WO2024093739A1 - 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
WO2024093739A1
WO2024093739A1 PCT/CN2023/126265 CN2023126265W WO2024093739A1 WO 2024093739 A1 WO2024093739 A1 WO 2024093739A1 CN 2023126265 W CN2023126265 W CN 2023126265W WO 2024093739 A1 WO2024093739 A1 WO 2024093739A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
condition
information
area
range
Prior art date
Application number
PCT/CN2023/126265
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 WO2024093739A1 publication Critical patent/WO2024093739A1/zh

Links

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0803—Configuration setting
    • H04L41/0823—Configuration setting characterised by the purposes of a change of settings, e.g. optimising configuration for enhancing reliability
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/16—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks using machine learning or artificial intelligence
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W16/00—Network planning, e.g. coverage or traffic planning tools; Network deployment, e.g. resource partitioning or cells structures
    • H04W16/22—Traffic simulation tools or models
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/02—Arrangements for optimising operational condition

Definitions

  • the present application relates to the field of communication technology, and in particular to a communication method and device.
  • AI Artificial intelligence
  • 3GPP 3rd Generation Partnership Project
  • the present application provides a communication method and apparatus for improving the collaboration efficiency between devices.
  • an embodiment of the present application provides a communication method.
  • the method includes: a first device may receive first information from a second device.
  • the first information may be used to indicate a correspondence between at least one condition and at least one AI model.
  • the first device may use a first AI model corresponding to the first condition, where the first condition is any one of the at least one condition.
  • the first device is a terminal device and the second device is a network device; or the first device is a network device and the second device is a terminal device.
  • the first device can select an AI model to use based on the correspondence between at least one condition indicated by the second device and at least one AI model, thereby improving the efficiency of the first device in determining and using the AI model, enabling the first device to use the AI model reasonably and effectively, and further improving the collaboration efficiency between devices.
  • the first device can select an AI model according to the corresponding relationship.
  • the second device does not need to select an AI model for the first device according to the measurement information of the first device, thereby reducing the overhead of the second device in obtaining the measurement information of the first device.
  • the first condition includes at least one of the following:
  • the signal quality of the first device satisfies a first signal quality condition
  • the application scenario of the first device meets the first application scenario condition
  • the resources of the first device meet the first resource condition
  • the resources of the second device served by the first device meet the second resource condition
  • the area where the first device is located meets the first area condition
  • the area where the second device served by the first device is located meets the second area condition
  • the capability of the first device satisfies the first capability condition
  • the performance of the second AI model currently used by the first device meets the first performance condition.
  • the first signal quality condition includes: the signal quality of the first device belongs to a first signal quality range;
  • the first application scenario condition includes: the application scenario of the first device belongs to the first application scenario range;
  • the first resource condition includes: the resources of the first device belong to the first resource range;
  • the second resource condition includes: the resources of the second device belong to the second resource range;
  • the first area condition includes: the area where the first device is located belongs to the first area range;
  • the second area condition includes: the area where the second device is located belongs to the second area range;
  • the first capability condition includes: the capability of the first device belongs to the first capability range; and/or
  • the first performance condition includes: the performance of the second AI model belongs to the first performance range.
  • the design provides a variety of possible conditions, enabling the first device to use the AI model reasonably and effectively, and is easy to implement.
  • the first information is further used to indicate a correspondence between at least one exit condition and at least one AI model.
  • the first device may stop using the first AI model, where the second condition is a condition in the at least one exit condition corresponding to the first AI model.
  • the second device can indicate the correspondence between at least one exit condition of the first device and at least one AI model.
  • the first device can reasonably exit the AI model, thereby improving the efficiency of the first device in determining and using the AI model, so that the first device can reasonably and effectively use the AI model, thereby improving the collaboration efficiency between devices.
  • the second condition includes at least one of the following:
  • the signal quality of the first device satisfies the second signal quality condition
  • the application scenario of the first device meets the second application scenario condition
  • the resources of the first device meet the third resource condition
  • the resources of the second device served by the first device meet the fourth resource condition
  • the area where the first device is located meets the third area condition
  • the area where the second device served by the first device is located meets the fourth area condition
  • the capability of the first device satisfies the second capability condition
  • the performance of the first AI model meets the second performance condition.
  • the second signal quality condition includes: the signal quality of the first device belongs to the second signal quality range;
  • the second application scenario condition includes: the application scenario of the first device belongs to the scope of the second application scenario;
  • the third resource condition includes: the resources of the first device belong to the third resource range;
  • the fourth resource condition includes: the resources of the second device belong to the fourth resource range;
  • the third area condition includes: the area where the first device is located belongs to the third area range;
  • the fourth area condition includes: the area where the second device is located belongs to the fourth area range;
  • the second capability condition includes: the capability of the first device belongs to the second capability range; and/or
  • the second performance condition includes: the performance of the first AI model falls within the second performance range.
  • the design provides a variety of possible exit conditions, allowing the first device to reasonably and effectively exit the AI model, and is easy to implement.
  • the first device may send a second message to the second device, and the second message is used to indicate that the AI model used by the first device has changed.
  • the first device can notify the second device of the information that the AI model has changed.
  • the second device can reasonably set the corresponding relationship between the condition and the AI model based on the second information fed back by the first device, so that the first device can reasonably select and use the AI model, thereby improving system performance.
  • the second information includes at least one of the following:
  • the network performance and/or AI model performance of the first device within a first period of time before the AI model used by the first device changes and/or within a second period of time after the change;
  • the first device can flexibly notify the second device of changes in the AI model.
  • the first device when the second device switches from the first device to the third device, the first device may send the first information to the third device.
  • the second device may be a terminal device
  • the first device may be a source network device of the terminal device
  • the third device may be a destination network device of the terminal device.
  • the source network device when the terminal device switches from the source network device to the destination network device, the source network device may send the first information from the terminal device to the destination network device.
  • the destination network device may also use the AI model according to the correspondence between the conditions provided by the terminal device and the AI model.
  • the first device before receiving the first information from the second device, the first device may send third information to the second device, the third information is used to indicate the capabilities of the first device and/or the application scenario of the first device, and the third information is used to determine the first information.
  • the second device can reasonably determine the correspondence between the AI model and each condition.
  • an embodiment of the present application provides a communication method.
  • the method includes: a second device may obtain first information, where the first information is used to indicate a corresponding relationship between at least one condition and at least one AI model. Then, the second device may send the first information to the first device.
  • the first device is a terminal device and the second device is a network device; or the first device is a network device and the second device is a terminal device.
  • the second device can send first information indicating the correspondence between at least one condition and at least one AI model to the first device.
  • the first device can select the AI model to use according to the correspondence, thereby improving the efficiency of the first device in determining and using the AI model, so that the first device can use the AI model reasonably and effectively, thereby improving the efficiency of collaboration between devices.
  • any one of the at least one condition includes at least one of the following:
  • the signal quality of the first device satisfies a first signal quality condition
  • the application scenario of the first device meets the first application scenario condition
  • the resources of the first device meet the first resource condition
  • the resources of the second device served by the first device meet the second resource condition
  • the area where the first device is located meets the first area condition
  • the area where the second device served by the first device is located meets the second area condition
  • the capability of the first device satisfies the first capability condition
  • the performance of the second AI model currently used by the first device meets the first performance condition.
  • the first signal quality condition includes: the signal quality of the first device belongs to a first signal quality range;
  • the first application scenario condition includes: the application scenario of the first device belongs to the first application scenario range;
  • the first resource condition includes: the resources of the first device belong to the first resource range;
  • the second resource condition includes: the resources of the second device belong to the second resource range;
  • the first area condition includes: the area where the first device is located belongs to the first area range;
  • the second area condition includes: the area where the second device is located belongs to the second area range;
  • the first capability condition includes: the capability of the first device belongs to the first capability range; and/or
  • the first performance condition includes: the performance of the second AI model belongs to the first performance range.
  • the design provides a variety of possible conditions, enabling the first device to use the AI model reasonably and effectively, and is easy to implement.
  • the first information is also used to indicate the corresponding relationship between at least one exit condition and at least one AI model.
  • the first device can reasonably exit the AI model, thereby improving the efficiency of the first device in determining and using the AI model, so that the first device can reasonably and effectively use the AI model, thereby improving the efficiency of collaboration between devices.
  • the second condition is one of the at least one exit condition, and the second condition includes at least one of the following:
  • the signal quality of the first device satisfies the second signal quality condition
  • the application scenario of the first device meets the second application scenario condition
  • the resources of the first device meet the third resource condition
  • the resources of the second device served by the first device meet the fourth resource condition
  • the area where the first device is located meets the third area condition
  • the area where the second device served by the first device is located meets the fourth area condition
  • the capability of the first device satisfies the second capability condition
  • the performance of the first AI model meets the second performance condition.
  • the second signal quality condition includes: the signal quality of the first device belongs to the second signal quality range;
  • the second application scenario condition includes: the application scenario of the first device belongs to the scope of the second application scenario;
  • the third resource condition includes: the resources of the first device belong to the third resource range;
  • the fourth resource condition includes: the resources of the second device belong to the fourth resource range;
  • the third area condition includes: the area where the first device is located belongs to the third area range;
  • the fourth area condition includes: the area where the second device is located belongs to the fourth area range;
  • the second capability condition includes: the capability of the first device belongs to the second capability range; and/or
  • the second performance condition includes: the performance of the first AI model falls within the second performance range.
  • the design provides a variety of possible exit conditions, allowing the first device to reasonably and effectively exit the AI model, and is easy to implement.
  • the second device may receive second information from the first device, and the second information is used to indicate that the AI model used by the first device has changed.
  • the first device can notify the second device of information about changes in the AI model.
  • the second device can reasonably set the corresponding relationship between the condition and the AI model based on the second information fed back by the first device, so that the first device can reasonably select and use the AI model, thereby improving system performance.
  • the second information may include at least one of the following:
  • the network performance and/or AI model performance of the first device within a first period of time before the AI model used by the first device changes and/or within a second period of time after the change;
  • the first device can flexibly notify the second device of changes in the AI model.
  • the second device when the first device switches from the second device to the fourth device, the second device may send the first information to the fourth device.
  • the first device may be a terminal device
  • the second device may be a source network device of the terminal device
  • the fourth device may be a destination network device of the terminal device.
  • the source network device may send the first information to the destination network device.
  • the destination network device can also obtain the correspondence between the conditions provided by the source network device and the AI model, so that it can cooperate with the terminal device that uses the correspondence for operation.
  • the second device may receive third information from the first device, the third information is used to indicate the capability of the first device and/or the application scenario of the first device, and the third information is used to determine the first information.
  • the second device can reasonably determine the correspondence between the AI model and each condition based on the third information.
  • an embodiment of the present application provides a communication method.
  • the method includes: a first device may receive fourth information from a second device, and the fourth information includes configuration information of at least one AI model.
  • the first device may use a third AI model after receiving fifth information from the second device.
  • the fifth information includes indication information of the third AI model, and the third AI model is one of the at least one AI model.
  • the second device after the second device sends the configuration information of at least one AI model to the first device, it can instruct the first device to use the third AI model through the indication information of the third AI model, thereby reducing the overhead and time required to instruct the first device to use the third AI model, thereby improving the collaboration efficiency between devices.
  • the indication information of the third AI model is the index of the third AI model. This design indicates the AI model by the index of the AI model, so that even if a third party obtains the index, it does not know the AI model corresponding to the index, thereby improving security.
  • an embodiment of the present application provides a communication method.
  • the method includes: a second device sends fourth information to a first device, the fourth information including configuration information of at least one AI model. Then, the second device may send fifth information to the first device, the fifth information including indication information of a third AI model, the third AI model being one of the at least one AI model, and the fifth information is used to instruct the first device to use the third AI model.
  • the second device after the second device sends the configuration information of at least one AI model to the first device, it can instruct the first device to use the third AI model through the indication information of the third AI model, thereby reducing the overhead and time required to instruct the first device to use the third AI model, thereby improving the collaboration efficiency between devices.
  • the indication information of the third AI model is the index of the third AI model. This design indicates the AI model by the index of the AI model, so that even if a third party obtains the index, it does not know the AI model corresponding to the index, thereby improving security.
  • an embodiment of the present application provides a communication device, comprising a unit for executing each step in any of the above aspects.
  • an embodiment of the present application provides a communication device, comprising at least one processing element and at least one storage element, wherein the at least one storage element is used to store programs and data, and the at least one processing element is used to read and execute the programs and data stored in the storage element, so that the method provided in any one of the above aspects of the present application is implemented.
  • an embodiment of the present application provides a communication system, comprising: a first device for executing the method provided in the first aspect, and a second device for executing the method provided in the second aspect.
  • an embodiment of the present application provides a communication system, comprising: a first device for executing the method provided in the third aspect, and a second device for executing the method provided in the fourth aspect.
  • an embodiment of the present application further provides a computer program, which, when executed on a computer, enables the computer to execute the method provided in any of the above aspects.
  • an embodiment of the present application further provides a computer-readable storage medium, in which a computer program is stored.
  • the computer program is executed by a computer, the computer executes the method provided in any of the above aspects.
  • an embodiment of the present application further provides a chip, which is used to read a computer program stored in a memory and execute the method provided in any of the above aspects.
  • an embodiment of the present application further provides a chip system, which includes a processor for supporting a computer device to implement the method provided in any of the above aspects.
  • the chip system also includes a memory, which is used to store the necessary programs and data of the computer device.
  • the chip system can be composed of a chip, or it can include a chip and other discrete devices.
  • FIG1 is a schematic diagram of the architecture of a communication system used in an embodiment of the present application.
  • FIG2 is a schematic diagram of a flow chart of a first communication method provided in an embodiment of the present application.
  • FIG3 is a schematic diagram of a flow chart of a second communication method provided in an embodiment of the present application.
  • FIG4 is a schematic diagram of a flow chart of a third communication method provided in an embodiment of the present application.
  • FIG5 is a schematic diagram of a flow chart of a fourth communication method provided in an embodiment of the present application.
  • FIG6 is a schematic diagram of the structure of a communication device provided in an embodiment of the present application.
  • FIG. 7 is a schematic diagram of the structure of another communication device provided in an embodiment of the present application.
  • NR new radio
  • GSM global system of mobile communications
  • LTE long term evolution
  • the interaction between the terminal device, the access network device and the core network device is described as an example. It should be noted that the method provided in the embodiment of the present application can be applied not only to the interaction between the terminal device and the network side, but also to the interaction between any two devices. The embodiment of the present application is not limited to this.
  • the terminal device can provide voice and/or data connectivity services for the user.
  • the terminal device can be a wireless terminal device or a wireless terminal device.
  • a wireless terminal device is a device with wireless transceiver function, for example, a device with wireless connection function or a device connected to a wireless modulator.
  • the terminal device can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; it can also be deployed on the water surface (such as a ship, etc.); it can also be deployed in the air (for example, on an airplane, a balloon, and a satellite, etc.).
  • the terminal device can also be called user equipment (UE), subscriber unit (SU), subscriber station (SS), mobile station (MB), mobile station (Mobile), remote station (RS), access point (AP), remote terminal (RT), access terminal (AT), user terminal (UT), user agent (UA), terminal device (UD).
