网元、用户面网元、接入管理网元、应用功能网元(通过网络能力开放功能网元)收集数据，并进行分析和预测。在5G通信系统中，网络数据分析网元可以是网络数据分析功能(network data analytics function，NWDAF)。在未来通信系统中，网络开放功能网元仍可以是NWDAF网元，或者，还可以有其它的名称，本申请不做限定。
统一数据存储网元，负责存储结构化的数据信息，其中包括签约信息，策略信息，以及有标准格式定义的网络数据或业务数据。在5G通信系统中，统一数据存储网元可以是统一数据存储(unified data repository，UDR)。在未来通信系统中，网络开放功能网元仍可以是UDR网元，或者，还可以有其它的名称，本申请不做限定。
可以理解的是，网元也可以称为“设备”、“实体”等。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的软件功能，或者是平台(例如，云平台)上实例化的虚拟化功能。上述网元或者功能可划分出一个或多个服务，进一步，还可能会出现独立于网络功能存在的服务。在本申请中，上述功能的实例、或上述功能中包括的服务的实例、或独立于网络功能存在的服务实例均可称为服务实例。
为便于理解本申请实施例，接下来对本请的应用场景进行介绍，本申请实施例描述的网络架构以及业务场景是为了更加清楚的说明本申请实施例的技术方案，并不构成对于本申请实施例提供的技术方案的限定，本领域普通技术人员可知，随着新业务场景的出现，本申请实施例提供的技术方案对于类似的技术问题，同样适用。
本申请实施例的技术方案可以应用于执行AI策略的场景中，尤其适用于执行AI策略出现性能下降的场景中，不仅仅限制于通信系统中。
如图2所示，介绍了一种AI模型应用架构示意图。其中，数据源(Data source)来自终端、接入网设备、gNB-CU、gNB-DU、操作、管理与维护(operation administration and maintenance，OAM)、或 其他管理实体等提供的数据(例如OAM检测的网络运行数据、例如网络负载、信道质量等)。模型训练主体通过对数据源(Data source)提供的训练数据(Training data)进行分析(例如模型建立、训练逼近、强化学习等)，得到训练好的AI模型。模型推理主体使用训练好的AI模型，基于数据源(Data source)提供的推测数据，得出合理的预测(例如接入网侧的性能预测、负载预测、UE轨迹预测等)。以指导网络做出策略调整，得到合理高效的策略(例如节省策略、移动性优化策略等)。策略调整由动作(Actor)实体统一规划，并发送到多个网络实体去运行。同时，多个网络实体运行了调整后的策略后，网络实体的相关数据会被再次收集到数据源(Data source)中。
在无线通信系统中，AI模型的训练过程可以在OAM中执行，也可以在接入网设备中执行，也可以在gNB-CU中执行，也可以是终端中执行，也可以是一个单独的网元实体(例如RAN侧智能控制中心(RAN intelligent controller，RIC))执行。基于AI模型进行推理的过程可以在接入网设备中执行，也可以在gNB-CU中执行，也可以在终端中执行，也可以是一个单独的网元实体AIC执行。示例性的，OAM与接入网设备之间可以通过当前的北向接口进行信息交互，例如模型的信息，用于训练模型的数据等。接入网设备、gNB-CU可以与其它网元之间可以复用当前的F1、Xn、Uu等接口进行信息交互。网元实体AIC与其它网元(例如OAM、接入网设备等)可以建立通信链路来通信，例如有线链路，或无线链路。当CU的CP和UP分离时，通常由CP负责接收AI模型以及后续AI的推理和策略生成功能。当CU-CP进一步切分为CU-CP1和CU-CP2时，通常由CU-CP1负责接收模型以及后续AI的推理功能，并生成具体的交互信令，由CU-CP2与其它网元进行交互。
为便于理解本申请实施例，以下对本申请实施例的部分用语进行解释说明，以便于本领域技术人员理解。
1、节能(Energy Saving)策略：其具体的参数包括但不限于以下的一项或多项：哪个或哪些小区去激活、关断(或打开)哪个或哪些载波、关断(或打开)哪个或哪些通道、关断(或打开)哪个或哪些时隙、发射功率的调整值(调整可以是降低或升高)数值、调整后的发射功率值。节能策略可以适用于接入网设备，例如，通过接入网设备收集本身和邻区的负载、能耗、能效信息，以及UE的轨迹信息、测量结果等，对本身负载的走向进行预测，并结合小区用途、KPI要求等，在不影响网络覆盖、用户接入的前提下，适时适当采取节能措施。最简单的节能策略包括直接将小区进行去激活，其他节能策略包括载波关断、通道关断、时隙关断、发射功率调整等，较为复杂的节能策略还包括将上述节能措施进行组合。AI模型可以对各小区的负载进行预测，并且制定合理的节能策略，从而保障在不影响用户业务和小区覆盖的前提下，降低基站的能耗，或提高基站的能效。当网络覆盖受到影响，或者无法满足UE接入、业务需求时，可以修改当前的节能策略，或者直接恢复到正常工作状态，并考虑对负载进行重新预测或更改使用的AI模型以进行重新推理。
2、负载均衡(Load Balancing)策略：其具体的参数包括但不限于以下的一项或多项：接入网设备选择接入自身的UE中的哪些UE切换到相邻接入网设备或相邻小区、当前接入网设备(或当前小区)小区和相邻接入网设备(或相邻小区)互相切换的切换门限或门限的调整值、本身的负载走向。负载均衡策略可以适用于接入网设备，例如，通过接入网设备收集本身和邻区的负载、能耗、能效信息，以及UE的轨迹信息、测量结果等，对本身负载的走向进行预测，并结合小区用途、KPI要求等，合理选择部分UE令其切换到邻区、或从邻区接收UE，使得整个片网的接入网设备间负载水平接近，减少出现部分接入网设备负载重影响正常业务而部分接入网设备资源闲置的情况。然而由于预测的准确性并非百分百，会导致出现UE选择不合理或者切换目标小区不合理，出现切换失败或UE业务收到影响，或，负载预测不准确导致负载均衡效果不佳，或，临时出现负载异常变动导致原有负载均衡策略不再适用的情况，此时可以退出或修改当前的负载均衡策略，并考虑对负载进行重新预测或更改使用的AI模型以进行重新推理。
3、移动性优化(Mobility Optimization)策略：其具体的参数包括但不限于以下的一项或多项：UE的预测轨迹、切换的时间节点，切换的目标小区等。移动性优化策略可以适用于接入网设备，例如，通过接入网设备对UE的历史轨迹信息收集，结合UE的测量信息，对UE的未来轨迹做出预测。基于预测的轨迹，提前判断UE是否切换，并且提前下发切换配置以及知会目标小区准备接入资源，减少UE在切换过程中的延迟并降低出现切换、接入失败的概率。然而由于轨迹预测准确率并非百分百，当预测轨迹错误时，会导致UE切换失败，业务中断。针对这种情况，可以考虑结合异常情况重新训练模型和推理，或者考虑更换模型，避免后续UE再度出现类似异常情况。
4、CSI-RS反馈增强(CSI-RS Feedback Enhancement)策略，其具体的参数包括但不限于以下的一项或多项：下行信道矩阵。主要流程如下：首先，接入网设备和UE间先交互一个字典，通常是接入网设备根据UE能力，以及本身的要求，提前训练好的模型，然后下发一个Encoder和Quantizer工具给UE；然后，UE可以根据下行信道矩阵，按照已有的字典，将待反馈矩阵进行压缩和量化，并且传递结果给到接入网设备侧；接下来，接入网设备侧根据字典和UE上报的数据，逆向恢复得到原始的信道矩阵。
5、波束管理增强(Beam Management Enhancement)策略，其具体的参数包括但不限于以下的一项或多项：稀疏扫描的矩阵。主要流程如下：接入网设备指示UE按照稀疏扫描的矩阵做波束扫描，UE根据此矩阵做波束扫描，并反馈扫描结果。接入网设备基于UE的稀疏扫描结果，推理出最优的CSI-RS波束，开始对UE进行下一阶段的扫描，由UE反馈最优的CSI-RS波束ID。
6、定位精度增强(Positioning Accuracy Enhancements)策略，其具体的参数包括但不限于以下的一项或多项：UE的位置。相比现有技术确定出的UE的位置，会预测出更精细的位置，例如UE位于市内或室外。
本申请实施例提供了多种通信方法，网元之间可以交互执行策略时的执行状态信息，任一网元可以基于对方执行策略时的执行状态来做出应对策略，这样，多个网元之间可以执行更优的策略，提高网络性能。
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行描述。各个实施例/示例之间可以相互参考。
实施例1：
如图3所示，介绍了一种通信流程示意图。该通信流程可以适用但不限于以下任一通信场景：
第一网元为第一接入网设备，第二网元为核心网设备，第三网元为第二接入网设备。
第一网元为第一接入网设备，第二网元为OAM，第三网元为第二接入网设备。
第一网元为第一终端，第二网元为第一接入网设备，第三网元为第二终端。
第一网元为第一终端，第二网元为第一接入网设备，第三网元为第二接入网设备。
接入网设备与核心网设备或OAM之间通过北向接口或NG接口进行通信，接入网设备与OAM之间可以直接通信，或通过核心网设备(例如AMF)进行通信。接入网设备与接入网设备之间通过Xn接口进行通信。终端与接入网设备之间通过Uu口进行通信。
其中，第一终端或第二终端可以为图1a、图1b介绍的终端设备；第一接入网设备、第二接入网设备可以为图1a介绍的接入网设备，或图1b介绍的gNB-CU、gNB-CU-CP、gNB-CU-CP1、gNB-CU-CP2、gNB-DU；核心网设备可以为图1a、图1b中介绍的任一网元，例如，AMF、UDM、SMF、UPF、PCF、NEF、UDR等。
包括以下步骤：
步骤301：第一网元获取待输入的第一测量量及获取第一AI模型。
AI模型即基于人工智能AI得到的模型。
本申请对于模型训练的主体不进行限制，可以是第一网元基于AI训练第一AI模型，也可以是除第一网元外的其它设备/网元基于AI训练第一AI模型，并将第一AI模型的信息(例如模型的参数，模型参数可以包括但不限于以下的一项或多项输入层、输出层、隐藏层等各个层的神经元的数量，连接关系、每一层的权重值等模型中涉及的参数)发送给第一网元。示例性的，第二网元基于AI训练模型，并将训练好的模型的信息发送给第一网元，第二网元还可以将训练好的模型的信息发送给第三网元和/或其它网元。示例性的，第一网元与第二网元分别基于AI训练各自的模型。示例性的，第一网元基于AI训练模型，并将训练好的模型发送给第二网元。
基于AI对模型进行训练的过程可以参考图2的介绍，不再重复赘述。训练好的模型可以适用于不同的用途，例如，节能、负载均衡、移动性优化、信道状态信息参考信号CSI-RS反馈增强、波束管理增强、定位精度增强等。
当第一网元为第一接入网设备时，第一测量量包括不限于以下的一项或多项：第一接入网设备测量的信息、相邻(与第一接入网设备相邻)接入网设备或相邻小区测量的信息、接入所述第一接入网设备的UE测量的信息，接入相邻接入网设备或相邻小区的UE测量的信息。需要注意是的，UE、相邻接入网设备、相邻小区可以将自身测量的信息发送给第一接入网设备。