  • UE user equipment
  • SU subscriber unit
  • SS subscriber station
  • MB mobile station
  • MB mobile station
  • RS remote station
  • AP access point
  • RT access terminal
  • AT user terminal
  • UT user agent
  • terminal device UD
  • the terminal device can be: a mobile phone, a tablet computer, a laptop computer, a PDA, a mobile internet device (MID), a mobile cellular phone, a cordless phone, a personal digital assistant (PDA), a customer-premises equipment (CPE), a smart point of sale (POS), a personal communication service (PCS) phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a wearable device (such as a smart watch, a smart phone), a rings, pedometers, etc.), vehicle-mounted equipment (for example, cars, bicycles, electric vehicles, airplanes, ships, trains, high-speed railways, etc.), virtual reality (VR) equipment, augmented reality (AR) equipment, wireless terminals in industrial control, smart home equipment (for example, refrigerators, televisions, air conditioners, electric meters, etc.), intelligent robots, workshop equipment, wireless terminals in unmanned driving, wireless terminals in remote surgery, wireless terminals in smart grids, wireless terminals in transportation
  • the network devices involved in the embodiments of the present application may be base stations in a radio access network (RAN) (such as next generation Node B (gNB)), etc.
  • RAN radio access network
  • gNB next generation Node B
  • the base station may be a centralized unit (CU) and a distributed unit (DU) separated architecture.
  • the RAN may be connected to a core network (for example, it may be a long term evolution (LTE) core network or a 5G core network, etc.).
  • LTE long term evolution
  • 5G core network etc.
  • the CU and DU may be understood as a division of the base station from a logical functional perspective.
  • the CU and DU may be physically separated or deployed together. Multiple DUs may share one CU.
  • a DU can also be connected to multiple CUs (not shown in the figure).
  • the CU and DU can be connected through an interface, such as an F1 interface.
  • the CU and DU can be divided according to the protocol layer of the wireless network.
  • the 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
  • the DU is used to perform the functions of the radio link control (RLC) layer, the media access control (MAC) layer, and the physical layer.
  • RRC radio resource control
  • SDAP service data adaptation protocol
  • PDCP packet data convergence protocol
  • RLC radio link control
  • MAC media access control
  • the CU or DU may have some processing functions of the protocol layer in the above division method.
  • some functions of the RLC layer and the functions of the protocol layers above the RLC layer are set in the CU, and the remaining functions of the RLC layer and the functions of the protocol layers below the RLC layer are set in the DU.
  • the functions of CU or DU can also be divided according to service type or other system requirements. For example, by latency, the functions whose processing time needs to meet the latency requirements are set in DU, and the functions that do not need to meet the latency requirements are set in CU.
  • CU can also have one or more functions of the core network.
  • One or more CUs can be set centrally or separately.
  • CU can be set on the network side for centralized management.
  • DU can have multiple radio frequency functions, or the radio frequency 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 centralized unit control plane (CU-CP) node or the centralized unit user plane (CU-UP) node is separated.
  • CU-CP and CU-UP can be implemented by different functional entities and connected through the E1 interface.
  • 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 can also be divided into CU-CP1 and CU-CP2.
  • CU-CP1 is used to implement the radio resource management function
  • CU-CP2 is used to implement the RRC function and the packet data convergence protocol control (PDCP-C) function (that is, the basic function of the control plane signaling at the PDCP layer).
  • PDCP-C packet data convergence protocol control
  • AI application scenarios include but are not limited to at least one of the following: energy saving, load balancing, mobility optimization, channel status information reference signal (CSI-RS) feedback enhancement, beam management enhancement, and positioning accuracy enhancement. These are explained below.
  • CSI-RS channel status information reference signal
  • the network equipment can predict the load of the network equipment based on its own and neighboring cell's load, energy consumption, energy efficiency and other information, as well as the terminal equipment's trajectory, measurement results and other information. Without affecting network coverage and user access, the network equipment can take appropriate energy-saving measures in a timely manner based on the prediction results.
  • the energy-saving measures may include at least one of the following: deactivating the cell, shutting down the carrier, shutting down the channel, shutting down the time slot, reducing the transmission power, etc.
  • the network equipment can predict the load of the network equipment based on its own and neighboring cell's load, energy consumption, energy efficiency and other information, as well as the terminal equipment's trajectory, measurement results and other information. Based on the prediction results, the network equipment can switch some terminal equipment to the neighboring cell, or switch some terminal equipment served by the neighboring cell to the local cell, so that the load of each network equipment in the network is close, avoiding the situation where some network equipment is overloaded, affecting the terminal equipment's business, while the resources of other network equipment are idle.
  • the network device can predict the future trajectory of the terminal device based on the historical trajectory information of the terminal device and the measurement information of the terminal device. Based on the prediction results, the network device can determine in advance whether the terminal device needs to be switched, configure the information required for the switching for the terminal device that needs to be switched in advance, and notify the target cell to prepare access resources for the terminal device, thereby reducing the delay of the terminal device during the switching process and improving the success rate of the terminal device switching.
  • the network device may provide an encoder and a quantization tool for the terminal device according to information such as the capabilities of the terminal device. After the terminal device compresses and quantizes the CSI-RS according to the encoder and the quantization tool, it may send the CSI-RS to the network device. The network device may determine the channel quality of the terminal device according to the CSI-RS.
  • the network device can obtain full beam scanning results from multiple terminal devices and train a sparse scanning matrix based on them. Then, the network device can After sending the sparse scanning matrix to the terminal device a, the sparse scanning result is obtained from the terminal device a, and the optimal CSI-RS beam of the terminal device a is determined according to the sparse scanning result.
  • Positioning enhancement is used to improve positioning accuracy.
  • Positioning enhancement may include at least one of the following: positioning enhancement based on access network equipment, positioning enhancement based on positioning management function network elements, and positioning enhancement based on terminal equipment.
  • the AI model is a specific implementation of the AI function, which can characterize the mapping relationship between the input and output of the model.
  • the AI model can be a neural network, a linear regression model, a decision tree model, a support vector machine (SVM), a Bayesian network, a Q learning model, or other machine learning models.
  • the AI function can include at least one of the following: data collection (collecting training data and/or inference data), data preprocessing, model training (or model learning), model information release (configuring model information), model verification, model inference, or inference result release. Among them, inference can also be called prediction.
  • the AI module has machine learning computing capabilities and is a module used to implement AI models.
  • the AI module can be located in operation administration and maintenance (OAM), in a network device, in a terminal device, or in a separate control layer, such as an AI control layer (AIC).
  • OAM operation administration and maintenance
  • AIC AI control layer
  • the main functions of the AI module may include: performing a series of AI calculations such as AI model building, training approximation, and reinforcement learning based on input data.
  • input data may include network operation data provided by network equipment or monitored by OAM, such as network load, channel quality, and other data.
  • the trained model provided by the AI module has the function of predicting changes in the RAN network, and can be used for load prediction and/or trajectory prediction of terminal devices, etc.
  • the AI module can also infer energy-saving strategies and/or mobility optimization strategies based on the prediction results of the trained model on RAN network performance.
  • the AI module When the AI module is located in OAM, the AI module can communicate with network equipment through the northbound interface; when the AI module is located in gNB or CU, the AI module can communicate with other devices through F1, Xn, Uu and other interfaces; when the AI module acts as an independent network entity, the AI module can establish a communication link from the AI module to the OAM and/or RAN side, which can be a wired link or a wireless link.
  • the CU-CP can be used to receive AI models and perform functions such as AI reasoning and policy generation.
  • the CU-CP includes CU-CP1 and CU-CP2
  • CU-CP1 can be used to receive AI models and perform functions such as AI reasoning, generate interactive signaling, and send the interactive signaling through CU-CP2.
  • the terminal device can support multiple AI models.
  • the network device can determine the AI model that the terminal device should currently use based on the measurement results of the terminal device and other information, and instruct the terminal device to use the AI model. Since the AI models used by the terminal device may be different in different situations, the network device needs to obtain the information used to determine the AI model from the terminal device in a timely manner, which has a large overhead and a long delay, and the collaboration efficiency between devices is low.
  • an embodiment of the present application provides a communication method, which can be applied to the communication system shown in Figure 1. Referring to the flowchart shown in Figure 2, the process of the method is specifically described below.
  • S201 The first device sends third information to the second device.
  • the second device receives the third information from the first device.
  • the first device is a terminal device, and the second device is a network device; or, the first device is a network device, and the second device is a terminal device.
  • the third information may be used to indicate the capability of the first device and/or the application scenario of the first device.
  • the third information may include the capability of the first device and/or the application scenario of the first device, and may also include information corresponding to the capability of the first device and/or the application scenario of the first device.
  • the capabilities of the first device include but are not limited to at least one of the following:
  • Cache capacity of the terminal device For example, the cache size of the terminal device is 50 megabytes (MB), which means that the terminal device can store an AI model with a capacity of 50MB.
  • MB megabytes
  • AI computing capability of the terminal device For example, the AI computing capability of the terminal device is 1 Tera (T, ie 10 12 ) floating-point operations per second (FLPS).
  • T Tera
  • FLPS floating-point operations per second
  • Power of the terminal device For example, the available power of the terminal device is 2000 mAh.
  • AI model information of the terminal device For example, the software library used to implement AI functions in the terminal device (for example, Tensflow). Another example is the AI model format supported by the terminal device (for example, Open Neural Network Exchange exchange, ONNX)).
  • the indication information may be information that corresponds to the AI model.
  • the indication information may be a model identifier (ID).
  • the model ID may be a bit stream consisting of multiple bits. The bit stream may not only indicate the AI model, but also indicate one or more of the manufacturer information, version information, compilation platform information, etc. of the AI model.
  • the indication information is an index to the AI model, and the index corresponds to the ID of the AI model.
  • the performance of the AI model expected by the terminal device For example, the target value of throughput that the terminal device expects to achieve after using the AI model. For another example, the prediction accuracy of the AI model expected by the terminal device is higher than 95%.
  • the capabilities of the first device include but are not limited to at least one of the following:
  • Cache capacity of the network device For example, if the cache size of a network device is 50MB, it means that the network device can store an AI model with a capacity of 50MB.
  • AI computing power of network equipment For example, the AI computing power of network equipment is 1T FLPS.
  • AI model information of network devices For example, the software library used to implement AI functions in network devices (for example, Tensflow). Another example is the AI model format supported by the network device (for example, ONNX).
  • the indication information may be information that corresponds to the AI model.
  • the indication information may be a model ID.
  • the model ID may be a bit stream consisting of multiple bits.
  • the bit stream may not only indicate the AI model, but also indicate one or more of the manufacturer information, version information, compilation platform information, etc. of the AI model.
  • the indication information is an index to the AI model, and the index corresponds to the ID of the AI model.
  • the performance of the AI model expected by the network device For example, the network device expects the target value of throughput to be achieved after using the AI model. For another example, the network device expects the prediction accuracy of the AI model to be higher than 95%.
  • Application scenarios of the first device include but are not limited to at least one of the following: energy saving, load balancing, mobility optimization, CSI-RS feedback enhancement, beam management enhancement, and positioning enhancement.
  • the third information may be carried in an existing message or in a new message.
  • the first device may actively send the third information to the second device, or may send the third information to the second device based on the request of the second device.
  • the first device may send the third information according to a set period, or may send the third information to the second device based on an event trigger.
  • the set period may be predefined or obtained by the first device from the second device.
  • the triggering event may be a change in the capability and/or application scenario of the first device.
  • S201 is an optional step.
  • S202 The second device sends first information to the first device.
  • the first device receives the first information from the second device.
  • the first information can be used to indicate the correspondence between at least one condition and at least one AI model.
  • At least one condition and at least one AI model may be in a one-to-one relationship.
  • at least one condition includes condition 1 and condition 2
  • at least one AI model includes AI model 1 and AI model 2, where condition 1 corresponds to AI model 1 and condition 2 corresponds to AI model 2.
  • At least one condition and at least one AI model may be in a many-to-one relationship.
  • at least one condition includes conditions 1 to 4
  • at least one AI model includes AI model 1 and AI model 2
  • conditions 1 and 4 correspond to AI model 1
  • conditions 2 and 3 correspond to AI model 2.
  • the first condition is any one of the at least one condition, and the first condition corresponds to the first AI model of the at least one AI model.
  • the first condition may include one or more conditions shown in Table 1.
  • the first signal quality condition may include: the signal quality of the first device belongs to the first signal quality range.
  • the signal quality of the first signal includes at least one of the following: the signal strength from the second device detected by the first device, the throughput of the first device, the transmission delay of the signal from the second device detected by the first device, the packet error rate of the signal from the second device detected by the first device, and the call drop rate of the first device.
  • the signal strength of the first device continues to exceed the first signal quality threshold (for example, 100 decibel relative to one milliwatt, dBm) for n consecutive seconds, and the maximum throughput of the first device is less than the first throughput threshold
  • the first device can use AI model 1, where n is a positive integer.
  • the first device can switch from AI model 1 to AI model 2. For another example, if the packet error rate of the signal from the second device detected by the first device and/or the call drop rate of the first device is greater than 20%, the first device can switch from AI model 1 to AI model 2. For another example, if the signal strength detected by the first device from the second device is greater than -100dBm, and the call drop rate of the first device is greater than 20%, the first device may use AI model 1.
  • the first application scenario condition includes: the application scenario of the first device belongs to the scope of the first application scenario.
  • the first application scenario condition can be used to indicate the correspondence between the application scenario or application sub-scenario and the AI model.
  • CSI-RS feedback enhancement can correspond to AI model a1 and AI model a2
  • beam management enhancement can correspond to AI model a3 and AI model a4.
  • the first device and the second device communicate using beam pair 1
  • the first device can use AI model a3
  • the first device and the second device communicate using beam pair 2 the first device can use AI model a4.
  • the first resource condition includes: the resources of the first device belong to the first resource range.
  • the first condition may include the first resource condition.
  • the uplink and downlink time-frequency resources used by the terminal device may correspond to the AI model. For example, when the center frequency of the bandwidth part (bandwidth part, BWP) used by the terminal device is the first frequency and/or the bandwidth used by the terminal device is the first bandwidth, the terminal device may use AI model b1.
  • BWP bandwidth part
  • the second resource condition includes: the resources of the second device belong to the second resource range.
  • the first condition may include the second resource condition.
  • the second resource range and the first resource range may be the same or different. For example, when the center frequency of the BWP used by the terminal device is the first frequency and/or the bandwidth used by the terminal device is the first bandwidth, the network device may use the AI model b1.
  • the first area condition includes: the area where the first device is located belongs to the first area range.
  • the first condition may include the first area condition.
  • the area where the terminal device is located may be the geographical location area where the terminal device is located, and the geographical location area may be determined by information such as administrative area, longitude and latitude, and/or altitude. For example, if the terminal device is located in city A, the terminal device may use the AI model c1 corresponding to city A.
  • the area where the terminal device is located can be determined by the public land mobile network (PLMN) accessed and/or the cell global identifier (CGI) of the accessed cell. For example, if the terminal device accesses PLMN1, the terminal device can use the AI model c2 corresponding to PLMN1. For another example, if the identifier of the cell accessed by the terminal device is CGI1, the terminal device can use the AI model c3 corresponding to CGI1.
  • PLMN public land mobile network
  • CGI cell global identifier
  • the second area condition includes: the area where the second device is located belongs to the second area range.
  • the first condition may include the second area condition.
  • the second area range and the first area range may be the same or different.
  • the region where the terminal device is located may be the geographical location region where the terminal device is located, and the geographical location region may be determined by information such as administrative region, longitude and latitude, and/or altitude. For example, if the terminal device is located in City A, the network device may use the AI model c1 corresponding to City A. The region where the terminal device is located may be sent by the terminal device to the network device.
  • the area where the terminal device is located can be determined by the public land mobile network (PLMN) accessed and/or the cell global identifier (CGI) of the accessed cell. For example, if the terminal device accesses PLMN1, the network device can use the AI model c2 corresponding to PLMN1. For another example, if the terminal device accesses a cell If the zone identifier is CGI1, the network device can use the AI model c3 corresponding to CGI1.
  • PLMN public land mobile network
  • CGI cell global identifier
  • the first capability condition includes: the capability of the first device belongs to the first capability range.
  • the capability of the first device may include the remaining power of the first device. When the remaining power of the first device is less than the first power threshold, the first device may use the AI model d1.
  • the capability of the first device includes the computing capability of the first device. When the computing rate of the first device is less than the first rate threshold, the first device may use the AI model d2.
  • the first performance condition includes: the performance of the second AI model belongs to the first performance range.
  • the second AI model may be the AI model currently used by the first device.
  • the performance of the second AI model may include the prediction accuracy of the second AI model and/or the network performance after network optimization using the prediction results of the second AI model. For example, when the prediction accuracy of the second AI model (for example, in beam management enhancement, the probability that the predicted optimal beam is the actual optimal beam) is less than the first accuracy threshold (for example, within 1 minute, the prediction accuracy of the second AI model is less than 95%), the first device may switch the currently used AI model to the first AI model.
  • the first device may switch the currently used AI model to the first AI model.
  • the first device may switch the currently used AI model to the first AI model.
  • the expected performance target is to increase the throughput of the first device to X. If the throughput of the first device is less than X after network optimization using the prediction result of the second AI model, the first device may switch the currently used AI model to the first AI model.
  • the first device When the first device satisfies the first condition, the first device can use the first AI model corresponding to the first condition. Therefore, the first condition can be called the entry condition of the first AI model. If the first device satisfies the first condition, the first device changes from using the second AI model to using the first AI model, and the first condition can also be called the change condition.
  • the second device can indicate the correspondence between at least one condition and at least one AI model of the first device, so that the first device can select the AI model to use based on the correspondence, thereby improving the efficiency of the first device in determining and using the AI model, enabling the first device to use the AI model reasonably and effectively, and thereby improving the collaboration efficiency between devices.
  • the first device can select an AI model according to the corresponding relationship.
  • the second device does not need to select an AI model for the first device according to the measurement information of the first device, thereby reducing the overhead of the second device in obtaining the measurement information of the first device.
  • the second device may determine the correspondence between at least one condition and at least one AI model based on at least one of the following: capability of the first device, application scenario of the first device, signal quality of the first device, and requirements of the second device.
  • the requirements of the second device may include the AI model existing in the second device and the target application scenario.
  • the second device may determine at least one AI model based on one or more of the capabilities of the first device, the application scenario of the first device, and the needs of the second device. For example, when the cache capacity of the first device is 50MB, the second device may determine that the size of any AI model in at least one AI model is less than 50MB. For another example, when the terminal device can support AI models in the ONNX format, the second device may determine that the format of any AI model in at least one AI model is ONNX. For another example, when the application scenario or target application scenario of the first device includes beam management enhancement, the second device may determine that at least one AI includes an AI model for beam management enhancement. For another example, at least one AI model includes an AI model existing in the second device.
  • the second device may determine the condition corresponding to at least one AI model according to one or more of the performance, overhead and signal quality of the at least one AI model, that is, determine the correspondence between at least one condition and at least one AI model. For example, when the computing power required for AI model 1 is 0.5T FLPS and the computing power required for AI model 2 is 0.3T FLPS, the second device may determine that the condition corresponding to AI model 1 includes that the computing power of the first device is greater than or equal to 0.5T FLPS (for example, the computing power of the first device is 0.4T FLPS), and the condition corresponding to AI model 1 includes that the computing power of the first device is greater than or equal to 0.3T FLPS (for example, the computing power of the first device is 0.4T FLPS).
  • the second device may determine that the condition corresponding to AI model 1 includes that the signal strength detected by the first device from the second device is greater than c dBm, where c is greater than or equal to a and less than or equal to b, and a, b and c are positive integers.
  • the first information may also include configuration information of at least one AI model.
  • the configuration information of the first AI model may include the model ID and/or model content information of the first AI model.
  • the configuration information of the first AI model may include the specific content of the first AI model and the model ID of the first model.
  • the specific content of the model ID can refer to the description of the model ID in S201; the specific content of the first AI model may include the first AI model and its sub-models.
  • the configuration information of the first AI model also includes an index of the first AI model.
  • the configuration information of the first AI model may include a model ID and an index of the first AI model.
  • the index of the first AI model is used to indicate the first AI model.
  • the index of the first AI model may correspond to the model ID of the first AI model.
  • the index of the first AI model may include at least one of the following: an index in the model list and/or an index in the model sub-list.
  • the index in the model list can be used to indicate the first AI model, and the index in the model sub-list can be used to indicate a sub-version of the first AI model.
  • index 1 can be used to indicate an AI model with a model ID of XXX
  • index 1-1 can be used to indicate an AI model with a model ID of XXXX1
  • index 1-2 can be used to indicate an AI model with a model ID of XXXX2
  • the AI model with a model ID of XXXX1 and the AI model with a model ID of XXXX1 are sub-versions of the AI model with a model ID of XXX.
  • indexes in the model lists and model sublists corresponding to different application scenarios can be numbered separately or consecutively.
  • the indexes in the model list corresponding to beam management enhancement are 1 and 2
  • the indexes in the model list corresponding to CSI-RS feedback enhancement are 3 and 4.
  • the first information further includes the shortest execution time of each AI model in the at least one AI model.
  • the shortest execution time of the first AI model may be 5 minutes.
  • the first information may also be used to indicate a corresponding relationship between at least one exit condition and at least one AI model.
  • the at least one exit condition and at least one AI model may be a one-to-one relationship or a many-to-one relationship.
  • the second condition may be a condition in at least one exit condition corresponding to the first AI model.
  • the second condition is taken as an example to illustrate at least one exit condition.
  • the second condition may include one or more conditions in Table 3.
  • the second signal quality condition includes: the signal quality of the first device belongs to the second signal quality range.
  • the signal quality of the first device can refer to the description in the first signal quality condition, which will not be repeated here.
  • the second signal quality range is different from the first signal quality range. For example, if the signal strength of the first device is lower than the first signal quality threshold, and/or the maximum throughput of the first device is higher than the first throughput threshold, the first device may stop using AI model 1. For another example, if the transmission delay of the signal from the second device detected by the first device is greater than 80ms, the first device may stop using AI model 1. For another example, if the packet error rate of the signal from the second device detected by the first device and/or the call drop rate of the first device is greater than 20%, the first device may stop using AI model 1.
  • the second application scenario condition includes: the application scenario of the first device belongs to the second application scenario range.
  • the second application scenario range is different from the first application scenario range.
  • the first application scenario range includes CSI-RS feedback enhancement
  • the second application scenario range includes beam management enhancement.
  • CSI-RS feedback enhancement may correspond to AI model a1 and AI model a2
  • beam management enhancement may correspond to AI model a3 and AI model a4. If the first AI model is AI model a1, and the application scenario of the first device is changed from CSI-RS feedback enhancement to beam management enhancement, the first device may stop using AI model a1.
  • the third resource condition includes: the resources of the first device belong to the third resource range.
  • the second condition may include the third resource condition.
  • the third resource range is different from the first resource range. For example, when the central frequency of the BWP used by the terminal device is the second frequency and the bandwidth used by the terminal device is the second bandwidth, the terminal device may stop using the AI model b1.
  • the fourth resource condition includes: the resources of the second device belong to the fourth resource range.
  • the second condition may include the fourth resource condition.
  • the fourth resource range may be the same as or different from the third resource range. For example, when the center frequency of the BWP used by the terminal device is the second frequency and the bandwidth used by the terminal device is the second bandwidth, the network device may stop using the AI model b1.
  • the third area condition includes: the area where the first device is located belongs to the third area range.
  • the second condition may include the third area condition.
  • the third area range is different from the first area range. For example, if the terminal device is located in City B, the terminal device may stop using the AI model c1 corresponding to City A. For another example, if the terminal device accesses PLMN2, the terminal device may stop using the AI model c2 corresponding to PLMN1. For another example, if the identifier of the cell accessed by the terminal device is CGI2, the terminal device may stop using the AI model c3 corresponding to CGI1.
  • the fourth area condition includes: the area where the second device is located belongs to the fourth area range.
  • the second condition may include the fourth area condition.
  • the fourth area range and the third area range may be the same or different. For example, if the terminal device is located in City B, the network device may stop using the AI model c1 corresponding to City A. For another example, if the terminal device accesses PLMN2, the network device may stop using the AI model c2 corresponding to PLMN1. For another example, if the cell accessed by the terminal device is identified as CGI2, the network device may stop using the AI model c3 corresponding to CGI1.
  • the second capability condition includes: the capability of the first device belongs to the second capability range.
  • the second capability range is different from the first capability range.
  • the capability of the first device may include the remaining power of the first device. When the remaining power of the first device is greater than or equal to the first power threshold, the first device may stop using the AI model d1.
  • the capability of the first device includes the computing capability of the first device. When the computing rate of the first device is greater than or equal to the first rate threshold, the first device may stop using the AI model d2.
  • the second performance condition includes: the performance of the first AI model belongs to the second performance range.
  • the performance of the first AI model may include the prediction accuracy of the first AI model and/or the network performance after the network is optimized using the prediction results of the first AI model. For example, when the prediction accuracy of the first AI model is less than the second accuracy threshold (for example, within 1 minute, the prediction accuracy of the first AI model is less than 95%), the first device may stop using the first AI model. For example, if the results of the first AI model's predictions for K consecutive times are inaccurate, K is a positive integer. For example, in beam management enhancement, the optimal beam predicted by the first AI model for 5 consecutive times is not the actual optimal beam, and the first device may stop using the first AI model.
  • the first device may stop using the first AI model.
  • the expected performance target is to increase the throughput of the first device to X. If the throughput of the first device after the network is optimized using the prediction results of the first AI model is less than X, the first device may stop using the first AI model.
  • the second device can indicate the correspondence between at least one exit condition of the first device and at least one AI model, so that the first device can reasonably exit the AI model, thereby improving the efficiency of the first device in determining and using the AI model, enabling the first device to reasonably and effectively use the AI model, thereby improving the collaboration efficiency between devices.
  • the application scenario may be included in at least one condition and at least one exit condition, or may exist as independent information in the first information.
  • the first information may include: application scenario, model list, model sublist, model ID and content information, at least one condition, at least one exit condition, and the shortest execution time.
  • index 1 can be used to indicate an AI model with a model ID of XXX
  • index 1-1 can be used to indicate an AI model with a model ID of XXXX1
  • index 1-2 can be used to indicate an AI model with a model ID of XXXX2
  • the AI model with model ID XXXX1 and the AI model with model ID XXXX1 are sub-versions of the AI model with model ID XXX.
  • Index 2 can be used to indicate an AI model with a model ID of YYY
  • index 2-1 can be used to indicate an AI model with a model ID of YYYY1
  • index 2-2 can be used to indicate an AI model with a model ID of YYYY2, wherein the AI model with model ID YYYY1 and the AI model with model ID YYYY1 are sub-versions of the AI model with model ID YYY.
  • At least one condition in Table 4 can be divided into two pieces of information, one for the entry condition of at least one AI model, and the other for the change condition of at least one AI model.
  • S204 is an optional step.
  • the second information may be used to indicate that the AI model used by the first device has changed.
  • the change in the AI model used by the first device may include at least one of the following: the first device changes from not using the AI model to using the AI model, the first device changes from using one AI model to using another AI model, and the first device changes from using the AI model to not using the AI model.
  • the second information includes at least one of the following:
  • the time when the AI model used by the first device changes for example, when the first device changes from not using the AI model to using the AI model at 19:00, the second information includes 19:00.
  • Conditions that trigger changes in the AI model used by the first device For example, when the first device uses the first AI model because the first condition is met, the second information includes the first condition.
  • the network performance and/or AI model performance of the first device within the first time period before and/or the second time period after the AI model used by the first device changes: for example, information such as the throughput, packet error rate and/or transmission delay of the first device within the first time period before and/or the second time period after the AI model used by the first device changes.
  • the first device changes from using AI model 1-1 to using AI model 2-2.
  • the prediction accuracy of AI model 1-1 within the first time period before the AI model used by the first device changes, and/or the prediction accuracy of AI model 2-2 within the second time period after the AI model used by the first device changes.
  • the first time period and the second time period may be pre-set or obtained by the first device from the second device.
  • the indication information may be an index of the AI model.
  • the first device changes from using AI model 1-1 to using AI model 2-2, and the second information includes the index of AI model 1-1 and/or the index of AI model 2-2.
  • Scenario 1 The first device is terminal device 1, and the second device is network device 1.
  • the terminal device 1 may send second information to the network device 1 when the AI model changes; the terminal device may also send second information to the network device 1 to indicate the multiple changes after the AI model changes multiple times, wherein the second information may indicate the multiple changes of the AI model in a list format.
  • terminal device 1 may save the second information. For example, terminal device 1 may save the second information in the form of a list in a network self-optimization report.
  • terminal device 1 may send the second information to network device 1.
  • network device 1 sends a network self-optimization report containing the second information to network device 1.
  • terminal device 1 When terminal device 1 enters a connected state, it may be connected to network device 1 or to other network devices (for example, network device 2).
  • terminal device 1 When terminal device 1 is connected to network device 1, terminal device 1 may send the second information directly to network device 1.
  • terminal device 1 is connected to network device 2, The terminal device 1 may send the second information to the network device 1 through the network device 2. Specifically, the terminal device 1 may send the second information to the network device 2, and then the network device 2 sends the second information to the network device 1.
  • Scenario 2 The first device is network device 1, and the second device is terminal device 1.
  • the network device 1 may send the second information to the terminal device 1 when the AI model changes; the network device may also send the second information indicating that the AI model has changed multiple times to the terminal device 1 after the AI model has changed multiple times, wherein the second information may indicate the information that the AI model has changed multiple times in a list manner.
  • the second information may be carried in an RRC message.
  • the network device 1 may save the second information, for example, the network device 1 may save the second information in the form of a list in the context information of the terminal device 1; or the network device 1 sends the context information of the terminal device 1 containing the second information to the core network.
  • the network device to which the terminal device 1 is connected may obtain the second information from the network device 1 or the core network, and send the second information to the terminal device 1.
  • the second device may optimize the configuration of the AI model in the first device according to the second information. For example, if the performance of the first device decreases or does not improve after the first device switches from AI model 1 to AI model 2 (such as an increase in the transmission delay of the first device), the second device may modify the conditions for switching from AI model 1 to AI model 2 to increase the difficulty of the first device switching from AI model 1 to AI model 2.
  • condition a for switching AI model 1 to AI model 2 includes: the transmission delay of the signal from the second device detected by the first device is greater than 80ms
  • the second device may modify condition a for switching AI model 1 to AI model 2 to condition b
  • condition b includes: the transmission delay of the signal from the second device detected by the first device is greater than or equal to 40ms and less than or equal to 60ms.