当第一网元为第一终端时，第一测量量包括不限于以下的一项或多项：第一终端接入的接入网设备测量的信息、第一终端接入的接入网设备的相邻(与第一接入网设备相邻)接入网设备或相邻小区测量的信息、接入所述第一接入网设备的其它UE测量的信息，接入相邻接入网设备或相邻小区的UE测量的信息。需要注意是的，相邻接入网设备、相邻小区、其它UE可以将自身测量的信息发送给第一终端。
测量的信息例如业务信息、负载信息等。
步骤302：第一网元将所述第一测量量输入至第一AI模型中，得到所述第一AI模型输出的第一策略。
所述第一策略的类型包括以下的任一项：节能策略、负载均衡策略、移动性优化策略、信道状态信息参考信号CSI-RS反馈增强策略、波束管理增强策略、定位精度增强策略。每种类型策略的具体内容(或用途或原理)，可以前文的介绍，不再重复赘述。
步骤303：第一网元执行所述第一策略。
可以理解的是，第一网元在某一时间，可以执行一个或多个策略，第一策略为一个或多个策略中的某一个策略。针对任一策略，第一网元均可以向第二网元发送执行该策略时的执行状态信息，本申请以第一策略为例进行介绍。
步骤304：第一网元向第二网元发送执行所述第一策略时的执行状态信息。
相应的，第二网元接收来自第一网元的所述第一网元执行第一策略时的执行状态信息。
所述执行状态信息用于表示所述第一网元执行所述第一策略时出现异常或未出现异常。
第一网元将自身执行第一策略时的执行状态告知给第二网元，以便于第二网元感知第一网元执行策略时的执行状态，任一网元可以基于对方执行策略时的执行状态来对自身所执行的策略进行自检或调整，这样，多个网元之间可以执行更优的策略，提高网络性能。
第一网元可以周期性地确定执行第一策略的执行状态(即是否正常)，周期可以是1分钟、或5分钟、或10分钟等。或者，第一网元可以基于事件触发(例如某一测量量超过设定阈值)来确定执行第一策略的执行状态(即是否正常)。
第一网元在确定执行状态后，就可以向第二网元发送执行状态信息。或者，若第一网元周期性发送执行状态信息，第一网元在确定执行状态后，可以在下一次最近的上报周期时刻到来时，向第二网元发送执行状态信息。
示例性的，第一网元中预先配置用于确定每种类型的策略是否执行异常的条件(或准则)，第一网元可以基于该条件(或准则)，确定每种类型的策略在被执行时是否出现异常。
当出现以下一项或多项情况(也可以理解为满足以下一项或多项条件)，确定节能策略执行异常：以下的多项情况通常适用于第一网元为第一接入网设备的场景。
相邻小区或相邻接入网设备的负载大于或等于设定负载阈值；例如，设定阈值为最大负载的80％、或85％等。例如相邻小区或相邻接入网设备可以将负载过高的信息发送给第一接入网设备，第一接入网设备接收到来自相邻接入网设备或相邻小区的负载过高的信息，需要第一接入网修改当前的节能策略，以从相邻接入网设备或相邻小区迁移负载到第一接入网设备，降低相邻接入网设备或相邻小区的负载水平。
UE接入数目、业务需求超过当前节能状态下的承受范围，导致UE没有足够的资源接入第一接入网设备或者是接入第一接入网设备的时间过长，例如，终端设备接入所述第一接入网设备的时长值大于或等于设定时长阈值。例如，大于或等于设定数量阈值的终端设备接入所述第一接入网设备时接入失败。设定数量阈值可以是一个具体的数值，例如100、1000等，设定数量阈值也可以是一个比例值，例如第一接入网设备能够接入的UE的最大数量的某一比例。
UE在第一接入网设备的小区内的业务无法正常开展，例如，接入所述第一接入网设备的终端设备的上行速率小于或等于设定第一速率阈值；例如，接入所述第一接入网设备的终端设备的下行速率小于或等于设定第二速率阈值；速率阈值可以针对每个业务进行设置。
基于AI的节能策略效果差，例如，在执行基于AI的节能策略后，第一接入网设备的能耗没有降低到预期，例如预期为非节能状态的50％以下；或者能量效率(系统负载/整体能耗)没有达到预期，例如预期为非节能状态的150％以上；或者没有传统节能策略效果好，例如第一接入网设备按一定周期(例如半小时)，退出基于AI的节能策略，执行传统节能策略时，发现执行传统节能策略能耗低或能效高。例如，所述第一接入网设备的能量效率小于或等于设定效率阈值。再例如，所述第一接入网设备 执行所述第一策略的能量效率小于或等于所述第一接入网设备执行第二策略的能量效率或不执行节能策略的能量效率，其中，所述第二策略为基于非AI得到的节能策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统节能策略，而非基于AI得到的节能策略。
当出现以下一项或多项情况(也可以理解为满足以下一项或多项条件)，确定负载均衡策略执行异常：以下的多项情况通常适用于第一网元为第一接入网设备的场景。
第一接入网设备的实际负载与预测负载差距过大，例如，所述第一接入网设备的实际负载与所述第一接入网设备基于第一策略预测的负载之间的负载差值大于或等于设定第一负载差值阈值。
相邻接入网设备(或相邻小区)的实际负载与预测负载差距过大，例如，相邻小区或相邻接入网设备的实际负载与所述相邻小区或相邻接入网设备预测的负载之间的负载差值大于或等于设定第二负载差值阈值。可以理解的是，相邻接入网设备或相邻小区可以把实际负载与预测负载差距过大的信息发送给第一接入网设备。
所述第一接入网设备的实际负载与相邻接入网设备的实际负载之间的负载差值大于或等于设定第三负载差值阈值。例如，第一接入网设备的实际负载与相邻接入网设备的实际负载之间的负载比值支持续超过设定比值(例如3:1)。可以理解的是，相邻接入网设备或相邻小区可以把实际负载发送给第一接入网设备。
当前基于AI的负载均衡策略效果差，或者不如传统负载均衡策略。例如在执行基于AI的负载均衡策略的过程中，可以周期性或事件触发(例如当第一接入网设备与相邻接入网设备(或相邻小区)负载差距大于一定门限时)回归传统负载均衡策略方案，如果出现传统负载均衡策略方案的负载均衡效果优于基于AI的负载均衡策略方案(例如第一接入网设备与相连接入网设备(或相邻小区)的负载比值更接近1:1)，则可以认为基于AI的负载均衡策略出现异常。例如，第一负载差值大于第二负载差值，其中，所述第一负载差值为所述第一接入网设备执行所述第一策略时所述第一接入网设备的实际负载与相邻接入网设备的实际负载之间的负载差值，所述第二负载差值为所述第一接入网设备执行第二策略或不执行负载均衡策略时所述第一接入网设备与相邻接入网设备之间的负载差值，所述第二策略为基于非AI得到负载均衡策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统负载均衡策略，而非基于AI得到的负载均衡策略。
UE出现RLF或潜在失败(SHR)等问题。在执行负载均衡策略的过程中，主要流程之一就是将UE从负载重的接入网设备切换到负载轻的接入网设备，但同时也要保证UE的业务不受影响且不出现RLF等链路问题。在执行基于AI的负载均衡策略的过程中，如果相比执行前出现了更多的UE掉话，或者潜在失败，或者UE的业务能力(例如吞吐量)出现了下降，则可以认为基于AI的负载均衡策略出现异常。例如，第一吞吐量小于第二吞吐量，其中，所述第一吞吐量为所述第一接入网设备执行所述第一策略时接入所述第一接入网设备的终端设备的吞吐量，所述第二吞吐量为所述第一接入网设备执行第二策略或不执行负载均衡策略时接入所述第一接入网设备的终端设备的吞吐量，所述第二策略为基于非AI得到的负载均衡策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统负载均衡策略，而非基于AI得到的负载均衡策略。例如，第一数量大于第二数量，其中，所述第一数量为所述第一接入网设备执行所述第一策略时终端设备切换过程中出现无线链路失败RLF或出现无线链路潜在失败或出现掉话的终端设备的第一数量，所述第二数量为所述第一接入网设备执行第二策略或不执行负载均衡策略时终端设备切换过程中出现无线链路失败RLF或出现无线链路潜在失败或出现掉话的终端设备的数量，所述第二策略为基于非AI得到的负载均衡策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统负载均衡策略，而非基于AI得到的负载均衡策略。
当出现以下一项或多项情况(也可以理解为满足以下一项或多项条件)，确定移动性优化策略执行异常：以下的多项情况通常适用于第一网元为第一接入网设备的场景。
第一接入网设备接收到的SON等移动性报告显示掉话率没有降低、切换/接入时延没有降低、接入成功率没有提升等。基于UE上报的报告，如果在执行基于AI的移动性优化策略一段时间后(例如5分钟)，UE上报的报告显示，掉话率并没有降低到预设值(例如降低到未执行移动性优化策略时掉话率的50％)，或切换/接入时延并没有降低(例如降低到未执行移动性优化策略时切换/接入时延的50％)，或接入成功率并没有提高到预测值(例如未执行移动性优化策略时的150％)，则可以认为当前的基于 AI的移动性优化策略异常。
例如，接入所述第一接入网设备的终端设备的掉话率大于或等于设定掉话率阈值，例如，该阈值为未执行移动性优化策略时掉话率的50％。例如，终端设备由其他接入网设备切换至所述第一接入网设备的时延大于或等于设定时延阈值，例如，该阈值为未执行移动性优化策略时切换时延的50％。例如，终端设备接入所述第一接入网设备的时延大于或等于设定时延阈值，例如，该阈值为未执行移动性优化策略时接入时延的50％。例如，终端设备接入所述第一接入网设备的接入成功率小于或等于设定阈值，例如该阈值为未执行移动性优化策略时的150％。
不如传统移动性优化策略方案效率好。例如在执行基于AI的移动性优化策略的过程中，可以周期性或事件触发(例如当掉话率高于一定门限时)回归传统移动性优化策略方案，如果出现传统移动性优化策略方案的切换效果好于基于AI的移动性优化策略的切换效果(例如掉话率更低)，则可以认为基于AI的移动性优化策略出现异常。
例如，第一掉话率大于或等于第二掉话率，其中，所述第一掉话率为所述第一接入网设备执行所述第一策略时接入所述第一接入网设备的终端设备的掉话率，第二掉话率为所述第一接入网设备执行第二策略或不执行移动性优化策略时接入所述第一接入网设备的终端设备的掉话率，所述第二策略为基于非AI得到的移动性优化策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统移动性优化策略，而非基于AI得到的移动性优化策略。
例如，第一切换时延大于或等于第二切换时延，其中，所述第一切换时延为所述第一接入网设备执行所述第一策略时终端设备由其他接入网设备切换至所述第一接入网设备的时延，第二切换时延为所述第一接入网设备执行第二策略或不执行移动性优化策略时终端设备由其他接入网设备切换至所述第一接入网设备的时延，所述第二策略为基于非AI得到的移动性优化策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统移动性优化策略，而非基于AI得到的移动性优化策略。
例如，第一接入时延大于或等于第二接入时延，其中，所述第一接入时延为所述第一接入网设备执行所述第一策略时终端设备接入所述第一接入网设备的时延，第二接入时延为所述第一接入网设备执行第二策略或不执行移动性优化策略时终端设备接入所述第一接入网设备的时延，所第二策略为基于非AI得到的移动性优化策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统移动性优化策略，而非基于AI得到的移动性优化策略。