  • the method shown in FIG2 may further include:
  • the first device may be a source network device of the terminal device
  • the third device may be a destination network device of the terminal device. That is, when the terminal device switches from the source network device to the destination network device, the source network device may send the first information from the terminal device to the destination network device.
  • the first information may be carried in an existing message or in a new message.
  • the source network device can send the first information from the terminal device to the destination network device.
  • the destination network device can also use the AI model according to the corresponding relationship between the conditions provided by the terminal device and the AI model.
  • the method shown in FIG2 further includes:
  • the second device may be a source network device of the terminal device
  • the fourth device may be a destination network device of the terminal device. That is, when the terminal device switches from the source network device to the destination network device, the source network device may send the first information to the destination network device.
  • the first information may be carried in an existing message or in a new message.
  • the source network device can send the first information to the destination network device.
  • the destination network device can also obtain the correspondence between the conditions provided by the source network device and the AI model, so that it can cooperate with the terminal device that uses the correspondence to operate.
  • an embodiment of the present application provides a communication method, which can be applied to the communication system shown in Figure 1. Referring to the flowchart shown in Figure 5, the process of the method is specifically described below.
  • S301 The second device sends fourth information to the first device.
  • the first device receives the fourth information from the second device.
  • the first device is a terminal device, and the second device is a network device; or, the first device is a network device, and the second device is a terminal device.
  • the fourth information may include configuration information of at least one AI model.
  • the specific content of the configuration information of at least one AI model can be referred to S202, which will not be repeated here.
  • the fourth information may be carried in an existing message or in a new message.
  • S302 The second device sends fifth information to the first device.
  • the first device receives the fifth information from the second device.
  • the fifth information may include indication information of a third AI model, where the third AI model is one of the at least one AI model.
  • the fifth information is used to instruct the first device to use the third AI model or switch from other AI models to the third AI model.
  • the indication information of the third AI model may be an index of the third AI model.
  • index of the third AI model For specific content of the index of the third AI model, reference may be made to the description of the index of the first AI model in S202, which will not be repeated here.
  • the correspondence between the model list, the model sub-list, and the model ID may be shown in Table 4.
  • the second device sends index 1 to the first device, it may instruct the first device to use the AI model with model ID XXX in the beam management enhanced scenario;
  • the second device sends index 1-1 to the first device, it may instruct the first device to use the model with model ID XXXX1 in the beam management enhanced scenario.
  • This method indicates the AI model through the index of the AI model. In this way, even if a third party obtains the index, it does not know the AI model corresponding to the index, thereby improving security.
  • the present application does not limit the method of sending the fifth information.
  • the fifth information can be carried in an RRC message, a MAC control element (MAC control element, MAC CE), downlink control information (downlink control Information, DCI) or uplink control information (uplink control information, UCI).
  • MAC control element MAC control element, MAC CE
  • DCI downlink control Information
  • UCI uplink control information
  • S303 The first device uses the third AI model.
  • S304 The second device sends the sixth information to the first device.
  • the first device receives the sixth information from the second device.
  • the sixth information may include indication information of the third AI model.
  • the sixth information may be used to instruct the first device to stop using the third AI device.
  • the specific content of the indication information of the third AI model may refer to S302, which will not be repeated here.
  • the sixth information can be carried in an RRC message, MAC CE, DCI or UCI.
  • S305 The first device stops using the third AI model.
  • S304-S305 are optional steps.
  • the second device after the second device sends the configuration information of at least one AI model to the first device, it can instruct the first device to use or stop using the third AI model through the indication information of the third AI model, thereby reducing the overhead and time required to instruct the first device to use or stop using the third AI model, thereby improving the collaboration efficiency between devices.
  • the embodiment of the present application provides a communication device through Figure 6, which can be used to perform the functions of the relevant steps in the above method embodiments.
  • the functions can be implemented by hardware, or by software or hardware executing corresponding software implementations.
  • the hardware or software includes one or more modules corresponding to the above functions.
  • the structure of the communication device is shown in Figure 6, including a communication unit 601 and a processing unit 602.
  • the communication device 600 can be applied to a network device (for example, the gNB in Figure 1) or a terminal device, and can implement the communication method provided in the above embodiments and examples of the present application.
  • the functions of each unit in the communication device 600 are introduced below.
  • the communication unit 601 is used to receive and send data.
  • the communication unit 601 can be implemented by a transceiver, for example, a mobile communication module.
  • the mobile communication module may include at least one antenna, at least one filter, a switch, a power amplifier, a low noise amplifier (LNA), etc.
  • LNA low noise amplifier
  • the processing unit 602 can be used to support the communication device 600 to perform the processing actions in the above method embodiment.
  • the processing unit 602 can be implemented by a processor.
  • the processor can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, transistor logic devices, hardware components or any combination thereof.
  • the general-purpose processor can be a microprocessor or any conventional processor.
  • the communication device 600 is applied to the first device in the embodiment of the present application shown in FIG. 2 or FIG. 3.
  • the first device may be a network device (for example, the gNB in FIG. 1) or a terminal device.
  • the specific functions of the processing unit 602 in this implementation are introduced below.
  • the processing unit 602 is used to receive first information from the second device through the communication unit 601, where the first information is used to indicate the correspondence between at least one condition and at least one AI model; when the first device meets the first condition, the first AI model corresponding to the first condition is used, and the first condition is any one of the at least one condition.
  • the first condition includes at least one of the following:
  • the signal quality of the first device satisfies a first signal quality condition
  • the application scenario of the first device meets the first application scenario condition
  • the resources of the first device meet the first resource condition
  • the resources of the second device served by the first device meet the second resource condition
  • the area where the first device is located meets the first area condition
  • the area where the second device served by the first device is located meets the second area condition
  • the capability of the first device satisfies the first capability condition
  • the performance of the second AI model currently used by the first device meets the first performance condition.
  • the first signal quality condition includes: the signal quality of the first device belongs to a first signal quality range;
  • the first application scenario condition includes: the application scenario of the first device belongs to the first application scenario range;
  • the first resource condition includes: the resources of the first device belong to the first resource range;
  • the second resource condition includes: the resources of the second device belong to the second resource range;
  • the first area condition includes: the area where the first device is located belongs to the first area range;
  • the second area condition includes: the area where the second device is located belongs to the second area range;
  • the first capability condition includes: the capability of the first device belongs to the first capability range; and/or
  • the first performance condition includes: the performance of the second AI model belongs to the first performance range.
  • the first information is also used to indicate the correspondence between at least one exit condition and at least one AI model
  • the processing unit 602 is specifically used to: after using the first AI model corresponding to the first condition, when the first device satisfies a second condition, stop using the first AI model, and the second condition is a condition in at least one exit condition corresponding to the first AI model.
  • the second condition includes at least one of the following:
  • the signal quality of the first device satisfies the second signal quality condition
  • the application scenario of the first device meets the second application scenario condition
  • the resources of the first device meet the third resource condition
  • the resources of the second device served by the first device meet the fourth resource condition
  • the area where the first device is located meets the third area condition
  • the area where the second device served by the first device is located meets the fourth area condition
  • the capability of the first device satisfies the second capability condition
  • the performance of the first AI model meets the second performance condition.
  • the second signal quality condition includes: the signal quality of the first device belongs to the second signal quality range;
  • the second application scenario condition includes: the application scenario of the first device belongs to the scope of the second application scenario;
  • the third resource condition includes: the resources of the first device belong to the third resource range;
  • the fourth resource condition includes: the resources of the second device belong to the fourth resource range;
  • the third area condition includes: the area where the first device is located belongs to the third area range;
  • the fourth area condition includes: the area where the second device is located belongs to the fourth area range;
  • the second capability condition includes: the capability of the first device belongs to the second capability range; and/or
  • the second performance condition includes: the performance of the first AI model falls within the second performance range.
  • the processing unit 602 is specifically used to: when the AI model used by the first device changes, send second information to the second device through the communication unit 601, and the second information is used to indicate that the AI model used by the first device has changed.
  • the second information includes at least one of the following:
  • the network performance and/or AI model performance of the first device within a first period of time before the AI model used by the first device changes and/or within a second period of time after the change;
  • the first device is a terminal device
  • the second device is a network device
  • the first device is a network device
  • the second device is a terminal device.
  • the processing unit 602 is specifically used to: when the second device switches from the first device to the third device, send the first information to the third device through the communication unit 601.
  • the processing unit 602 is specifically configured to: before receiving the first information from the second device, through the communication unit 601 Send third information to the second device, where the third information is used to indicate the capability of the first device and/or the application scenario of the first device, and the third information is used to determine the first information.
  • the communication device 600 is applied to the second device in the embodiment of the present application shown in FIG. 2 or FIG. 4.
  • the second device may be a network device (for example, the gNB in FIG. 1) or a terminal device.
  • the specific functions of the processing unit 602 in this implementation are introduced below.
  • the processing unit 602 is used to obtain first information, where the first information is used to indicate the correspondence between at least one condition and at least one AI model; and send the first information to the first device through the communication unit 601.
  • any of the at least one condition includes at least one of the following:
  • the signal quality of the first device satisfies a first signal quality condition
  • the application scenario of the first device meets the first application scenario condition
  • the resources of the first device meet the first resource condition
  • the resources of the second device served by the first device meet the second resource condition
  • the area where the first device is located meets the first area condition
  • the area where the second device served by the first device is located meets the second area condition
  • the capability of the first device satisfies the first capability condition
  • the performance of the second AI model currently used by the first device meets the first performance condition.
  • the first signal quality condition includes: the signal quality of the first device belongs to a first signal quality range;
  • the first application scenario condition includes: the application scenario of the first device belongs to the first application scenario range;
  • the first resource condition includes: the resources of the first device belong to the first resource range;
  • the second resource condition includes: the resources of the second device belong to the second resource range;
  • the first area condition includes: the area where the first device is located belongs to the first area range;
  • the second area condition includes: the area where the second device is located belongs to the second area range;
  • the first capability condition includes: the capability of the first device belongs to the first capability range; and/or
  • the first performance condition includes: the performance of the second AI model belongs to the first performance range.
  • the first information is also used to indicate the correspondence between at least one exit condition and at least one AI model.
  • the second condition is one of the at least one exit condition, and the second condition includes at least one of the following:
  • the signal quality of the first device satisfies the second signal quality condition
  • the application scenario of the first device meets the second application scenario condition
  • the resources of the first device meet the third resource condition
  • the resources of the second device served by the first device meet the fourth resource condition
  • the area where the first device is located meets the third area condition
  • the area where the second device served by the first device is located meets the fourth area condition
  • the capability of the first device satisfies the second capability condition
  • the performance of the first AI model meets the second performance condition.
  • the second signal quality condition includes: the signal quality of the first device belongs to the second signal quality range;
  • the second application scenario condition includes: the application scenario of the first device belongs to the scope of the second application scenario;
  • the third resource condition includes: the resources of the first device belong to the third resource range;
  • the fourth resource condition includes: the resources of the second device belong to the fourth resource range;
  • the third area condition includes: the area where the first device is located belongs to the third area range;
  • the fourth area condition includes: the area where the second device is located belongs to the fourth area range;
  • the second capability condition includes: the capability of the first device belongs to the second capability range; and/or
  • the second performance condition includes: the performance of the first AI model falls within the second performance range.
  • the processing unit 602 is specifically used to: receive second information from the first device through the communication unit 601, and the second information is used to indicate that the AI model used by the first device has changed.
  • the second information includes at least one of the following:
  • a condition that triggers a change in the AI model used by the first device
  • the network performance and/or AI model performance of the first device within a first period of time before the AI model used by the first device changes and/or within a second period of time after the change;
  • the first device is a terminal device
  • the second device is a network device
  • the first device is a network device
  • the second device is a terminal device.
  • the processing unit 602 is specifically used to: when the first device switches from the second device to the fourth device, send the first information to the fourth device through the communication unit 601.
  • the processing unit 602 is specifically used to: before obtaining the first information, receive third information from the first device through the communication unit 601, the third information is used to indicate the capability of the first device and/or the application scenario of the first device, and the third information is used to determine the first information.
  • the communication device 600 is applied to the first device in the embodiment of the present application shown in FIG5.
  • the first device may be a network device (for example, the gNB in FIG1) or a terminal device.
  • the specific functions of the processing unit 602 in this embodiment are introduced below.
  • the processing unit 602 is used to receive fourth information from the second device through the communication unit 601, the fourth information including configuration information of at least one artificial intelligence AI model; receive fifth information from the second device through the communication unit 601, the fifth information including indication information of a third AI model, the third AI model is one model of at least one AI model; and use the third AI model.
  • the indication information of the third AI model is an index of the third AI model.
  • the communication device 600 is applied to the second device in the embodiment of the present application shown in FIG5.
  • the second device may be a network device (for example, the gNB in FIG1) or a terminal device.
  • the specific functions of the processing unit 602 in this embodiment are introduced below.
  • the processing unit 602 is used to send fourth information to the first device through the communication unit 601, the fourth information including configuration information of at least one artificial intelligence AI model; send fifth information to the first device through the communication unit 601, the fifth information including indication information of the third AI model, the third AI model is one model of at least one AI model, and the fifth information is used to instruct the first device to use the third AI model.
  • the indication information of the third AI model is an index of the third AI model.
  • each functional unit in each embodiment of the present application may be integrated into a processing unit, or may exist physically separately, or two or more units may be integrated into one unit.
  • the above 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 the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product.
  • the computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) or a processor (processor) 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.
  • the embodiment of the present application provides a communication device as shown in FIG7, which can be used to perform the relevant steps in the above method embodiment.
  • the communication device can be applied to a network device (for example, the gNB in FIG1) or a terminal device, and can implement the communication method provided by the above embodiment and example of the present application, and has the function of the communication device shown in FIG6.
  • the communication device 700 includes: a communication module 701, a processor 702, and a memory 703. Among them, the communication module 701, the processor 702, and the memory 703 are interconnected.
  • the communication module 701, the processor 702 and the memory 703 are interconnected via a bus 704.
  • the bus 704 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus.
  • PCI peripheral component interconnect
  • EISA extended industry standard architecture
  • the bus may be divided into an address bus, a data bus, a control bus, and a control bus. For the convenience of representation, only one thick line is used in FIG7 , but it does not mean that there is only one bus or one type of bus.
  • the communication module 701 is used to receive and send data to realize communication interaction with other devices.
  • the communication module 701 can be realized through a physical interface, a communication module, a communication interface, and an input/output interface.
  • the processor 702 may be used to support the communication device 700 in executing the processing actions in the above method embodiment. When the communication device 700 is used to implement the above method embodiment, the processor 702 may also be used to implement the functions of the above processing unit 602.
  • the processor 702 may be a CPU, or other general-purpose processors, DSPs, ASICs, FPGAs, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof.
  • a general-purpose processor may be a microprocessor, or any conventional processor.
  • the communication device 700 is applied to the first device in the embodiment of the present application shown in FIG. 2 or 3.
  • the processor 702 is specifically used to: receive first information from the second device through the communication module 701, the first information is used to indicate the correspondence between at least one condition and at least one AI model; when the first device meets the first condition, use the first AI model corresponding to the first condition, the first condition is any one of the at least one condition.
  • the communication device 700 is applied to the second device in the embodiment of the present application shown in Figure 2 or Figure 4.
  • the processor 702 is specifically used to: obtain first information, where the first information is used to indicate the corresponding relationship between at least one condition and at least one AI model; and send the first information to the first device through the communication module 701.
  • the communication device 700 is applied to the first device in the embodiment of the present application shown in Figure 5.
  • the processor 702 is specifically used to: receive fourth information from the second device through the communication module 701, the fourth information including configuration information of at least one artificial intelligence AI model; receive fifth information from the second device through the communication module 701, the fifth information including indication information of a third AI model, the third AI model being one of the at least one AI model; and use the third AI model.
  • the communication device 700 is applied to the second device in the embodiment of the present application shown in Figure 5.
  • the processor 702 is specifically used to: send fourth information to the first device through the communication module 701, the fourth information includes configuration information of at least one artificial intelligence AI model; send fifth information to the first device through the communication module 701, the fifth information includes indication information of a third AI model, the third AI model is one of the at least one AI model, and the fifth information is used to instruct the first device to use the third AI model.
  • processor 702 can refer to the description of the communication method provided in the above embodiments and examples of the present application, as well as the specific functional description of the communication device 600 in the embodiment of the present application shown in Figure 6, which will not be repeated here.
  • the memory 703 is used to store program instructions and data.
  • the program instructions may include program codes, and the program codes include computer operation instructions.
  • the memory 703 may include RAM, and may also include non-volatile memory (non-volatile memory), such as at least one disk memory.
  • the processor 702 executes the program instructions stored in the memory 703, and uses the data stored in the memory 703 to implement the above functions, thereby realizing the communication method provided in the above embodiment of the present application.
  • the memory 703 in FIG. 7 of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories.
  • the non-volatile memory can be a ROM, a programmable read-only memory (Programmable ROM, PROM), an erasable programmable read-only memory (Erasable PROM, EPROM), an electrically erasable programmable read-only memory (Electrically EPROM, EEPROM) or a flash memory.
  • the volatile memory can be a RAM, which is used as an external cache.
  • RAM Direct Rambus RAM
  • SRAM Static RAM
  • DRAM Dynamic RAM
  • SDRAM Synchronous DRAM
  • DDR SDRAM Double Data Rate SDRAM
  • ESDRAM Enhanced SDRAM
  • SLDRAM Synchlink DRAM
  • DR RAM Direct Rambus RAM
  • the embodiments of the present application further provide a computer program, which, when executed on a computer, enables the computer to execute the methods provided in the above embodiments.
  • the embodiments of the present application further provide a computer-readable storage medium, in which a computer program is stored.
  • the computer program is executed by a computer, the computer executes the method provided in the above embodiments.
  • the storage medium may be any available medium that can be accessed by a computer.
  • a computer-readable medium may include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage medium or other magnetic storage device, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer.
  • an embodiment of the present application further provides a chip, which is used to read a computer program stored in a memory to implement the method provided in the above embodiments.
  • the embodiments of the present application provide a chip system, which includes a processor for supporting a computer device to implement the functions involved in each device in the above embodiments.
  • the chip system also includes a memory, which is used to store the necessary programs and data for the computer device.
  • the chip system can be composed of a chip, or it can include a chip and other discrete devices.
  • a first device may receive first information from a second device.
  • the first information may be used to indicate the correspondence between at least one condition and at least one AI model.
  • the first device may use the first AI model corresponding to the first condition, wherein the first condition is any one of the at least one condition.
  • the first device may select the AI model to be used based on the correspondence between at least one condition indicated by the second device and at least one AI model, thereby improving the efficiency of the first device in determining and using the AI model, so that the first device can use the AI model reasonably and effectively, thereby improving the efficiency of collaboration between devices.
  • the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.
  • a computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
  • These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.
  • These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

Landscapes

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

Abstract

本申请公开了一种通信方法及装置。该方法包括：第一设备可接收来自第二设备的第一信息。其中，第一信息可用于指示至少一个条件和至少一个AI模型的对应关系。当第一设备满足第一条件时，第一设备可使用与第一条件对应的第一AI模型，其中，第一条件为至少一个条件中的任一条件。通过该方法，第一设备可根据第二设备指示的至少一个条件和至少一个AI模型的对应关系，选择使用的AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。

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11月04日提交中国专利局、申请号为202211379923.3、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)是一种模拟人脑进行复杂计算的技术。随着数据存储和计算能力的提升，AI技术得到了越来越多的运用。第三代合作伙伴计划(3rd generation partnership project，3GPP)提出将AI运用到新空口(new radio，NR)通信系统中，通过智能收集和数据分析，提升网络性能和用户体验。
然而，如何在通信系统中合理、有效地使用AI模型，以提升设备之间的协作效率，是需要解决的问题。
发明内容
本申请提供了一种通信方法及装置，用以提升设备之间的协作效率。
第一方面，本申请实施例提供了一种通信方法。该方法包括：第一设备可接收来自第二设备的第一信息。其中，第一信息可用于指示至少一个条件和至少一个AI模型的对应关系。当第一设备满足第一条件时，第一设备可使用与第一条件对应的第一AI模型，第一条件为至少一个条件中的任一条件。
可选的，第一设备为终端设备，第二设备为网络设备；或者第一设备为网络设备，第二设备为终端设备。
通过该方法，第一设备可根据第二设备指示的至少一个条件和至少一个AI模型的对应关系，选择使用的AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。
并且，在该方法中，第一设备可根据对应关系选择AI模型，这样，无需第二设备根据第一设备的测量信息等为第一设备选择AI模型，从而可降低第二设备获取第一设备的测量信息的开销。
在一种可能的设计中，第一条件包括以下至少一项：
第一设备的信号质量满足第一信号质量条件；
第一设备的应用场景满足第一应用场景条件；
第一设备的资源满足第一资源条件；
第一设备服务的第二设备的资源满足第二资源条件；
第一设备所处的区域满足第一区域条件；
第一设备服务的第二设备所处的区域满足第二区域条件；
第一设备的能力满足第一能力条件；
第一设备当前使用的第二AI模型的性能满足第一性能条件。
可选的，第一信号质量条件包括：第一设备的信号质量属于第一信号质量范围；
第一应用场景条件包括：第一设备的应用场景属于第一应用场景范围；
第一资源条件包括：第一设备的资源属于第一资源范围；
第二资源条件包括：第二设备的资源属于第二资源范围；
第一区域条件包括：第一设备所处的区域属于第一区域范围；
第二区域条件包括：第二设备所处的区域属于第二区域范围；
第一能力条件包括：第一设备的能力属于第一能力范围；和/或
第一性能条件包括：第二AI模型的性能属于第一性能范围。
该设计提供了多种可能的条件，使得第一设备能够合理、有效的使用AI模型，且易于实现。
在一种可能的设计中，第一信息还用于指示至少一个退出条件和至少一个AI模型的对应关系。在使用与第一条件对应的第一AI模型之后，当第一设备满足第二条件时，第一设备可停止使用第一AI模型，第二条件为至少一个退出条件中与第一AI模型对应的条件。
通过该设计，第二设备可指示第一设备至少一个退出条件和至少一个AI模型的对应关系，这样，第一设备可合理退出AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。
在一种可能的设计中，第二条件包括以下至少一项：
第一设备的信号质量满足第二信号质量条件；
第一设备的应用场景满足第二应用场景条件；
第一设备的资源满足第三资源条件；
第一设备服务的第二设备的资源满足第四资源条件；
第一设备所处的区域满足第三区域条件；
第一设备服务的第二设备所处的区域满足第四区域条件；
第一设备的能力满足第二能力条件；
第一AI模型的性能满足第二性能条件。
可选的，第二信号质量条件包括：第一设备的信号质量属于第二信号质量范围；
第二应用场景条件包括：第一设备的应用场景属于第二应用场景范围；
第三资源条件包括：第一设备的资源属于第三资源范围；
第四资源条件包括：第二设备的资源属于第四资源范围；
第三区域条件包括：第一设备所处的区域属于第三区域范围；
第四区域条件包括：第二设备所处的区域属于第四区域范围；
第二能力条件包括：第一设备的能力属于第二能力范围；和/或