例如，第一接入成功率小于或等于第二接入成功率，其中，所述第一接入成功率为所述第一接入网设备执行所述第一策略时终端设备接入所述第一接入网设备的接入成功率，第二接入成功率为所述第一接入网设备执行第二策略或不执行移动性优化策略时终端设备接入所述第一接入网设备的接入成功率，所第二策略为基于非AI得到的移动性优化策略，所述第二策略与所述第一策略的类型相同；第二策略可以理解为传统移动性优化策略，而非基于AI得到的移动性优化策略。
基于UE的测量信息，判断UE的实际轨迹与预测轨迹不相符等。例如，所述第一接入网设备执行所述第一策略时对终端设备的预测的移动轨迹与所述终端设备的实际移动轨迹不同。例如，预测出UE的轨迹为从小区1到小区2，但根据UE上报的测量结果，UE实际上是从小区1移动至小区3，则对UE的轨迹预测不准。
当出现以下一项或多项情况(也可以理解为满足以下一项或多项条件)，确定CSI-RS反馈增强策略执行异常：以下的多项情况通常适用于第一网元为第一终端或第一接入网设备的场景。
UE下行吞吐量并未达到预期，例如预期为达到传统CSI-RS反馈增强策略方案的120％，或者UE下行吞吐量出现下降，例如降为传统CSI-RS反馈增强策略方案的80％。例如，接入所述第一接入网设备的终端设备的吞吐量小于或等于设定吞吐量阈值。
如果第一网元为第一接入网设备，UE可以把吞吐量发送给第一接入网设备，以便接入网设备进行判断。
当出现以下一项或多项情况(也可以理解为满足以下一项或多项条件)，确定CSI-RS反馈增强策略执行异常：以下的多项情况通常适用于第一网元为第一终端或第一接入网设备的场景。
UE波束扫描时长未降低到预期值，例如预期为传统波束扫描方案的80％，例如，接入所述第一接入网设备的终端设备的波束扫描时长值大于或等于设定扫描时长阈值。
UE波束扫描时长未超过传统波束扫描方案。例如，第一时长大于或等于第二时长，其中，第一时长为所述第一接入网设备执行所述第一策略时接入所述第一接入网设备的终端设备的波束扫描时长，第二时长为所述第一接入网设备执行第二策略或不执行波束管理增强策略时接入所述第一接入网设备的终端设备的波束扫描时长，所述第二策略为基于非AI得到的波束管理增强策略。
UE的接入成功率降低，例如，终端设备接入所述第一接入网设备的接入成功率小于或等于设定成功率阈值；例如阈值为传统波束扫描方案的80％。例如，第一接入成功率小于或等于第二接入成功率，其中，所述第一接入成功率为所述第一接入网设备执行所述第一策略时终端设备接入所述第一接入网设备的接入成功率，第二接入成功率为所述第一接入网设备执行第二策略或不执行波束管理增强策略时终端设备接入所述第一接入网设备的接入成功率，所第二策略为基于非AI得到的波束管理增强策略；
UE的吞吐量降低，例如因为稀疏扫描的目标波束不当导致UE并未选择到最优的波束，与传统波束扫描方案比吞吐量出现下降。例如，接入所述第一接入网设备的终端设备的吞吐量小于或等于设定吞吐量阈值。例如，第一吞吐量小于第二吞吐量，其中，所述第一吞吐量为所述第一接入网设备执行所述第一策略时接入所述第一接入网设备的终端设备的吞吐量，所述第二吞吐量为所述第一接入网设备执行第二策略或不执行波束管理增强策略时接入所述第一接入网设备的终端设备的吞吐量，所述第二策略为基于非AI得到的波束管理增强策略。
如果第一网元为第一接入网设备，UE可以把波束扫描时长值、接入成功率和吞吐量中的一项或多项发送给第一接入网设备，以便接入网设备进行判断。
当出现以下一项或多项情况(也可以理解为满足以下一项或多项条件)，确定定位精度增强策略执行异常：以下的多项情况通常适用于第一网元为第一终端或第一接入网设备的场景。
出现一定比例(例如80％以上)的UE的LOS/NLOS判断结果错误，例如，第一接入网设备对终端设备进行视线传输los或非视线传输Nlos的结果错误的数量大于设定数量阈值。
需要注意的是，以上针对每种类型的策略介绍了用于判断该策略是否出现异常的条件，这些条件只是举例说明，不应造成判断该策略是否出现异常的限定。本申请中多种类型的策略之间，可以互相使用用于判断是否出现异常的条件。另外，可以理解的是，当第一网元无法直接测量到上述介绍的用于判断该策略是否出现异常的条件中的参数时，可以由能够直接测量到这些参数的网元测量后发送给第一网元。
一种具体的示例中，第一网元在确定执行第一策略正常时，向第二网元发送执行第一策略正常的指示信息(即执行状态信息为执行第一策略正常的指示信息)。所述执行所述第一策略正常的指示信息包括但不限于以下的一项或多项：执行正常的指示、用于判断策略正常的测量参数、用于判断策略正常的配置参数、配置参数的生效时间、配置参数的生效时间、第一AI模型的标识(用于标识第一AI模型的，例如编号或索引)、第一AI模型的参数、第一策略的标识(用于标识第一策略，例如编号或索引)、第一策略、第一策略的生效时间。
示例性的，通过1bit或更多比特的取值来表示执行正常或执行异常。例如，当1bit的取值为0时，表示执行正常，当1bit的取值为1时，表示执行异常。
用于判断策略正常常的测量参数可以理解为第一网元执行第一策略时，各个网元(第一网元和/或与第一网元进行通信的其它网元)的性能参数，包括：步骤301中获取的第一测量量中的一项或多项，和/或，上述介绍的确定执行异常时出现的情况(或执行异常时满足的条件)中涉及的一个或多个参数(不包括设定阈值)。示例性的，与执行正常相关的性能参数包括但不限于以下的一项或多项：当前各小区负载，各小区预测负载，切换成功率，接入成功率等。
用于判断策略正常的配置参数可以理解为用于确定是否异常而设定的各种阈值/门限值等，包括但不限于前文介绍的用于确定策略是否执行异常的条件中的一个或多个设定阈值。例如，用于判断策略正常的配置参数包括但不限于以下的一项或多项：设定负载阈值、设定时长阈值、设定数值阈值、设定第一速率阈值、设定第二速率阈值、设定效率阈值、设定第一负载差值阈值、设定第二负载差值阈值、设定第三负载差值阈值、设定吞吐量阈值、设定扫描时长阈值、设定成功率阈值、各种阈值/门限值的生效时间。
一种具体的示例中，第一网元在确定执行第一策略出现异常时，向第二网元发送执行所述第一策略 出现异常的指示信息(即执行状态信息为执行第一策略出现异常的指示信息)。所述执行所述第一策略出现异常的指示信息包括但不限于以下的一项或多项：执行异常的指示、执行异常的时间、执行异常的原因、用于判断策略出现异常的测量参数、用于判断策略出现异常的配置参数、所述配置参数的生效时间、所述第一网元针对执行异常期望采用的修正方式、所述第一网元针对执行异常期望采用的修正方式所需的信息、第一AI模型的标识(用于标识第一AI模型的，例如编号或索引)、第一AI模型的参数、第一策略的标识(用于标识第一策略，例如编号或索引)、第一策略、所述第一策略的生效时间。
执行异常的时间例如为19:00、或下午7:00等。
执行异常的原因可以是某一类型的策略异常，例如，节能策略异常、或负载均衡策略异常、或移动性优化策略、或信道状态信息参考信号CSI-RS反馈增强策略异常、或波束管理增强策略异常、或定位精度增强策略异常。执行异常的原因可以精细到具体的原因，具体的原因可以理解为前文介绍的确定执行异常时出现的情况(或执行异常时满足的条件)。
用于判断策略出现异常的测量参数可以理解为第一网元执行第一策略时，各个网元(第一网元和/或与第一网元进行通信的其它网元)的性能参数，包括：步骤301中获取的第一测量量中的一项或多项，和/或，上述介绍的确定执行异常时出现的情况(或执行异常时满足的条件)中涉及的一个或多个参数(不包括设定阈值)。可以参考前文介绍的与用于判断策略正常的配置参数，不再重复赘述。
用于判断策略出现异常的配置参数可以理解为用于确定是否异常而设定的各种阈值/门限值等，包括但不限于前文介绍的用于确定策略是否执行异常的条件中的一个或多个设定阈值。可以参考前文介绍的用于判断策略正常的配置参数，不再重复赘述。
所述第一网元针对执行异常期望采用的修正方式包括但不限于以下的任一项：
方式一：执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同。可以理解为期望采用的修正方式为执行传统策略(第二策略即为传统策略)，而非基于AI得到的策略。
方式二：执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同。可以理解为期望采用的修正方式为采用新的AI模型确定新的AI策略(第三策略即为新的AI策略)，第三策略可以基于以下方式得到：基于新的第二AI模型，将待输入的第二测量量输入至新的第二AI模型中，得到第二AI模型输出的第三策略。在该方式中，第一网元还可以向第二网元上报第二AI模型的信息，例如第二AI模型的编号或索引、或第二AI模型的参数。其中，得让AI模型可以基于对第一AI模型进行修正得到的，也可以是新训练的AI模型。
方式三：执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同。可以理解为期望采用的修正方式为回退到之前执行的策略(第四策略即为之前执行的策略)。
方式四：不执行与所述第一策略类型相同的任何策略。可以理解为期望采用的修正方式为退出第一策略。
所述第一网元针对执行异常期望采用的修正方式所需的信息可以是用于修正第一AI模型的信息，或用于修正第一策略的信息。可以理解的是，修正后的第一AI模型即为第二AI模型，可以用于确定所述第三策略；修正后的第一策略即为所述第三策略(新的AI策略)。所述第一网元针对执行异常期望采用的修正方式所需的信息包括但不限于以下的一项或多项：新的AI模型的标识(例如索引或编号)、新的AI模型的参数、新的AI策略(即第三策略的)的标识(例如策略的编号或索引)、新的AI策略、其它网元的性能信息、UE历史轨迹、邻区负载信息、网络配置(例如网络节能配置)等。
可选的，步骤305：第二网元向第一网元发送第一信息。
相应的，第一网元接收来自第二网元的第一信息。
步骤305为可选的步骤，可以不执行，也可以与后续的步骤309合并执行。
第一信息可以指示以下的一项或多项：允许第一网元针对执行异常进行修正、不允许第一网元针对执行异常进行修正、针对执行异常允许采用的修正方式、需要第一网元再次发送的信息(例如第一测量量的信息、模型相关信息，在该情况下，第一网元可以再次向第二网元发送对应的信息)。
可以通过1bit甚至更多比特的取值来表示允许修正或不允许进行修整。例如，当1bit的取值为0时，表示允许进行修正，当1bit的取值为1时，表示不允许进行修正。
如果第一网元上报了第一网元期望采用的修正方式，则第二网元针对执行异常允许采用的修正方式 可以是第一网元期望采用的修正方式，当然，也可以不是第一网元期望采用的修正方式。第二网元可以基于第一网元期望采用的修正方式和/或其它网元的性能参数等确定允许采用的修正方式。
所述允许采用的修正方式包括但不限于以下任一方式：
方式一：执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同。可以理解为允许采用的修正方式为执行传统策略(第二策略即为传统策略)，而非基于AI得到的策略。