第二性能条件包括：第一AI模型的性能属于第二性能范围。
该设计提供了多种可能的退出条件，使得第一设备能够合理、有效的退出AI模型，且易于实现。
在一种可能的设计中，当第一设备使用的AI模型发生变动时，第一设备可向第二设备发送第二信息，第二信息用于指示第一设备使用的AI模型发生变动。通过该设计，第一设备可通知第二设备AI模型发生变动的信息。这样，第二设备可根据第一设备反馈的第二信息合理设置条件与AI模型的对应关系，使得第一设备合理选择和使用AI模型，从而提高系统性能。
在一种可能的设计中，第二信息包括以下至少一项：
第一设备使用的AI模型发生变动的时间；
触发第一设备使用的AI模型发生变动的条件；
第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，第一设备的网络性能和/或AI模型性能；
第一设备使用的AI模型发生变动前和/或变化后，第一设备所使用的AI模型的指示信息。
通过该设计，第一设备可灵活通知第二设备AI模型发生变动的信息。
在一种可能的设计中，若第一设备为网络设备，当第二设备从第一设备向第三设备切换时，第一设备可向第三设备发送第一信息。其中，第二设备可为终端设备，第一设备可为终端设备的源网络设备，第三设备可为终端设备的目的网络设备。通过该设计，当终端设备从源网络设备向目的网络设备切换时，源网络设备可向目的网络设备发送来自终端设备的第一信息。这样，目的网络设备也可根据终端设备提供的条件和AI模型的对应关系来使用AI模型。
在一种可能的设计中，在接收来自第二设备的第一信息之前，第一设备可向第二设备发送第三信息，第三信息用于指示第一设备的能力和/或第一设备的应用场景，第三信息用于确定第一信息。通过该设计，第二设备可合理确定AI模型和各条件的对应关系。
第二方面，本申请实施例提供了一种通信方法。该方法包括：第二设备可获取第一信息，第一信息用于指示至少一个条件和至少一个AI模型的对应关系。然后，第二设备可向第一设备发送第一信息。
可选的，第一设备为终端设备，第二设备为网络设备；或者第一设备为网络设备，第二设备为终端设备。
通过该方法，第二设备可向第一设备发送用于指示的至少一个条件和至少一个AI模型的对应关系的第一信息。这样，第一设备可根据该对应关系选择使用的AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。
在一种可能的设计中，至少一个条件中的任一条件包括以下至少一项：
第一设备的信号质量满足第一信号质量条件；
第一设备的应用场景满足第一应用场景条件；
第一设备的资源满足第一资源条件；
第一设备服务的第二设备的资源满足第二资源条件；
第一设备所处的区域满足第一区域条件；
第一设备服务的第二设备所处的区域满足第二区域条件；
第一设备的能力满足第一能力条件；
第一设备当前使用的第二AI模型的性能满足第一性能条件。
可选的，第一信号质量条件包括：第一设备的信号质量属于第一信号质量范围；
第一应用场景条件包括：第一设备的应用场景属于第一应用场景范围；
第一资源条件包括：第一设备的资源属于第一资源范围；
第二资源条件包括：第二设备的资源属于第二资源范围；
第一区域条件包括：第一设备所处的区域属于第一区域范围；
第二区域条件包括：第二设备所处的区域属于第二区域范围；
第一能力条件包括：第一设备的能力属于第一能力范围；和/或
第一性能条件包括：第二AI模型的性能属于第一性能范围。
该设计提供了多种可能的条件，使得第一设备能够合理、有效的使用AI模型，且易于实现。
在一种可能的设计中，第一信息还用于指示至少一个退出条件和至少一个AI模型的对应关系。这样，第一设备可合理退出AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。
在一种可能的设计中，第二条件为至少一个退出条件中的一个条件，第二条件包括以下至少一项：
第一设备的信号质量满足第二信号质量条件；
第一设备的应用场景满足第二应用场景条件；
第一设备的资源满足第三资源条件；
第一设备服务的第二设备的资源满足第四资源条件；
第一设备所处的区域满足第三区域条件；
第一设备服务的第二设备所处的区域满足第四区域条件；
第一设备的能力满足第二能力条件；
第一AI模型的性能满足第二性能条件。
可选的，第二信号质量条件包括：第一设备的信号质量属于第二信号质量范围；
第二应用场景条件包括：第一设备的应用场景属于第二应用场景范围；
第三资源条件包括：第一设备的资源属于第三资源范围；
第四资源条件包括：第二设备的资源属于第四资源范围；
第三区域条件包括：第一设备所处的区域属于第三区域范围；
第四区域条件包括：第二设备所处的区域属于第四区域范围；
第二能力条件包括：第一设备的能力属于第二能力范围；和/或
第二性能条件包括：第一AI模型的性能属于第二性能范围。
该设计提供了多种可能的退出条件，使得第一设备能够合理、有效的退出AI模型，且易于实现。
在一种可能的设计中，第二设备可接收来自第一设备的第二信息，第二信息用于指示第一设备使用的AI模型发生变动。通过该设计，第一设备可通知第二设备AI模型发生变动的信息。这样，第二设备可根据第一设备反馈的第二信息合理设置条件与AI模型的对应关系，使得第一设备合理选择和使用AI模型，从而提高系统性能。
在一种可能的设计中，第二信息可包括以下至少一项：
第一设备使用的AI模型发生变动的时间；
触发第一设备使用的AI模型发生变动的条件；
第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，第一设备的网络性能和/或AI模型性能；
第一设备使用的AI模型发生变动前和/或变化后，第一设备所使用的AI模型的指示信息。
通过该设计，第一设备可灵活通知第二设备AI模型发生变动的信息。
在一种可能的设计中，若第二设备为网络设备，当第一设备从第二设备向第四设备切换时，第二设备可向第四设备发送第一信息。其中，第一设备可为终端设备，第二设备可为终端设备的源网络设备，第四设备可为终端设备的目的网络设备。通过该设计，当终端设备从源网络设备向目的网络设备切换时，源网络设备可向目的网络设备发送第一信息。这样，目的网络设备也可获取源网络设备提供的条件和AI模型的对应关系，从而可与使用该对应关系进行操作的终端设备进行协作。
在一种可能的设计中，在获取第一信息之前，第二设备可接收来自第一设备的第三信息，第三信息用于指示第一设备的能力和/或第一设备的应用场景，第三信息用于确定第一信息。通过该设计，第二设备可根据第三信息合理确定AI模型和各条件的对应关系。
第三方面，本申请实施例提供了一种通信方法。该方法包括：第一设备可接收来自第二设备的第四信息，第四信息包括至少一个AI模型的配置信息。第一设备可在接收来自第二设备的第五信息后，使用第三AI模型。其中，第五信息包括第三AI模型的指示信息，第三AI模型为至少一个AI模型中的一个模型。
通过该方法，第二设备在向第一设备发送至少一个AI模型的配置信息后，通过第三AI模型的指示信息即可指示第一设备使用第三AI模型，从而可降低指示第一设备使用第三AI模型所需的开销和时间，进而能够提升设备之间的协作效率。
在一种可能的设计中，第三AI模型的指示信息为第三AI模型的索引。该设计通过AI模型的索引来指示AI模型，这样，即便第三方获取到索引，也不知道该索引对应的AI模型，从而可提高安全性。
第四方面，本申请实施例提供了一种通信方法。该方法包括：第二设备向第一设备发送第四信息，第四信息包括至少一个AI模型的配置信息。然后，第二设备可向第一设备发送第五信息，第五信息包括第三AI模型的指示信息，第三AI模型为至少一个AI模型中的一个模型，第五信息用于指示第一设备使用第三AI模型。
通过该方法，第二设备在向第一设备发送至少一个AI模型的配置信息后，通过第三AI模型的指示信息即可指示第一设备使用第三AI模型，从而可降低指示第一设备使用第三AI模型所需的开销和时间，进而能够提升设备之间的协作效率。
在一种可能的设计中，第三AI模型的指示信息为第三AI模型的索引。该设计通过AI模型的索引来指示AI模型，这样，即便第三方获取到索引，也不知道该索引对应的AI模型，从而可提高安全性。
第五方面，本申请实施例提供了一种通信装置，包括用于执行以上任一方面中各个步骤的单元。
第六方面，本申请实施例提供了一种通信装置，包括至少一个处理元件和至少一个存储元件，其中该至少一个存储元件用于存储程序和数据，该至少一个处理元件用于读取并执行存储元件存储的程序和数据，以使得本申请以上任一方面提供的方法被实现。
第七方面，本申请实施例提供了一种通信系统，包括：用于执行第一方面提供的方法的第一设备，用于执行第二方面提供的方法的第二设备。
第八方面，本申请实施例提供了一种通信系统，包括：用于执行第三方面提供的方法的第一设备，用于执行第四方面提供的方法的第二设备。
第九方面，本申请实施例还提供了一种计算机程序，当所述计算机程序在计算机上运行时，使得所述计算机执行上述任一方面提供的方法。
第十方面，本申请实施例还提供了一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机程序，当所述计算机程序被计算机执行时，使得所述计算机执行上述任一方面提供的方法。
第十一方面，本申请实施例还提供了一种芯片，所述芯片用于读取存储器中存储的计算机程序，执行上述任一方面提供的方法。
第十二方面，本申请实施例还提供了一种芯片系统，该芯片系统包括处理器，用于支持计算机装置实现上述任一方面提供的方法。在一种可能的设计中，所述芯片系统还包括存储器，所述存储器用于保存该计算机装置必要的程序和数据。该芯片系统可以由芯片构成，也可以包含芯片和其他分立器件。
上述第五方面至第十二方面中任一方面可以达到的技术效果可以参照上述第一方面至第四方面中任一方面中任一种可能设计可以达到的技术效果说明，重复之处不予论述。
附图说明
图1为本申请实施例应用的通信系统的架构示意图；
图2为本申请实施例提供的第一种通信方法的流程示意图；
图3为本申请实施例提供的第二种通信方法的流程示意图；
图4为本申请实施例提供的第三种通信方法的流程示意图；
图5为本申请实施例提供的第四种通信方法的流程示意图；
图6为本申请的实施例提供的一种通信装置的结构示意图；
图7为本申请的实施例提供的另一种通信装置的结构示意图。
具体实施方式
下面结合说明书附图对本申请实施例做详细描述。
本申请实施例可以应用于各种移动通信系统，例如：新无线(new radio，NR)系统、全球移动通讯(global system of mobile communications，GSM)系统、长期演进(long term evolution，LTE)系统以及未来通信系统等其它通信系统，在此不做限制。
本申请实施例中，以终端设备、接入网设备以及核心网设备之间的交互为例进行描述，需要说明的是，本申请实施例提供的方法，不仅可以应用于终端设备与网络侧之间的交互，还可以应用于任意两个设备之间的交互中，本申请实施例对此并不限定。
本申请实施例中，终端设备可为用户提供语音和/或数据连通性服务。终端设备可以是无线终端设备，也可以为无线终端设备。无线终端设备是一种具有无线收发功能的设备，例如，具有无线连接功能的设备或连接到无线调制器的设备。终端设备可以部署在陆地上，包括室内或室外、手持或车载；也可以部署在水面上(如轮船等)；还可以部署在空中(例如飞机、气球和卫星上等)。终端设备还可称为用户设备(user equipment，UE)、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)、终端设备(user device，UD)。
示例性的，终端设备可以是：手机、平板电脑、笔记本电脑、掌上电脑、移动互联网设备(mobile internet device，MID)、移动蜂窝电话、无绳电话、个人数字助理(personal digital assistant，PDA)、客户终端设备(customer-premises equipment，CPE)、智能销售点(point of sale，POS)机、个人通信业务(personal communication service，PCS)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、可穿戴设备(例如智能手表、智能手环、计步器等)、车载设备(例如，汽车、自行车、电动车、飞机、船舶、火车、高铁等)、虚拟现实(virtual reality，VR)设备、增强现实(augmented reality，AR)设备、工业控制中的无线终端、智能家居设备(例如，冰箱、电视、空调、电表等)、智能机器人、车间设备、无人驾驶中的无线终端、远程手术中的无线终端、智能电网(smart grid)中的无线终端、运输安全中的无线终端、智慧城市中的无线终端，智能电话、笔记本电脑、平板电脑、无线数据卡、无线调制解调器(modulator demodulator，Modem)、飞行设备(例如，智能机器人、热气球、无人机、飞机)、家用电器、交通工具、工具设备、服务设备或服务设施等。
本申请实施例中所涉及到的网络设备，可以是无线接入网(radio access network，RAN)中的基站(如下一代基站(generation Node B，gNB))等。如图1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(long term evolution，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，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)层以及物理(physical)层等的功能。该划分方式仅是一种举例，CU和DU也可以按照其他方式进行划分。例如，与上述方式相比，CU或者DU可具有更多协议层的功能。还例如，CU或DU可具有上述划分方式中的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，即集中式单元控制面(central unit-control plane，CU-CP)节点或者集中式单元用户面(central unit-userl plane，CU-UP)节点分离。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，CU-CP和CU-UP可以与DU相耦合，共同完成基站的功能。可选的，CU的控制面CU-CP还可切分为CU-CP1和CU-CP2。其中，CU-CP1用于实现无线资源管理功能，CU-CP2用于实现RRC功能和分组数据汇聚层协议控制(packet data convergence protocol control，PDCP-C)功能(即控制面信令在PDCP层的基本功能)。
为便于理解本申请，下面针对本申请中涉及的一些名词或术语进行说明。
一、AI应用场景(use case)
AI应用场景包括但不限于以下至少一项：节能(energy saving)、负载均衡(load balancing)、移动性优化(mobility optimization)、信道状态信息参考信号(channel status information reference signal，CSI-RS)反馈增强(CSI-RS feedback enhancement)、波束管理增强(beam management enhancement)和定位增强(positioning accuracy enhancement)。下面分别对此进行说明。
1)节能：
网络设备可根据自身和邻区的负载、能耗、能效等信息，以及终端设备的轨迹、测量结果等信息，对网络设备的负载进行预测。在不影响网络覆盖和用户接入的前提下，网络设备可根据预测结果适时适当采取节能措施。其中，节能措施可包括以下至少一项：将小区去激活、关断载波、关断通道、关断时隙、降低发射功率等。
2)负载均衡：
网络设备可根据自身和邻区的负载、能耗、能效等信息，以及终端设备的轨迹、测量结果等信息，对网络设备的负载进行预测。网络设备可根据预测结果，将部分终端设备切换到邻区，或将邻区服务的部分终端设备切换到本小区，从而使得网络中各网络设备的负载接近，避免出现部分网络设备的负载过重，影响终端设备的业务，而另一部分的网络设备的资源又处于闲置状态。
3)移动性优化：
网络设备可根据终端设备的历史轨迹信息和终端设备的测量信息，对终端设备的未来轨迹进行预测。网络设备可根据预测结果，提前判断终端设备是否需要切换，并提前为需要进行切换的终端设备配置切换所需的信息，以及通知目标小区为该终端设备准备接入资源，从而可降低终端设备在切换过程中的延迟，提高终端设备切换的成功率。
4)CSI-RS反馈增强：
网络设备可根据终端设备的能力等信息为终端设备提供编码器和量化工具。终端设备根据该编码器和量化工具对CSI-RS进行压缩和量化后，可向网络设备发送CSI-RS。网络设备可根据CSI-RS确定终端设备的信道质量。
5)波束管理增强：
网络设备可从多个终端设备获取全波束扫描结果，并据此训练出稀疏扫描矩阵。然后，网络设备可 在向终端设备a发送稀疏扫描矩阵之后，从终端设备a获取稀疏扫描结果，并根据稀疏扫描结果确定终端设备a的最优CSI-RS波束。
6)定位增强：
定位增强用于提升定位准确率。定位增强可以包括以下至少一项：基于接入网设备的定位增强、基于定位管理功能网元的定位增强、基于终端设备的定位增强。
二、AI模型
AI模型是AI功能的具体实现，可表征模型的输入和输出之间的映射关系。AI模型可以是神经网络、线性回归模型、决策树模型、支持向量机(support vector machine，SVM)、贝叶斯网络、Q学习模型或者其他机器学习模型等。AI功能可以包括以下至少一项：数据收集(收集训练数据和/或推理数据)、数据预处理、模型训练(或称为，模型学习)、模型信息发布(配置模型信息)、模型校验、模型推理、或推理结果发布。其中，推理又可以称为预测。
AI模块具有机器学习计算能力，是用于实现AI模型的模块。在通信系统中，AI模块可位于操作维护管理(operation administration and maintenance，OAM)中，也可位于网络设备中，也可以位于终端设备中，还可以是单独的控制层，例如，AI控制层(AI control layer，AIC)。