第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元跳过所述第一策略，执行所述第二策略。
方式二：执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同。可以理解为允许采用的修正方式为采用新的AI模型(第二模型即为新的AI模型)确定新的AI策略(第三策略即为新的AI策略)，第三策略可以基于以下方式得到：将待输入的第二测量量输入至第二AI模型中，得到第二AI模型输出的第三策略。第二AI模型可以是在第一AI模型的基础上进行修正(训练)得到的AI模型，也可以是重新训练的AI模型。
第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元跳过所述第一策略，执行所述第三策略。
方式三：执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同。可以理解为允许采用的修正方式为回退到之前执行的策略(第四策略即为之前执行的策略)。
第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元跳过所述第一策略，执行所述第四策略。
方式四：不执行与所述第一策略类型相同的任何策略。可以理解为允许采用的修正方式为退出第一策略。
第一网元基于所述修正方式对所述第一策略产生的异常进行修正包括：第一网元跳过所述第一策略。
可以通过2bit甚至更多比特的取值来表示允许采用的修正方式。
在所述允许采用的修正方式为执行第三策略(新的AI策略)时，所述第一信息包括但不限于以下至少一项：所述第三策略的索引、所述第三策略、所述第二AI模型的索引、所述第二AI模型的参数或者所述第三网元的性能信息；其中，所述第三网元的性能信息用于确定所述第三策略。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第三策略的标识时，执行所述第三策略的标识指示的所述第三策略，以对所述第一策略产生的异常进行修正。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第三策略时，执行所述第三策略，以对所述第一策略产生的异常进行修正。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第二AI模型的标识时，将第二测量量输入所述第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正。所述第二测量量与所述第一测量量可以相同，也可以不同。其中，不同可以是指测量量中包括的测量信息不同，例如第一测量量包括第一接入网设备的测量信息和UE的测量信息，第二测量量包括UE的测量信息，不包括第一接入网设备的测量信息。不同可以是指测量量的取值不同，例如针对负载信息，第一测量量对应的负载信息为80％，第二测量量对应的负载信息为85％。通常，所述第二测量量的测量时间不早于所述第一测量量的测量时间。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第二AI模型的参数时，将第二测量量输入所述参数对应的第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正。
在一种可能的实现中，第一网元基于所述修正方式对所述第一策略产生的异常进行修正，包括：当所述第一信息包括所述第三网元的性能信息时，基于所述性能信息修正所述第一AI模型，得到第二AI模型；将第三测量量输入所述第二AI模型，得到所述第三策略；并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第三测量量与所述第一测量量相同，也可以不同，其中，不同可以是指测量量中包括的测量信息不同和/或测量量的取值不同，可以参考第一测量量与第二测量量不同 的介绍，原理类似，不再重复赘述。通常，所述第三测量量的测量时间不早于所述第一测量量的测量时间。
一种可能的实现方式，基于性能信息修正第一AI模型包括：将性能信息作为训练数据，对第一AI模型进行再次训练得到第二AI模型。
第一网元在确定执行第一策略出现异常时，可以等待第二网元通知允许采用的修正方式；当接收到来自第二网元的允许采用的修正方式后，基于该允许采用的修正方式进行修整。或者，第一网元在确定执行第一策略出现异常时，无需等待第二网元通知允许采用的修正方式，而自身来决定采用的修正方式，并基于自身决定采用的修正方式进行修正。再或者，第一网元在确定执行第一策略出现异常时，第一网元可以先自身来决定采用的修正方式，并基于自身决定采用的修正方式进行修正，当接收到来自第二网元的允许采用的修正方式后，再基于该允许采用的修正方式进行修正。
步骤305与步骤306的先后顺序不进行限制。
步骤306：第二网元在基于所述执行状态信息确定所述第一网元执行所述第一策略出现异常时，第二网元向第三网元发送第二信息。
相应的，第三网元接收来自第二网元的第二信息。
若规定只有在执行策略异常时，才会发送执行状态信息，则执行状态信息为执行第一策略出现异常的指示信息，则第二网元接收到执行第一策略出现异常的指示信息后，可以向第三网元发送第二信息，可以省略基于执行状态信息确定所述第一网元执行所述第一策略是否出现异常的过程。
所述第二信息用于指示以下的一项或多项：
第一策略被执行时出现异常、第一AI模型输出的第一类型的策略被执行时出现异常(所述第一策略的类型为所述第一类型)、第一网元的标识信息(例如小区ID)、第一网元向第二网元发送的一项或多项信息、第二网元向第一网元发送的一项或多项信息、第三网元对当前执行的策略进行自检、对第三网元的策略指导方案、需要第三网元发送的信息。
第三网元的策略指导方案与第二网元向第一网元发送的允许采用的修正方式类似，例如包括以下的任一项：执行基于非AI得到的策略(可以理解为传统策略)、执行新的AI策略、执行之前的策略、退出当前的策略。
需要第三网元发送的信息可以理解为第二网元请求的信息，包括但不限于：第三网元的性能信息、第一网元向第二网元请求的信息。第一网元向第二网元请求的信息包括但不限于第一网元针对执行异常期望采用的修正方式所需的信息。
一种示例中，所述第三网元为所述第二网元管理或服务的任一网元；一种示例中，所述第三网元正在执行所述第一AI模型输出的策略(即第一网元与第二网元为使用同一模型的网元)；一种示例中，所述第三网元正在执行所述第一策略；一种示例中，所述第三网元正在执行第五策略，所述第五策略与所述第一策略的类型相同，第五策略可以是基于AI得到的策略，也可以是基于非AI得到的策略(即传统策略)。一种示例中，第三网元为第二网元与第一网元有关联、且存在潜在AI策略执行异常的网元。
步骤307与步骤308的先后顺序不进行限制。
可选的，步骤307：第三网元在接收来自第二网元的第二信息后，确定自身执行的策略是否出现异常。
第二信息可以理解为警告信息。第三网元根据第二网元的警告信息，进行自检，排除风险。
可选的，步骤308：第三网元在接收来自第二网元的第二信息后，第三网元向第二网元发送第二网元请求的信息(即需要第三网元发送的信息)。
相应的，第二网元接收来自第三网元的第二网元请求的信息(即需要第三网元发送的信息)。
可选的，步骤309：第二网元向第一网元发送第一网元需要的信息。
相应的，第一网元接收来自第二网元的第一网元需要的信息。
第二网元可以基于步骤304中执行第一策略异常的指示信息和/或步骤308中第三网元向第二网元发送的第二网元请求的信息(即需要第三网元发送的信息)，向第一网元发送第一网元需要的信息。第一网元需要的信息可以参考步骤305中第一信息的内容。
一种具体的示例中，第二网元可以根据第三网元发送的第二网元请求的信息(即需要第三网元发送的信息)，确定(或重新确定)允许第一网元针对执行异常进行修正、或不允许第一网元针对执行异常 进行修正、或针对执行异常允许采用的修正方式，并向第一网元指示允许第一网元针对执行异常进行修正、或不允许第一网元针对执行异常进行修正、或针对执行异常允许采用的修正方式。可以参考前文步骤305中有关第一信息的描述，不再重复赘述。
实施例2：与实施例1的区别包括：第一网元与第三网元之间直接通信，无需第二网元的参与。
如图4所示，介绍了一种通信流程示意图。该通信流程可以适用但不限于以下任一通信场景：
第一网元为第一终端、第三网元为第二终端。
第一网元为第一接入网设备，第三网元为第二接入网设备。
接入网设备与接入网设备之间通过Xn接口进行通信。终端与终端之间通过Sidelink接口，例如PC5接口进行通信。
其中，第一终端或第二终端可以为图1a、图1b介绍的终端设备；第一接入网设备、第二接入网设备可以为图1a介绍的接入网设备，或图1b介绍的gNB-CU、gNB-CU-CP、gNB-CU-CP1、gNB-CU-CP2、gNB-DU。
包括以下步骤：
步骤401：获取待输入的第一测量量、及获取第一AI模型。
AI模型为基于人工智能AI得到的模型。
步骤402：将所述待输入的第一测量量输入至第一AI模型中，得到所述第一AI模型输出的第一策略。
步骤403：执行所述第一策略。
步骤404：第一网元向第三网元发送执行所述第一策略时的执行状态信息。
相应的，第三网元接收来自第一网元的所述第一网元执行第一策略时的执行状态信息。
所述执行状态信息用于表示所述第一网元执行所述第一策略时出现异常或未出现异常。
步骤401至步骤404的过程可以参考步骤301至步骤304的过程，不再重复赘述。
需要注意的是，步骤404与步骤304的不同之处包括：第一网元向第三网元发送的执行状态信息可以包括步骤304中第一网元向第二网元发送的执行状态信息中的一项或多项，和/或，步骤306中第二网元向第三网元发送的第二信息中的一项或多项。
一种具体的示例中，第一网元向第三网元发送的执行状态信息包括但不限于以下的一项或多项：
执行异常的指示、执行异常的时间、执行异常的原因、用于判断策略出现异常的测量参数、用于判断策略出现异常的配置参数、配置参数的生效时间、所述第一网元针对执行异常期望采用的修正方式、所述第一网元针对执行异常期望采用的修正方式所需的信息、所述第一AI模型的标识、所述第一AI模型的参数、所述第一策略的标识、所述第一策略或者所述第一策略的生效时间、第一网元的标识信息、第三网元对当前执行的策略进行自检、对第三网元的策略指导方案。
可选的，步骤405：第三网元在接收来自第一网元的执行状态信息后，确定自身执行的策略是否出现异常。
进一步可选的，第三网元在确定执行状态信息为执行第一策略异常的指示信息时，确定自身执行的策略是否出现异常。
可选的，步骤406：第三网元在接收来自第一网元的执行状态信息后，第三网元向第一网元发送执行状态信息中第一网元请求的信息(即需要第三网元发送的信息)。
进一步可选的，第三网元在确定执行状态信息为执行第一策略异常的指示信息时，向第一网元发送执行状态信息中第一网元请求的信息(即需要第三网元发送的信息)。
示例性的，需要第三网元发送的信息包括但不限于以下的一项或多项：