AI模块的主要功能可包括：根据输入数据，进行AI模型建立、训练逼近、强化学习等一系列AI计算。在通信系统中，输入数据可包括网络设备提供的或OAM监测的网络运行数据，例如，网络负载、信道质量等数据。
AI模块提供的已训练完成的模型具有针对RAN侧网络变化的预测功能，可以用于负载预测和/或终端设备的轨迹预测等。此外，AI模块还可以根据训练完成的模型对RAN网络性能的预测结果，推测出节能策略和/或移动性优化策略等。
当AI模块位于OAM中时，AI模块可通过北向接口与网络设备通信；当AI模块位于gNB或CU中时，AI模块可通过F1、Xn、Uu等接口与其他设备进行通信；当AI模块作为一个独立的网络实体时，AI模块可建立AI模块到OAM和/或RAN侧的通信链路，该通信链路可为有线链路或无线链路。
另外，当CU的CP和UP分离时，CU-CP可用于接收AI模型以及执行AI推理和策略生成等功能。当CU-CP包括CU-CP1和CU-CP2时，CU-CP1可用于接收AI模型以及执行AI推理等功能，生成交互信令，并通过CU-CP2发送该交互信令。
终端设备可支持多个AI模型。网络设备可根据终端设备的测量结果等信息确定终端设备当前应使用的AI模型，并指示终端设备使用该AI模型。由于终端设备在不同情况下使用的AI模型可能不同，因此，网络设备需及时从终端设备获取用于确定AI模型的信息，开销较大，且时延较长，设备之间的协作效率较低。
为解决上述问题，本申请实施例提供了一种通信方法，该方法可应用于图1所示的通信系统中。下面参阅图2所示的流程图，对该方法的流程进行具体说明。
S201：第一设备向第二设备发送第三信息。相应的，第二设备接收来自第一设备的第三信息。
可选的，第一设备为终端设备，第二设备为网络设备；或者，第一设备为网络设备，第二设备为终端设备。
其中，第三信息可用于指示第一设备的能力和/或第一设备的应用场景。第三信息可包括第一设备的能力和/或第一设备的应用场景，也可包括与第一设备的能力和/或第一设备的应用场景存在对应关系的信息。
当第一设备为终端设备时，第一设备的能力包括但不限于以下至少一项：
1、终端设备的缓存能力：例如，终端设备的缓存大小为50兆字节(MB)，表示终端设备可存储容量大小为50MB的AI模型。
2、终端设备的AI计算能力：例如，终端设备的AI计算能力为1太(Tera，T，即1012)每秒浮点运算次数(Floating-point operations per second，FLPS)。
3、终端设备的电量：例如，终端设备的可用电量为2000毫安。
4、终端设备的AI模型信息：例如，终端设备中用于实现AI功能的软件库(例如，张量流系统(Tensflow))。又例如，终端设备支持的AI模型格式(例如，开放神经网络交换(open neural network  exchange，ONNX))。
5、终端设备本地已存储的AI模型的指示信息：该指示信息可为与AI模型存在对应关系的信息。例如，该指示信息可为模型标识(identifier，ID)。其中，模型ID可以是多个比特组成的比特流。该比特流不仅可指示AI模型，还可以指示AI模型的厂商信息、版本信息、编译平台信息等中的一个或多个。又例如，该指示信息为与AI模型的索引，该索引与AI模型的ID对应。
6、终端设备期望的AI模型的性能：例如，终端设备期望使用AI模型后吞吐量可达到的目标值。又例如，终端设备期望使用AI模型的预测准确率高于95％。
当第一设备为网络设备时，第一设备的能力包括但不限于以下至少一项：
1、网络设备的缓存能力：例如，网络设备的缓存大小为50MB，表示网络设备可存储容量大小为50MB的AI模型。
2、网络设备的AI计算能力：例如，网络设备的AI计算能力为1T FLPS。
3、网络设备的AI模型信息：例如，网络设备中用于实现AI功能的软件库(例如，Tensflow)。又例如，网络设备支持的AI模型格式(例如，ONNX)。
4、网络设备本地已存储的AI模型的指示信息：该指示信息可为与AI模型存在对应关系的信息。例如，该指示信息可为模型ID。其中，模型ID可以是多个比特组成的比特流。该比特流不仅可指示AI模型，还可以指示AI模型的厂商信息、版本信息、编译平台信息等中的一个或多个。又例如，该指示信息为与AI模型的索引，该索引与AI模型的ID对应。
5、网络设备期望的AI模型的性能：例如，网络设备期望使用AI模型后吞吐量可达到的目标值。又例如，网络设备期望使用AI模型的预测准确率高于95％。
第一设备的应用场景包括但不限于以下至少一项：节能、负载均衡、移动性优化、CSI-RS反馈增强、波束管理增强和定位增强。
本申请中，第三信息可携带在现有的消息中，也可以携带在新的消息中。
第一设备可主动向第二设备发送第三信息，也可以基于第二设备的请求向第二设备发送第三信息。当第一设备主动向第二设备发送第三信息时，第一设备可按照设定周期发送第三信息，也可基于事件触发向第二设备发送第三信息。其中，设定周期可为预定义的，也可以为第一设备从第二设备获取的。触发事件可为第一设备的能力和/或应用场景发生变化。
在本申请中，S201为可选的步骤。
S202：第二设备向第一设备发送第一信息。相应的，第一设备接收来自第二设备的第一信息。
其中，第一信息可用于指示至少一个条件和至少一个AI模型的对应关系。
在一些可能的方式中，至少一个条件和至少一个AI模型可以是一对一的关系。例如，至少一个条件包括条件1和条件2，至少一个AI模型包括AI模型1和AI模型2，条件1和AI模型1对应，条件2和AI模型2对应。
在另一些可能的方式中，至少一个条件和至少一个AI模型可以是多对一的关系。例如，至少一个条件包括条件1-条件4，至少一个AI模型包括AI模型1和AI模型2，条件1和条件4与AI模型1对应，条件2和条件3与AI模型2对应。
第一条件为至少一个条件中的任一条件，第一条件与至少一个AI模型中的第一AI模型对应。可选的，第一条件可包括表1所示的一个或多个条件。
表1

下面分别对第一条件可包括的每个条件进行说明。
第一信号质量条件可包括：第一设备的信号质量属于第一信号质量范围。其中，第一信号的信号质量包括以下至少一项：第一设备检测到的来自第二设备的信号强度、第一设备的吞吐量、第一设备检测到的来自第二设备的信号的传输时延、第一设备检测到的来自第二设备的信号的误包率、第一设备的掉话率。例如，若第一设备的信号强度在连续n秒内持续超过第一信号质量阈值(例如，100分贝毫瓦(decibel relative to one milliwatt，dBm))，且第一设备的最大吞吐量小于第一吞吐量阈值，则第一设备可使用AI模型1，其中，n为正整数。还例如，若第一设备检测到的来自第二设备的信号的传输时延大于80毫秒(ms)，则第一设备可从AI模型1切换到AI模型2。又例如，若第一设备检测到的来自第二设备的信号的误包率和/或第一设备的掉话率大于20％，则第一设备可从AI模型1切换到AI模型2。再例如，若第一设备检测到的来自第二设备的信号强度大于-100dBm，且第一设备的掉话率大于20％，则第一设备可使用AI模型1。
第一应用场景条件包括：第一设备的应用场景属于第一应用场景范围。该第一应用场景条件可用于指示应用场景或应用子场景与AI模型的对应关系。例如，CSI-RS反馈增强可对应于AI模型a1和AI模型a2，波束管理增强可对应于AI模型a3和AI模型a4。波束管理增强场景中，第一设备和第二设备使用波束对1进行通信时，第一设备可使用AI模型a3；第一设备和第二设备使用波束对2进行通信时，第一设备可使用AI模型a4。
第一资源条件包括：第一设备的资源属于第一资源范围。可选的，第一设备为终端设备时，第一条件可包括第一资源条件。此时，终端设备使用的上下行时频资源可与AI模型存在对应关系。例如，当终端设备使用的带宽部分(bandwidth part，BWP)的中心频点为第一频点和/或终端设备使用的带宽为第一带宽时，终端设备可使用AI模型b1。
第二资源条件包括：第二设备的资源属于第二资源范围。可选的，第一设备为网络设备且第二设备为终端设备时，第一条件可包括第二资源条件。第二资源范围和第一资源范围可以相同，也可以不同。例如，当终端设备使用的BWP的中心频点为第一频点和/或终端设备使用的带宽为第一带宽时，网络设备可使用AI模型b1。
第一区域条件包括：第一设备所处的区域属于第一区域范围。可选的，第一设备为终端设备时，第一条件可包括第一区域条件。
在一种可能的方式中，终端设备所处的区域可为终端设备所在的地理位置区域，该地理位置区域可由行政区域、经纬度和/或高度等信息确定。例如，若终端设备位于A市，则终端设备可使用与A市对应的AI模型c1。
在另一些可能的方式中，终端设备所处的区域可通过接入的公共陆地移动网(public land mobile network，PLMN)和/或接入小区的小区全球标识(Cell Global Identifier，CGI)来确定。例如，若终端设备接入到PLMN1，则终端设备可使用与PLMN1对应的AI模型c2。又例如，若终端设备接入的小区的标识为CGI1，则终端设备可使用与CGI1对应的AI模型c3。
第二区域条件包括：第二设备所处的区域属于第二区域范围。可选的，第一设备为网络设备且第二设备为终端设备时，第一条件可包括第二区域条件。第二区域范围和第一区域范围可以相同，也可以不同。
在一种可能的方式中，终端设备所处的区域可为终端设备所在的地理位置区域，该地理位置区域可由行政区域、经纬度和/或高度等信息确定。例如，若终端设备位于A市，则网络设备可使用与A市对应的AI模型c1。其中，终端设备所处的区域可以是终端设备发给网络设备的。
在另一些可能的方式中，终端设备所处的区域可通过接入的公共陆地移动网(public land mobile network，PLMN)和/或接入小区的小区全球标识(Cell Global Identifier，CGI)来确定。例如，若终端设备接入到PLMN1，则网络设备可使用与PLMN1对应的AI模型c2。又例如，若终端设备接入的小 区的标识为CGI1，则网络设备可使用与CGI1对应的AI模型c3。
第一能力条件包括：第一设备的能力属于第一能力范围。例如，第一设备的能力可包括第一设备的剩余电量。当第一设备的剩余电量小于第一电量阈值时，第一设备可使用AI模型d1。又例如，第一设备的能力包括第一设备的计算能力。当第一设备的计算速率小于第一速率阈值时，第一设备可使用AI模型d2。
第一性能条件包括：第二AI模型的性能属于第一性能范围。其中，第二AI模型可为第一设备当前使用的AI模型。第二AI模型的性能可包括第二AI模型的预测准确率和/或使用第二AI模型的预测结果进行网络优化后的网络性能。例如，当第二AI模型的预测准确率(例如，在波束管理增强中，预测出的最优波束是实际最优波束的概率)小于第一准确率阈值(例如，1分钟内，第二AI模型的预测准确率小于95％)时，第一设备可将当前使用的AI模型切换为第一AI模型。还例如，若第二AI模型连续M次预测的结果都不准确，M为正整数，例如，在波束管理增强中，第二AI模型连续5次预测出的最优波束都不是实际最优波束，第一设备可将当前使用的AI模型切换为第一AI模型。又例如，使用第二AI模型的预测结果进行网络优化后的网络性能未达到预期的性能目标时，第一设备可将当前使用的AI模型切换为第一AI模型。示例性的，预期的性能目标为将第一设备的吞吐量提升至X，若使用第二AI模型的预测结果进行网络优化后第一设备的吞吐量小于X，则第一设备可将当前使用的AI模型切换为第一AI模型。
当第一设备满足第一条件时，第一设备可使用第一条件对应的第一AI模型。因此，第一条件可称为第一AI模型的进入条件。若第一设备满足第一条件时，第一设备从使用第二AI模型变更为使用第一AI模型，第一条件也可称为变更条件。
通过该方法，第二设备可指示第一设备至少一个条件和至少一个AI模型的对应关系，这样，第一设备可根据该对应关系选择使用的AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。
并且，在该方法中，第一设备可根据对应关系选择AI模型，这样，无需第二设备根据第一设备的测量信息等为第一设备选择AI模型，从而可降低第二设备获取第一设备的测量信息的开销。
可选的，第二设备可根据以下至少一项，确定至少一个条件和至少一个AI模型的对应关系：第一设备的能力、第一设备的应用场景、第一设备的信号质量、第二设备的需求。其中，第二设备的需求可包括第二设备中存在的AI模型以及目标应用场景。
第二设备可根据第一设备的能力、第一设备的应用场景和第二设备的需求中的一个或多个确定出至少一个AI模型。例如，当第一设备的缓存能力为50MB时，第二设备可确定至少一个AI模型中任一AI模型的大小小于50MB。又例如，当终端设备可支持格式为ONNX的AI模型时，第二设备可确定至少一个AI模型中任一AI模型的格式为ONNX。又例如，当第一设备的应用场景或目标应用场景包括波束管理增强时，第二设备可确定至少一个AI中包括用于进行波束管理增强的AI模型。再例如，至少一个AI模型包括第二设备中存在的AI模型。
第二设备可根据至少一个AI模型的性能、开销以及第一设备的信号质量中的一个或多个，确定与至少一个AI模型对应的条件，即确定至少一个条件和至少一个AI模型的对应关系。例如，当AI模型1所需的计算能力为0.5T FLPS，AI模型2所需的计算能力为0.3T FLPS时，第二设备可确定与AI模型1对应的条件包括第一设备的计算能力大于或等于0.5T FLPS(例如，第一设备的计算能力为0.4T FLPS)，与AI模型1对应的条件包括第一设备的计算能力大于或等于0.3T FLPS(例如，第一设备的计算能力为0.4T FLPS)。又例如，第一设备检测到的来自第二设备的信号强度范围为a～b dBm，则第二设备可确定与AI模型1对应的条件包括第一设备检测到的来自第二设备的信号强度大于c dBm，其中，c大于或等于a，且小于或等于b，a、b和c为正整数。
可选的，第一信息还可包含至少一个AI模型的配置信息。下面以第一AI模型为例进行说明。如表1所示，第一AI模型的配置信息可包括第一AI模型的模型ID和/或模型内容信息。例如，当第一设备中的AI模型的格式为在ONNX格式或3GPP定义的AI模型格式时，第一AI模型的配置信息可包括第一AI模型的具体内容以及第一模型的模型ID。其中，模型ID的具体内容可参考S201中对模型ID的说明；第一AI模型的具体内容可包括第一AI模型及其子模型。
可选的，第一AI模型的配置信息还包括第一AI模型的索引。例如，若第二设备获知第一设备本地存储有第一AI模型的模型内容信息时，第一AI模型的配置信息可包括模型ID和第一AI模型的索引。第一AI模型的索引用于指示第一AI模型。例如，第一AI模型的索引可与第一AI模型的模型ID对应。
第一AI模型的索引可包括以下至少之一：模型列表中的索引和/或模型子列表中的索引。模型列表中的索引可用于指示第一AI模型，模型子列表中的索引可用于指示第一AI模型的子版本。例如，如表2所示，索引1可用于指示模型ID为XXX的AI模型，索引1-1可用于指示模型ID为XXXX1的AI模型，索引1-2可用于指示模型ID为XXXX2的AI模型，其中，模型ID为XXXX1的AI模型和模型ID为XXXX1的AI模型为模型ID为XXX的AI模型的子版本。
表2
另外，不同应用场景对应的模型列表和模型子列表中的索引可单独编号，也可以连续编号。例如，波束管理增强对应的模型列表中的索引为1和2，CSI-RS反馈增强对应的模型列表中的索引为3和4。
可选的，第一信息还包括至少一个AI模型中每个AI模型的最短执行时间。例如，第一AI模型的最短执行时间可为5分钟。通过该方法，可避免第一设备在不同AI模型之间频繁切换，从而避免因频繁切换AI模型引起的性能抖动和下降。
可选的，第一信息还可用于指示至少一个退出条件和至少一个AI模型的对应关系。至少一个退出条件和至少一个AI模型可以是一对一的关系，也可以是多对一的关系。
第二条件可为至少一个退出条件中与第一AI模型对应的条件。下面以第二条件为例，对至少一个退出条件进行说明。可选的，第二条件可包括表3中的一个或多个条件。
表3
下面分别对第二条件可包括的每个条件进行说明。
第二信号质量条件包括：第一设备的信号质量属于第二信号质量范围。其中，第一设备的信号质量可参考第一信号质量条件中的说明，此处不再赘述。第二信号质量范围与第一信号质量范围不同。例如，若第一设备的信号强度低于第一信号质量阈值，和/或第一设备的最大吞吐量高于第一吞吐量阈值，则第一设备可停止使用AI模型1。还例如，若第一设备检测到的来自第二设备的信号的传输时延大于80ms，则第一设备可停止使用AI模型1。又例如，若第一设备检测到的来自第二设备的信号的误包率和/或第一设备的掉话率大于20％，则第一设备可停止使用AI模型1。
第二应用场景条件包括：第一设备的应用场景属于第二应用场景范围。第二应用场景范围与第一应用场景范围不同。例如，第一应用场景范围包括CSI-RS反馈增强，第二应用场景范围包括波束管理增强。CSI-RS反馈增强可对应于AI模型a1和AI模型a2，波束管理增强可对应于AI模型a3和AI模型a4。若第一AI模型为AI模型a1，第一设备的应用场景从CSI-RS反馈增强更改为波束管理增强，则第一设备可停止使用AI模型a1。
第三资源条件包括：第一设备的资源属于第三资源范围。可选的，第一设备为终端设备时，第二条件可包括第三资源条件。第三资源范围与第一资源范围不同。例如，当终端设备使用的BWP的中心频点为第二频点，终端设备使用的带宽为第二带宽时，终端设备可停止使用AI模型b1。
第四资源条件包括：第二设备的资源属于第四资源范围。可选的，第一设备为网络设备且第二设备为终端设备时，第二条件可包括第四资源条件。第四资源范围与第三资源范围可以相同，也可以不同。例如，当终端设备使用的BWP的中心频点为第二频点，终端设备使用的带宽为第二带宽时，网络设备可停止使用AI模型b1。