允许第一网元针对执行异常进行修正、不允许第一网元针对执行异常进行修正、针对执行异常允许采用的修正方式、需要第一网元再次发送的信息(例如测量量信息、模型相关信息，在该情况下，第一网元可以再次向第三网元发送对应的信息)、所述第一网元针对执行异常期望采用的修正方式所需的信息。所述第一网元针对执行异常期望采用的修正方式所需的信息包括但不限于以下的一项或多项：新的AI模型的标识(例如索引或编号)、新的AI模型的参数、新的AI策略(即第三策略的)的标识(例如策略的编号或索引)、新的AI策略、其它网元的性能信息、UE历史轨迹、邻区负载信息、网络配置(例如网络节能配置)等。
可选的，第一网元还可以基于所述第三网元的发送的第一网元请求的信息，对所述第一策略进行修 正或对所述第一AI模型进行修正。
第一网元对第一策略进行修正后，得到第二策略，第一网元可以跳过第一策略，执行第二策略。
第一网元对第一AI模型进行修正得到第二AI模型；将测量量输入所述第二AI模型，得到所述第三策略；第一网元可以跳过所述第一策略，执行所述第三策略。其中，基于性能信息修正第一AI模型的一种方式为，将性能信息作为训练数据，对第一AI进行再次训练得到第二AI模型。
可选的，第三网元或第一网元还可以将第一网元与第三网元之间交互的信息发送给其它网元(例如核心网设备、OAM等)。
实施例3：
第一网元在与第二网元进行通信时，也可以与第三网元进行通信。即对于第一网元来说，可以同时执行图3的流程和图4的流程。一种可能的实现方式中，当第一网元同时执行步骤304和步骤404时，步骤304仅仅起到通知的作用，无需第二网元向第一网元进行反馈，例如，无需执行步骤305或步骤309。另一种可能的实现方式中，当第一网元同时执行步骤304和步骤404时，步骤404仅仅起到通知的作用，无需第三网元向第一网元进行反馈，例如，无需执行步骤406。另一种可能的实现方式中，当第一网元同时执行步骤304和步骤404时，第一网元可以基于第二网元的反馈和第三网元的反馈，来做出决定。
前文介绍了本申请实施例的方法，下文中将介绍本申请实施例中的装置。方法、装置是基于同一技术构思的，由于方法、装置解决问题的原理相似，因此装置与方法的实施可以相互参见，重复之处不再赘述。
本申请实施例可以根据上述方法示例，对装置进行功能模块的划分，例如，可以对应各个功能划分为各个功能模块，也可以将两个或两个以上的功能集成在一个模块中。这些模块既可以采用硬件的形式实现，也可以采用软件功能模块的形式实现。需要说明的是，本申请实施例中对模块的划分是示意性的，仅仅为一种逻辑功能划分，具体实现时可以有另外的划分方式。
基于与上述方法的同一技术构思，参见图5，提供了一种通信装置500结构示意图，该通信装置500可以包括：处理模块510，可选的，还包括接收模块520a、发送模块520b、存储模块530。处理模块510可以分别与存储模块530和接收模块520a和发送模块520b相连，所述存储模块530也可以与接收模块520a和发送模块520b相连。
在一种示例中，上述的接收模块520a和发送模块520b也可以集成在一起，定义为收发模块。
在一种示例中，该通信装置500可以为第一网元，也可以为应用于第一网元中的芯片或功能单元。该通信装置500具有上述方法中第一网元的任意功能，例如，该通信装置500能够执行上述图3、图4的方法中由第一网元执行的各个步骤。
所述接收模块520a，可以执行上述方法实施例中第一网元执行的接收动作。
所述发送模块520b，可以执行上述方法实施例中第一网元执行的发送动作。
所述处理模块510，可以执行上述方法实施例中第一网元执行的动作中，除发送动作和接收动作外的其它动作。
在一种示例中，所述处理模块510，用于获取待输入的第一测量量、以及获取第一人工智能AI模型；将所述第一测量量输入至所述第一AI模型中，得到所述第一AI模型输出的第一策略；执行所述第一策略；在确定执行所述第一策略出现异常时，所述发送模块520b，用于向第二网元发送执行所述第一策略出现异常的指示信息。
在一种示例中，所述第一策略的类型包括以下的任一项：节能策略、负载均衡策略、移动性优化策略、信道状态信息参考信号CSI-RS反馈增强策略、波束管理增强策略、定位精度增强策略。
在一种示例中，所述执行所述第一策略出现异常的指示信息包括以下的一项或多项：执行异常的指示、执行异常的时间、执行异常的原因、用于判断策略出现异常的测量参数、用于判断策略出现异常的配置参数、配置参数的生效时间、所述第一网元针对执行异常期望采用的修正方式、所述第一网元针对执行异常期望采用的修正方式所需的信息、所述第一AI模型的标识、所述第一AI模型的参数、所述第一策略的标识、所述第一策略或者所述第一策略的生效时间。
在一种示例中，所述接收模块520a，用于接收来自所述第二网元的第一信息，所述第一信息用于 指示针对执行异常允许采用的修正方式；所述处理模块510，用于基于所述修正方式对所述第一策略产生的异常进行修正。
在一种示例中，所述允许采用的修正方式包括以下任一项：执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同；执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同；执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同。
在一种示例中，在所述允许采用的修正方式为执行第三策略时，所述第一信息包括以下至少一项：所述第三策略的标识、所述第三策略、所述第二AI模型的标识、所述第二AI模型的参数或者所述第三网元的性能信息；其中，所述第三网元的性能信息用于确定所述第三策略。
在一种示例中，所述处理模块510，具体用于当所述第一信息包括所述第三策略的标识时，执行所述第三策略的标识指示的所述第三策略，以对所述第一策略产生的异常进行修正；或者，当所述第一信息包括所述第三策略时，执行所述第三策略，以对所述第一策略产生的异常进行修正；或者，当所述第一信息包括所述第二AI模型的标识时，将第二测量量输入所述第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第二测量量与所述第一测量量相同，或所述第二测量量的测量时间不早于所述第一测量量的测量时间；或者，当所述第一信息包括所述第二AI模型的参数时，将第二测量量输入所述参数对应的第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第二测量量与所述第一测量量相同，或所述第二测量量的测量时间不早于所述第一测量量的测量时间；或者，当所述第一信息包括所述第三网元的性能信息时，基于所述性能信息修正所述第一AI模型，得到第二AI模型；将第三测量量输入所述第二AI模型，得到所述第三策略；并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第三测量量与所述第一测量量相同，或所述第三测量量的测量时间不早于所述第一测量量的测量时间。
在一种示例中，所述装置为第一接入网设备，所述第二网元为核心网网元、或操作、管理与维护OAM网元；或者，所述装置为第一终端设备，所述第二网元为接入网设备；或者，所述装置为第一接入网设备，第二网元为第二接入网设备；或者，所述装置为第一终端设备，第二网元为第二终端设备。
在一种示例中，在确定执行所述第一策略出现异常时，所述发送模块520b，用于向第三网元发送执行所述第一策略出现异常的指示信息；所述接收模块520a，用于接收来自第三网元的性能信息；所述处理模块510，用于基于所述第三网元的性能信息，对所述第一策略进行修正或对所述第一AI模型进行修正。
在一种示例中，所述存储模块530，可以存储第一网元执行的方法的计算机执行指令，以使处理模块510和接收模块520a和发送模块520b执行上述示例中第一网元执行的方法。
在一种示例中，该通信装置500可以为第二网元，也可以为应用于第二网元中的芯片或功能单元。该通信装置500具有上述方法中第二网元的任意功能，例如，该通信装置500能够执行上述图3的方法中由第二网元执行的各个步骤。
所述接收模块520a，可以执行上述方法实施例中第二网元执行的接收动作。
所述发送模块520b，可以执行上述方法实施例中第二网元执行的发送动作。
所述处理模块510，可以执行上述方法实施例中第二网元执行的动作中，除发送动作和接收动作外的其它动作。
在一种示例中，所述接收模块520a，用于接收来自第一网元的所述第一网元执行第一策略出现异常的指示信息，所述第一策略为第一网元将待输入的第一测量量输入至第一AI模型后所述第一AI模型输出的策略；所述发送模块520b，用于向第三网元发送第二信息，所述第二信息用于指示第一策略被执行时出现异常或第一AI模型输出的第一类型的策略被执行时出现异常，所述第一策略的类型为所述第一类型。
在一种示例中，所述第三网元满足以下任一条件：所述第三网元正在执行所述第一AI模型输出的策略；或者，所述第三网元正在执行所述第一策略；或者，所述第三网元正在执行第五策略，所述第五策略与所述第一策略的类型相同。
在一种示例中，所述处理模块510，用于确定针对执行异常允许采用的修正方式；所述发送模块 520b，还用于向所述第一网元发送第一信息，所述第一信息用于指示针对执行异常允许采用的修正方式。
在一种示例中，所述允许采用的修正方式包括：执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同；或者，执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同；或者，执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同；不执行任何策略。
在一种示例中，在所述允许采用的修正方式为执行第三策略时，所述第一信息包括以下至少一项：所述第三策略的标识、所述第三策略、所述第二AI模型的标识、所述第二AI模型的参数或者所述第三网元的性能信息；其中，所述第三网元的性能信息用于确定所述第三策略。
在一种示例中，所述接收模块520a，用于接收来自所述第三网元的性能信息；所述处理模块510，用于将所述第三网元的性能信息确定为用于修正所述第一AI模型的信息；或者，将所述第三网元的性能信息确定为用于修正所述第一策略的信息；或者，基于所述第三网元的性能信息及AI确定所述第二AI模型的信息；或者，基于所述第三网元的性能信息及AI确定所述第三策略的信息。
在一种示例中，所述接收模块520a，还用于接收第三网元的性能信息。
在一种示例中，所述存储模块530，可以存储第二网元执行的方法的计算机执行指令，以使处理模块510和接收模块520a和发送模块520b执行上述示例中第二网元执行的方法。
在一种示例中，该通信装置500可以为第三网元，也可以为应用于第三网元中的芯片或功能单元。该通信装置500具有上述方法中第三网元的任意功能，例如，该通信装置500能够执行上述图3、图4的方法中由第三网元执行的各个步骤。
所述接收模块520a，可以执行上述方法实施例中第三网元执行的接收动作。
所述发送模块520b，可以执行上述方法实施例中第三网元执行的发送动作。
所述处理模块510，可以执行上述方法实施例中第三网元执行的动作中，除发送动作和接收动作外的其它动作。
在一种示例中，所述接收模块520a，用于接收来自第二网元的第二信息，所述第二信息用于指示第一策略被执行时出现异常或第一AI模型输出的第一类型的策略被执行时出现异常，所述第一策略属于所述第一类型；所述发送模块520b，用于向第二网元发送所述第三网元的性能信息；所述处理模块510，用于确定所述第三网元执行的策略是否出现异常。
在一种示例中，所述接收模块520a，用于接收来自第一网元的所述第一网元执行第一策略出现异常的指示信息，所述第一策略为第一网元将待输入的第一测量量输入至第一AI模型后所述第一AI模型输出的策略；所述发送模块520b，用于向所述第一网元发送所述第三网元的性能信息，所述处理模块510，用于确定所述第三网元执行的策略是否出现异常。
在一种示例中，所述存储模块530，可以存储第三网元执行的方法的计算机执行指令，以使处理模块510和接收模块520a和发送模块520b执行上述示例中第三网元执行的方法。
示例的，存储模块可以包括一个或者多个存储器，存储器可以是一个或者多个设备、电路中用于存储程序或者数据的器件。存储模块可以是寄存器、缓存或者RAM等，存储模块可以和处理模块集成在一起。存储模块可以是ROM或者可存储静态信息和指令的其他类型的静态存储设备，存储模块可以与处理模块相独立。
所述收发模块可以是输入或者输出接口、管脚或者电路等。
作为一种可能的产品形态，装置可以由一般性的总线体系结构来实现。
如图6所示，提供了一种通信装置600的示意性框图。
该通信装置600可以包括：处理器610，可选的，还包括收发器620、存储器630。该收发器620，可以用于接收程序或指令并传输至所述处理器610，或者，该收发器620可以用于该通信装置600与其他通信设备进行通信交互，比如交互控制信令和/或业务数据等。该收发器620可以为代码和/或数据读写收发器，或者，该收发器620可以为处理器与收发机之间的信号传输收发器。所述处理器610和所述存储器630之间电耦合。
一种示例中，该通信装置600可以为第一网元，也可以为应用于第一网元中的芯片。应理解，该装置具有上述方法中第一网元的任意功能，例如，所述通信装置600能够执行上述图3、图4的方法中由 第一网元执行的各个步骤。示例的，所述存储器630，用于存储计算机程序；所述处理器610，可以用于调用所述存储器630中存储的计算机程序或指令，执行上述示例中第一网元执行的方法，或者通过所述收发器620执行上述示例中第一网元执行的方法。
一种示例中，该通信装置600可以为第二网元，也可以为应用于第二网元中的芯片。应理解，该装置具有上述方法中第二网元的任意功能，例如，所述通信装置600能够执行上述图3的方法中由第二网元执行的各个步骤。示例的，所述存储器630，用于存储计算机程序；所述处理器610，可以用于调用所述存储器630中存储的计算机程序或指令，执行上述示例中第二网元执行的方法，或者通过所述收发器620执行上述示例中第二网元执行的方法。
一种示例中，该通信装置600可以为第三网元，也可以为应用于第三网元中的芯片。应理解，该装置具有上述方法中第三网元的任意功能，例如，所述通信装置600能够执行上述图3、图4的方法中由第三网元执行的各个步骤。示例的，所述存储器630，用于存储计算机程序；所述处理器610，可以用于调用所述存储器630中存储的计算机程序或指令，执行上述示例中第三网元执行的方法，或者通过所述收发器620执行上述示例中第三网元执行的方法。
图5中的处理模块510可以通过所述处理器610来实现。
图5中的接收模块520a和发送模块520b可以通过所述收发器620来实现。或者，收发器620分为接收器和发送器，接收器执行接收模块的功能，发送器执行发送模块的功能。
图5中的存储模块530可以通过所述存储器630来实现。
作为一种可能的产品形态，装置可以由通用处理器(通用处理器也可以称为芯片或芯片系统)来实现。
一种可能的实现方式中，实现应用于第一网元的装置或第二网元或第三网元的装置的通用处理器包括：处理电路(处理电路也可以称为处理器)；可选的，还包括：与所述处理电路内部连接通信的输入输出接口、存储介质(存储介质也可以称为存储器)，所述存储介质用于存储处理电路执行的指令，以执行上述示例中第一网元或第二网元或第三网元执行的方法。
图5中的处理模块510可以通过处理电路来实现。
图5中的接收模块520a和发送模块520b可以通过输入输出接口来实现。或者，输入输出接口分为输入接口和输出接口，输入接口执行接收模块的功能，输出接口执行发送模块的功能。
图5中的存储模块530可以通过存储介质来实现。
作为一种可能的产品形态，本申请实施例的装置，还可以使用下述来实现：一个或多个FPGA(现场可编程门阵列)、PLD(可编程逻辑器件)、控制器、状态机、门逻辑、分立硬件部件、任何其它适合的电路、或者能够执行本申请通篇所描述的各种功能的电路的任意组合。
本申请实施例还提供了一种计算机可读存储介质，存储有计算机程序，该计算机程序被计算机执行时，可以使得所述计算机用于执行上述通信方法。或者说：所述计算机程序包括用于实现上述通信的方法的指令。
本申请实施例还提供了一种计算机程序产品，包括：计算机程序代码，当所述计算机程序代码在计算机上运行时，使得计算机可以执行上述提供的通信的方法。
本申请实施例还提供了一种通信的系统，所述通信系统包括：执行上述通信方法的第一网元、第二网元、第三网元中的一项或多项。
另外，本申请实施例中提及的处理器可以是中央处理器(central processing unit，CPU)，基带处理器，基带处理器和CPU可以集成在一起，或者分开，还可以是网络处理器(network processor，NP)或者CPU和NP的组合。处理器还可以进一步包括硬件芯片或其他通用处理器。上述硬件芯片可以是专用集成电路(application-specific integrated circuit，ASIC)，可编程逻辑器件(programmable logic device，PLD)或其组合。上述PLD可以是复杂可编程逻辑器件(complex programmable logic device，CPLD)，现场可编程逻辑门阵列(field-programmable gate array，FPGA)，通用阵列逻辑(generic array logic，GAL)及其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件等或其任意组合。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。
本申请实施例中提及的存储器可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性 存储器两者。其中，非易失性存储器可以是只读存储器(Read-Only Memory，ROM)、可编程只读存储器(Programmable ROM，PROM)、可擦除可编程只读存储器(Erasable PROM，EPROM)、电可擦除可编程只读存储器(Electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(Random Access Memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用，例如静态随机存取存储器(Static RAM，SRAM)、动态随机存取存储器(Dynamic RAM，DRAM)、同步动态随机存取存储器(Synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(Double Data Rate SDRAM，DDR SDRAM)、增强型同步动态随机存取存储器(Enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(Synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(Direct Rambus RAM，DR RAM)。应注意，本申请描述的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
本申请实施例中提及的收发器中可以包括单独的发送器，和/或，单独的接收器，也可以是发送器和接收器集成一体。收发器可以在相应的处理器的指示下工作。可选的，发送器可以对应物理设备中发射机，接收器可以对应物理设备中的接收机。
本领域普通技术人员可以意识到，结合本文中所公开的实施例中描述的各方法步骤和单元，能够以电子硬件、计算机软件或者二者的结合来实现，为了清楚地说明硬件和软件的可互换性，在上述说明中已经按照功能一般性地描述了各实施例的步骤及组成。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。本领域普通技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，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，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另外，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口、装置或单元的间接耦合或通信连接，也可以是电的，机械的或其它的形式连接。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本申请实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各个单元单独物理存在，也可以是两个或两个以上单元集成在一个单元中。上述集成的单元既可以采用硬件的形式实现，也可以采用软件功能单元的形式实现。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分，或者该技术方案的全部或部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(read-only memory，ROM)、随机存取存储器(random access memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
本申请中的“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这三种情况。字符“/”一般表示前后关联对象是一种“或”的关系。本申请中所涉及的多个，是指两个或两个以上。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。
尽管已描述了本申请的优选实施例，但本领域内的技术人员一旦得知了基本创造性概念，则可对这些实施例作出另外的变更和修改。所以，所附权利要求意欲解释为包括优选实施例以及落入本申请范围的所有变更和修改。
显然，本领域的技术人员可以对本申请实施例进行各种改动和变型而不脱离本申请实施例的精神和范围。这样，倘若本申请实施例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，则本申请也意图包括这些改动和变型在内。

Claims (27)

  1. 一种通信方法，其特征在于，应用于第一网元，包括：
    获取待输入的第一测量量、以及获取第一人工智能AI模型；
    将所述第一测量量输入至所述第一AI模型中，得到所述第一AI模型输出的第一策略；
    执行所述第一策略；
    在确定执行所述第一策略出现异常时，向第二网元发送指示信息，所述指示信息用于指示执行所述第一策略出现异常。
  2. 如权利要求1所述的方法，其特征在于，所述第一策略的类型包括以下的任一项：
    节能策略、负载均衡策略、移动性优化策略、信道状态信息参考信号CSI-RS反馈增强策略、波束管理增强策略或者定位精度增强策略。
  3. 如权利要求1或2所述的方法，其特征在于，所述执行所述第一策略出现异常的指示信息包括以下的一项或多项：
    执行异常的指示、执行异常的时间、执行异常的原因、用于判断策略出现异常的测量参数、用于判断策略出现异常的配置参数、配置参数的生效时间、所述第一网元针对执行异常期望采用的修正方式、所述第一网元针对执行异常期望采用的修正方式所需的信息、所述第一AI模型的标识、所述第一AI模型的参数、所述第一策略的标识、所述第一策略或者所述第一策略的生效时间。
  4. 如权利要求1-3任一项所述的方法，其特征在于，还包括：
    接收来自所述第二网元的第一信息，所述第一信息用于指示针对执行异常允许采用的修正方式；
    基于所述修正方式对所述第一策略产生的异常进行修正。
  5. 如权利要求4所述的方法，其特征在于，所述允许采用的修正方式包括以下任一项：
    执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同；
    执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同；或者，
    执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同。
  6. 如权利要求5所述的方法，其特征在于，在所述允许采用的修正方式为执行第三策略时，所述第一信息包括以下至少一项：
    所述第三策略的标识、所述第三策略、所述第二AI模型的标识、所述第二AI模型的参数或者所述第三网元的性能信息；其中，所述第三网元的性能信息用于确定所述第三策略。
  7. 如权利要求6所述的方法，其特征在于，基于所述修正方式对所述第一策略产生的异常进行修正，包括：
    当所述第一信息包括所述第三策略的标识时，执行所述第三策略的标识指示的所述第三策略，以对所述第一策略产生的异常进行修正；或者，
    当所述第一信息包括所述第三策略时，执行所述第三策略，以对所述第一策略产生的异常进行修正；或者，
    当所述第一信息包括所述第二AI模型的标识时，将第二测量量输入所述第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第二测量量与所述第一测量量相同，或所述第二测量量的测量时间不早于所述第一测量量的测量时间；或者，
    当所述第一信息包括所述第二AI模型的参数时，将第二测量量输入所述参数对应的第二AI模型，得到所述第三策略，并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第二测量量与所述第一测量量相同，或所述第二测量量的测量时间不早于所述第一测量量的测量时间；或者，
    当所述第一信息包括所述第三网元的性能信息时，基于所述性能信息修正所述第一AI模型，得到第二AI模型；将第三测量量输入所述第二AI模型，得到所述第三策略；并执行所述第三策略，以对所述第一策略产生的异常进行修正，其中，所述第三测量量与所述第一测量量相同，或所述第三测量量的测量时间不早于所述第一测量量的测量时间。
  8. 如权利要求1-7任一项所述的方法，其特征在于，所述第一网元为第一接入网设备，所述第二网元为核心网网元、或操作、管理与维护OAM网元；或者，
    所述第一网元为第一终端设备，所述第二网元为接入网设备；或者，
    所述第一网元为第一接入网设备，所述第二网元为第二接入网设备；或者，
    所述第一网元为第一终端设备，所述第二网元为第二终端设备。
  9. 一种通信方法，其特征在于，应用于第二网元，包括：
    接收来自第一网元的指示信息，所述指示信息用于指示所述第一网元执行第一策略出现异常，所述第一策略为第一网元将待输入的第一测量量输入至第一人工智能AI模型后所述第一AI模型输出的策略；
    向第三网元发送第二信息，所述第二信息用于指示所述第一策略被执行时出现异常或第一AI模型输出的第一类型的策略被执行时出现异常，所述第一策略的类型为所述第一类型。
  10. 如权利要求9所述的方法，其特征在于，所述第一策略的类型包括以下的任一项：
    节能策略、负载均衡策略、移动性优化策略、信道状态信息参考信号CSI-RS反馈增强策略、波束管理增强策略或者定位精度增强策略。
  11. 如权利要求9或10所述的方法，其特征在于，所述执行所述第一策略出现异常的指示信息包括以下的一项或多项：
    执行异常的指示、执行异常的时间、执行异常的原因、用于判断策略出现异常的测量参数、用于判断策略出现异常的配置参数、配置参数的生效时间、所述第一网元针对执行异常期望采用的修正方式、所述第一网元针对执行异常期望采用的修正方式所需的信息、所述第一AI模型的标识、所述第一AI模型的参数、所述第一策略的标识、所述第一策略或者所述第一策略的生效时间。
  12. 如权利要求9-11任一项所述的方法，其特征在于，所述第三网元满足以下任一条件：
    所述第三网元正在执行所述第一AI模型输出的策略；
    所述第三网元正在执行所述第一策略；或者，
    所述第三网元正在执行第五策略，所述第五策略与所述第一策略的类型相同。
  13. 如权利要求9-12任一项所述的方法，其特征在于，还包括：
    确定针对执行异常允许采用的修正方式；
    向所述第一网元发送第一信息，所述第一信息用于指示针对执行异常允许采用的修正方式。
  14. 如权利要求13所述的方法，其特征在于，所述允许采用的修正方式包括以下任一项：
    执行第二策略，所述第二策略为基于非AI得到的策略，所述第二策略与所述第一策略的类型相同；
    执行第三策略，所述第三策略为基于第二AI模型确定的策略，所述第三策略与所述第一策略的类型相同；或者，
    执行第四策略，所述第四策略为执行所述第一策略之前执行的策略，所述第四策略与所述第一策略的类型相同。
  15. 如权利要求14所述的方法，其特征在于，在所述允许采用的修正方式为执行所述第三策略时，所述第一信息包括以下至少一项：
    所述第三策略的标识、所述第三策略、所述第二AI模型的标识、所述第二AI模型的参数或者所述第三网元的性能信息；其中，所述第三网元的性能信息用于确定所述第三策略。
  16. 如权利要求15所述的方法，其特征在于，所述第三网元的性能信息用于确定所述第三策略，包括：
    所述第三网元的性能信息用于修正所述第一AI模型，修正后的第一AI模型用于确定所述第三策略。
  17. 如权利要求15或16所述的方法，其特征在于，还包括：
    接收来自所述第三网元的性能信息。
  18. 如权利要求9-17任一项所述的方法，其特征在于，所述第一网元为第一接入网设备，所述第二网元为核心网网元、或操作、管理与维护OAM网元，所述第三网元为第二接入网设备；或者，
    所述第一网元为第一终端设备，所述第二网元为接入网设备，所述第三网元为第二终端设备；
    所述第一网元为终端设备，所述第二网元为第一接入网设备，所述第三网元为第二接入网设备。
  19. 一种通信方法，其特征在于，应用于第三网元，包括：
    接收来自第二网元的第二信息，所述第二信息用于指示第一策略被执行时出现异常或第一AI模型输出的第一类型的策略被执行时出现异常，所述第一策略属于所述第一类型；其中，所述第一策略基于第一AI模型得到；
    向第二网元发送所述第三网元的性能信息，和/或，确定所述第三网元执行的策略是否出现异常。
  20. 一种通信方法，其特征在于，应用于第一网元，包括：
    获取待输入的第一测量量及获取第一AI模型；
    将所述第一测量量输入至所述第一AI模型中，得到所述第一AI模型输出的第一策略；
    执行所述第一策略；
    在确定执行所述第一策略出现异常时，向第三网元发送指示信息，所述指示信息用于指示执行所述第一策略出现异常；
    接收来自第三网元的性能信息；
    基于所述第三网元的性能信息，对所述第一策略进行修正或对所述第一AI模型进行修正。
  21. 一种通信方法，其特征在于，应用于第三网元，包括：
    接收来自第一网元的指示信息，所述指示信息用于指示所述第一网元执行第一策略出现异常，所述第一策略为第一网元将待输入的第一测量量输入至第一AI模型后所述第一AI模型输出的策略；
    向所述第一网元发送所述第三网元的性能信息，和/或，确定所述第三网元执行的策略是否出现异常。
  22. 一种通信装置，其特征在于，包括：实现如权利要求1-21任一项所述的方法的功能模块。
  23. 一种通信装置，其特征在于，包括处理器，所述处理器与存储器耦合；
    所述存储器，用于存储计算机程序或指令；
    所述处理器，用于执行所述存储器中的部分或者全部计算机程序或指令，当所述部分或者全部计算机程序或指令被执行时，用于实现如权利要求1-21任一项所述的方法。
  24. 一种通信装置，其特征在于，包括处理器和存储器；
    所述存储器，用于存储计算机程序或指令；
    所述处理器，用于执行所述存储器中的部分或者全部计算机程序或指令，当所述部分或者全部计算机程序或指令被执行时，用于实现如权利要求1-21任一项所述的方法。
  25. 一种芯片系统，其特征在于，所述芯片系统包括：处理电路；所述处理电路与存储介质耦合；
    所述处理电路，用于执行所述存储介质中的部分或者全部计算机程序或指令，当所述部分或者全部计算机程序或指令被执行时，用于实现如权利要求1-21任一项所述的方法。
  26. 一种计算机可读存储介质，其特征在于，用于存储计算机程序，所述计算机程序包括用于实现权利要求1-21任一项所述的方法的指令。
  27. 一种计算机程序产品，其特征在于，所述计算机程序产品包括：计算机程序代码，当所述计算机程序代码在计算机上运行时，使得计算机执行如权利要求1-21任一项所述的方法。
PCT/CN2023/116672 2022-09-27 2023-09-04 一种通信方法及装置 WO2024066930A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211186062.7A CN117793766A (zh) 2022-09-27 2022-09-27 一种通信方法及装置
CN202211186062.7 2022-09-27

Publications (1)

Publication Number Publication Date
WO2024066930A1 true WO2024066930A1 (zh) 2024-04-04

Family

ID=90378705

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/116672 WO2024066930A1 (zh) 2022-09-27 2023-09-04 一种通信方法及装置

Country Status (2)

Country Link
CN (1) CN117793766A (zh)
WO (1) WO2024066930A1 (zh)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110380888A (zh) * 2019-05-29 2019-10-25 华为技术有限公司 一种网络异常检测方法和装置
CN111461231A (zh) * 2020-04-02 2020-07-28 腾讯云计算（北京）有限责任公司 一种短信息的发送控制方法、装置及存储介质
CN114071484A (zh) * 2020-07-30 2022-02-18 华为技术有限公司 基于人工智能的通信方法和通信装置
KR20220055378A (ko) * 2020-10-26 2022-05-03 주식회사 인피닉스 딥러닝 기반으로 이상 행동을 자동으로 감지하는 온디바이스 ai 장치 및 이의 동작 방법

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110380888A (zh) * 2019-05-29 2019-10-25 华为技术有限公司 一种网络异常检测方法和装置
CN111461231A (zh) * 2020-04-02 2020-07-28 腾讯云计算（北京）有限责任公司 一种短信息的发送控制方法、装置及存储介质
CN114071484A (zh) * 2020-07-30 2022-02-18 华为技术有限公司 基于人工智能的通信方法和通信装置
KR20220055378A (ko) * 2020-10-26 2022-05-03 주식회사 인피닉스 딥러닝 기반으로 이상 행동을 자동으로 감지하는 온디바이스 ai 장치 및 이의 동작 방법

Also Published As

Publication number Publication date
CN117793766A (zh) 2024-03-29

Similar Documents

Publication Publication Date Title
US11451452B2 (en) Model update method and apparatus, and system
US11129086B2 (en) System message notification and sending method and apparatus
KR101561483B1 (ko) 네트워크의 액세스 노드들에 의한 결함 검출 응답들을 조정하는 방법
US10966108B2 (en) Optimizing radio cell quality for capacity and quality of service using machine learning techniques
US10827430B2 (en) Signaling for controlling power usage in radio access networks
US20220338106A1 (en) Slice control method and apparatus
US20230388817A1 (en) Activating intelligent wireless communciation device reporting in a wireless network
US11057863B2 (en) Terminal device and network device for exchanging information on selected beam
US20160105894A1 (en) eMBMS Management Method, Multimedia Broadcast Multicast Service Coordination Entity, and Base Station
WO2022058013A1 (en) Methods and apparatuses for handover procedures
US20230209450A1 (en) Mobility management method and apparatus
US20240015534A1 (en) Model processing method, communication apparatus, and system
CN114765789A (zh) 无线通信网络中数据处理方法和装置
US20230350724A1 (en) Node determination method for distributed task and communication device
US20230079241A1 (en) Intelligent cellular channel management
WO2024066930A1 (zh) 一种通信方法及装置
US11595286B2 (en) Systems and methods for modifying parameters of a wireless network based on granular energy efficiency metrics
US20230319597A1 (en) Network node and a method performed in a wireless communication network for handling configuration of radio network nodes using reinforcement learning
EP3200540B1 (en) Interference adjusting and processing method and device
US20230413116A1 (en) Load balancing method and apparatus, and readable storage medium
WO2023201746A1 (en) Method, device, and system for resource status report in wireless networks
WO2023169101A1 (zh) 一种通信方法和通信装置
CN114205059B (zh) 信息交互方法、装置及存储介质
US20220338105A1 (en) Method and system for controlling user equipment policy in wireless network
US20240211768A1 (en) Signaling of training policies

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: 23870170

Country of ref document: EP

Kind code of ref document: A1