第三区域条件包括：第一设备所处的区域属于第三区域范围。可选的，第一设备为终端设备时，第二条件可包括第三区域条件。第三区域范围与第一区域范围不同。例如，若终端设备位于B市，则终端设备可停止使用与A市对应的AI模型c1。还例如，若终端设备接入到PLMN2，则终端设备可停止使用与PLMN1对应的AI模型c2。又例如，若终端设备接入的小区的标识为CGI2，则终端设备可停止使用与CGI1对应的AI模型c3。
第四区域条件包括：第二设备所处的区域属于第四区域范围。可选的，第一设备为网络设备且第二设备为终端设备时，第二条件可包括第四区域条件。第四区域范围和第三区域范围可以相同，也可以不同。例如，若终端设备位于B市，则网络设备可停止使用与A市对应的AI模型c1。还例如，若终端设备接入到PLMN2，则网络设备可停止使用与PLMN1对应的AI模型c2。又例如，若终端设备接入的小区的标识为CGI2，则网络设备可停止使用与CGI1对应的AI模型c3。
第二能力条件包括：第一设备的能力属于第二能力范围。第二能力范围与第一能力范围不同。例如，第一设备的能力可包括第一设备的剩余电量。当第一设备的剩余电量大于或等于第一电量阈值时，第一设备可停止使用AI模型d1。又例如，第一设备的能力包括第一设备的计算能力。当第一设备的计算速率大于或等于第一速率阈值时，第一设备可停止使用AI模型d2。
第二性能条件包括：第一AI模型的性能属于第二性能范围。第一AI模型的性能可包括第一AI模型的预测准确率和/或使用第一AI模型的预测结果进行网络优化后的网络性能。例如，当第一AI模型的预测准确率小于第二准确率阈值(例如，1分钟内，第一AI模型的预测准确率小于95％)时，第一设备可停止使用第一AI模型。还例如，若第一AI模型连续K次预测的结果都不准确，K为正整数，例如，在波束管理增强中，第一AI模型连续5次预测出的最优波束都不是实际最优波束，第一设备可停止使用第一AI模型。又例如，使用第一AI模型的预测结果进行网络优化后的网络性能未达到预期的性能目标时，第一设备可停止使用第一AI模型。示例性的，预期的性能目标为将第一设备的吞吐量提升至X，若使用第一AI模型的预测结果进行网络优化后第一设备的吞吐量小于X，则第一设备可停止使用第一AI模型。
通过该方法，第二设备可指示第一设备至少一个退出条件和至少一个AI模型的对应关系，这样，第一设备可合理退出AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。
可选的，应用场景可包含在至少一个条件和至少一个退出条件中，也可作为第一信息中独立的信息存在。例如，如表4所示。第一信息可包括：应用场景、模型列表、模型子列表、模型ID和内容信息、至少一个条件、至少一个退出条件、最短执行时间。
表4

其中，索引1可用于指示模型ID为XXX的AI模型，索引1-1可用于指示模型ID为XXXX1的AI模型，索引1-2可用于指示模型ID为XXXX2的AI模型，其中，模型ID为XXXX1的AI模型和模型ID为XXXX1的AI模型为模型ID为XXX的AI模型的子版本。索引2可用于指示模型ID为YYY的AI模型，索引2-1可用于指示模型ID为YYYY1的AI模型，索引2-2可用于指示模型ID为YYYY2的AI模型，其中，模型ID为YYYY1的AI模型和模型ID为YYYY1的AI模型为模型ID为YYY的AI模型的子版本。
可选的，表4中的至少一个条件可拆成两个信息，一个为至少一个AI模型的进入条件，另一个为至少一个AI模型的变更条件。
S203：当第一设备满足第一条件时，第一设备使用与第一条件对应的第一AI模型。
S203的具体内容可参考对第一条件和第一AI模型的对应关系的说明，此处不再赘述。
S204：当第一设备满足第二条件时，第一设备停止使用第一AI模型，第二条件为至少一个退出条件中与第一AI模型对应的条件。
其中，S204为可选的步骤。
S204的具体内容可参考对第二条件和第一AI模型的对应关系的说明，此处不再赘述。
S205：当第一设备使用的AI模型发生变动时，第一设备向第二设备发送第二信息。相应的，第二设备接收来自第一设备的第二信息。
其中，第二信息可用于指示第一设备使用的AI模型发生变动。第一设备使用的AI模型发生变动可包括以下至少一项：第一设备从不使用AI模型变为使用AI模型，第一设备从使用一个AI模型变为使用另一个AI模型，第一设备从使用AI模型变为不使用AI模型。
可选的，第二信息包括以下至少一项：
1、第一设备使用的AI模型发生变动的时间：例如，当第一设备在19:00从不使用AI模型变为使用AI模型时，第二信息包括19:00。
2、触发第一设备使用的AI模型发生变动的条件：例如，当第一设备因满足第一条件而使用第一AI模型时，第二信息包括第一条件。
3、第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，第一设备的网络性能和/或AI模型性能：例如，第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，第一设备的吞吐量、误包率和/或传输时延等信息。又例如，第一设备从使用AI模型1-1变为使用AI模型2-2，第一设备使用的AI模型发生变动前第一时长内AI模型1-1的预测准确性，和/或第一设备使用的AI模型发生变动后第二时长内AI模型2-2的预测准确性。其中，第一时长和第二时长可为预先设置的，也可为第一设备从第二设备获取的。
4、第一设备使用的AI模型发生变动前和/或变化后使用的AI模型的指示信息：该指示信息可为AI模型的索引。例如，第一设备从使用AI模型1-1变为使用AI模型2-2，第二信息包括AI模型1-1的索引和/或AI模型2-2的索引。
下面结合不同的场景，对S205进行说明。
场景1：第一设备为终端设备1，第二设备为网络设备1。
当终端设备1处于连接态时，终端设备1可在AI模型发生变动时，向网络设备1发送第二信息；终端设备也可在AI模型发生多次变动后，向网络设备1发送用于指示该多次变动的第二信息，其中，第二信息可通过列表方式指示AI模型发生多次变动的信息。
当终端设备1处于非连接态时，终端设备1可保存第二信息，例如，终端设备1可在网络自优化报告中以列表的形式保存第二信息。在终端设备1进入连接态之后，终端设备1可向网络设备1发送第二信息，例如，终端设备1向网络设备1发送包含第二信息的网络自优化报告。其中，终端设备1进入连接态时，可能与网络设备1连接，也可能与其他网络设备(例如网络设备2)连接。当终端设备1与网络设备1连接时，终端设备1可直接向网络设备1发送第二信息。当终端设备1与网络设备2连接时， 终端设备1可通过网络设备2向网络设备1发送第二信息。具体的，终端设备1可向网络设备2发送第二信息，然后，网络设备2向网络设备1发送第二信息。
场景2：第一设备为网络设备1，第二设备为终端设备1。
当终端设备1处于连接态时，网络设备1可在AI模型发生变动时，向终端设备1发送第二信息；网络设备也可在AI模型发生多次变动后，向终端设备1发送用于指示AI模型发生多次变动的第二信息，其中，第二信息可通过列表方式指示AI模型发生多次变动的信息。可选的，第二信息可承载在RRC消息中。
当终端设备1处于非连接态时，网络设备1可保存第二信息，例如，网络设备1可在终端设备1的上下文信息中以列表的形式保存第二信息；或者，网络设备1向核心网发送包含第二信息的终端设备1的上下文信息。在终端设备1进入连接态之后，终端设备1接入的网络设备可从网络设备1或核心网获取第二信息，并向终端设备1发送第二信息。
可选的，第二设备在接收到第二信息后，可根据第二信息优化第一设备中的AI模型的配置。例如，若第一设备从AI模型1切换到AI模型2后，第一设备的性能出现下降或并未提升(如第一设备的传输时延增大)，则第二设备可修改从AI模型1切换到AI模型2的条件，提高第一设备从AI模型1切换到AI模型2的难度。假设在接收到第二信息之前，AI模型1切换到AI模型2的条件a包括：第一设备检测到的来自第二设备的信号的传输时延大于80ms，则在接收到第二信息后，第二设备可将AI模型1切换到AI模型2的条件a修改为条件b，条件b包括：第一设备检测到的来自第二设备的信号的传输时延大于或等于40ms且小于或等于60ms。通过该方法，第二设备可根据第一设备反馈的第二信息合理设置条件与AI模型的对应关系，使得第一设备合理选择和使用AI模型，从而提高系统性能。
可选的，当第一设备为网络设备，第二设备为终端设备时，如图3所示，图2所示方法还可包括：
S206：第二设备从第一设备向第三设备切换时，第一设备向第三设备发送第一信息。
其中，第一设备可为终端设备的源网络设备，第三设备可为终端设备的目的网络设备。也就是说，当终端设备从源网络设备向目的网络设备切换时，源网络设备可向目的网络设备发送来自终端设备的第一信息。
其中，第一信息可承载在现有的消息中，也可以承载在新的消息中。
通过该方法，当终端设备从源网络设备向目的网络设备切换时，源网络设备可向目的网络设备发送来自终端设备的第一信息。这样，目的网络设备也可根据终端设备提供的条件和AI模型的对应关系来使用AI模型。
可选的，当第一设备为终端设备，第二设备为网络设备时，如图4所示，图2所示方法还包括：
S207：第一设备从第二设备向第四设备切换时，第二设备向第四设备发送第一信息。
其中，第二设备可为终端设备的源网络设备，第四设备可为终端设备的目的网络设备。也就是说，当终端设备从源网络设备向目的网络设备切换时，源网络设备可向目的网络设备发送第一信息。
其中，第一信息可承载在现有的消息中，也可以承载在新的消息中。
通过该方法，当终端设备从源网络设备向目的网络设备切换时，源网络设备可向目的网络设备发送第一信息。这样，目的网络设备也可获取源网络设备提供的条件和AI模型的对应关系，从而可与使用该对应关系进行操作的终端设备进行协作。
为解决上述问题，本申请实施例提供了一种通信方法，该方法可应用于图1所示的通信系统中。下面参阅图5所示的流程图，对该方法的流程进行具体说明。
S301：第二设备向第一设备发送第四信息。相应的，第一设备接收来自第二设备的第四信息。
可选的，第一设备为终端设备，第二设备为网络设备；或者，第一设备为网络设备，第二设备为终端设备。
其中，第四信息可包括至少一个AI模型的配置信息。至少一个AI模型的配置信息的具体内容可参考S202，此处不再赘述。
第四信息可承载在现有的消息中，也可承载在新的消息中。
S302：第二设备向第一设备发送第五信息。相应的，第一设备接收来自第二设备的第五信息。
其中，第五信息可包括第三AI模型的指示信息，第三AI模型为至少一个AI模型中的一个模型。第五信息用于指示第一设备使用第三AI模型或从其他AI模型切换为第三AI模型。
可选的，第三AI模型的指示信息可为第三AI模型的索引。该第三AI模型的索引的具体内容可参考S202中对第一AI模型的索引的说明，此处不再赘述。
例如，模型列表、模型子列表和模型ID的对应关系可如表4所示。当第二设备向第一设备发送索引1时，可指示第一设备在波束管理增强场景下使用模型ID为XXX的AI模型；当第二设备向第一设备发送索引1-1时，可指示第一设备在波束管理增强场景下使用模型ID为XXXX1的模型。
该方法通过AI模型的索引来指示AI模型，这样，即便第三方获取到索引，也不知道该索引对应的AI模型，从而可提高安全性。
本申请对发送第五信息的方式不作限定，例如，第五信息可承载在RRC消息、MAC控制元素(MAC control element，MAC CE)、下行控制信息(downlink control Information，DCI)或上行控制信息(uplink control information，UCI)中。
S303：第一设备使用第三AI模型。
S304：第二设备向第一设备发送第六信息。相应的，第一设备接收来自第二设备的第六信息。
其中，第六信息可包括第三AI模型的指示信息。第六信息可用于指示第一设备停止使用第三AI设备。第三AI模型的指示信息的具体内容可参考S302，此处不再赘述。
本申请对发送第六信息的方式不作限定，例如，第六信息可承载在RRC消息、MAC CE、DCI或UCI中。
S305：第一设备停止使用第三AI模型。
S304-S305为可选的步骤。
通过该方法，第二设备在向第一设备发送至少一个AI模型的配置信息后，通过第三AI模型的指示信息即可指示第一设备使用或停止使用第三AI模型，从而可降低指示第一设备使用或停止使用第三AI模型所需的开销和时间，进而能够提升设备之间的协作效率。
基于与图2至图5方法实施例相同的技术构思，本申请实施例通过图6提供了一种通信装置，可用于执行上述方法实施例中相关步骤的功能。所述功能可以通过硬件实现，也可以通过软件或者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模块。该通信装置的结构如图6所示，包括通信单元601和处理单元602。所述通信装置600可以应用于网络设备(例如，图1中的gNB)或终端设备，并可以实现以上本申请实施例以及实例提供的通信方法。下面对所述通信装置600中的各个单元的功能进行介绍。
所述通信单元601，用于接收和发送数据。所述通信单元601可以通过收发器实现，例如，移动通信模块。其中，移动通信模块可以包括至少一个天线、至少一个滤波器，开关，功率放大器，低噪声放大器(low noise amplifier，LNA)等。
所述处理单元602可用于支持所述通信装置600执行上述方法实施例中的处理动作。所述处理单元602可以是通过处理器实现。例如，所述处理器可以为中央处理单元(central processing unit，CPU)，还可以是其它通用处理器、数字信号处理器(digital signal processor，DSP)、专用集成电路(application specific integrated circuit，ASIC)、现场可编程门阵列(field programmable gate array，FPGA)或者其它可编程逻辑器件、晶体管逻辑器件，硬件部件或者其任意组合。通用处理器可以是微处理器，也可以是任何常规的处理器。
在一种实施方式中，所述通信装置600应用于图2或图3所示的本申请实施例中的第一设备。第一设备可为网络设备(例如，图1中的gNB)或终端设备。下面对该实施方式中的所述处理单元602的具体功能进行介绍。
所述处理单元602，用于通过所述通信单元601接收来自第二设备的第一信息，第一信息用于指示至少一个条件和至少一个AI模型的对应关系；当第一设备满足第一条件时，使用与第一条件对应的第一AI模型，第一条件为至少一个条件中的任一条件。
可选的，第一条件包括以下至少一项：
第一设备的信号质量满足第一信号质量条件；
第一设备的应用场景满足第一应用场景条件；
第一设备的资源满足第一资源条件；
第一设备服务的第二设备的资源满足第二资源条件；
第一设备所处的区域满足第一区域条件；
第一设备服务的第二设备所处的区域满足第二区域条件；
第一设备的能力满足第一能力条件；
第一设备当前使用的第二AI模型的性能满足第一性能条件。
可选的，第一信号质量条件包括：第一设备的信号质量属于第一信号质量范围；
第一应用场景条件包括：第一设备的应用场景属于第一应用场景范围；
第一资源条件包括：第一设备的资源属于第一资源范围；
第二资源条件包括：第二设备的资源属于第二资源范围；
第一区域条件包括：第一设备所处的区域属于第一区域范围；
第二区域条件包括：第二设备所处的区域属于第二区域范围；
第一能力条件包括：第一设备的能力属于第一能力范围；和/或
第一性能条件包括：第二AI模型的性能属于第一性能范围。
可选的，第一信息还用于指示至少一个退出条件和至少一个AI模型的对应关系，所述处理单元602具体用于：使用与第一条件对应的第一AI模型之后，当第一设备满足第二条件时，停止使用第一AI模型，第二条件为至少一个退出条件中与第一AI模型对应的条件。
可选的，第二条件包括以下至少一项：
第一设备的信号质量满足第二信号质量条件；
第一设备的应用场景满足第二应用场景条件；
第一设备的资源满足第三资源条件；
第一设备服务的第二设备的资源满足第四资源条件；
第一设备所处的区域满足第三区域条件；
第一设备服务的第二设备所处的区域满足第四区域条件；
第一设备的能力满足第二能力条件；
第一AI模型的性能满足第二性能条件。
可选的，第二信号质量条件包括：第一设备的信号质量属于第二信号质量范围；
第二应用场景条件包括：第一设备的应用场景属于第二应用场景范围；
第三资源条件包括：第一设备的资源属于第三资源范围；
第四资源条件包括：第二设备的资源属于第四资源范围；
第三区域条件包括：第一设备所处的区域属于第三区域范围；
第四区域条件包括：第二设备所处的区域属于第四区域范围；
第二能力条件包括：第一设备的能力属于第二能力范围；和/或
第二性能条件包括：第一AI模型的性能属于第二性能范围。
可选的，所述处理单元602具体用于：当第一设备使用的AI模型发生变动时，通过所述通信单元601向第二设备发送第二信息，第二信息用于指示第一设备使用的AI模型发生变动。
可选的，第二信息包括以下至少一项：
第一设备使用的AI模型发生变动的时间；
触发第一设备使用的AI模型发生变动的条件；
第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，第一设备的网络性能和/或AI模型性能；
第一设备使用的AI模型发生变动前和/或变化后，第一设备所使用的AI模型的指示信息。
可选的，第一设备为终端设备，第二设备为网络设备；或者
第一设备为网络设备，第二设备为终端设备。
可选的，当第一设备为网络设备时，所述处理单元602具体用于：第二设备从第一设备向第三设备切换时，通过所述通信单元601向第三设备发送第一信息。
可选的，所述处理单元602具体用于：接收来自第二设备的第一信息之前，通过所述通信单元601 向第二设备发送第三信息，第三信息用于指示第一设备的能力和/或第一设备的应用场景，第三信息用于确定第一信息。
在另一种实施方式中，所述通信装置600应用于图2或图4所示的本申请实施例中的第二设备。第二设备可为网络设备(例如，图1中的gNB)或终端设备。下面对该实施方式中的所述处理单元602的具体功能进行介绍。
所述处理单元602，用于获取第一信息，第一信息用于指示至少一个条件和至少一个AI模型的对应关系；通过所述通信单元601向第一设备发送第一信息。
可选的，至少一个条件中的任一条件包括以下至少一项：
第一设备的信号质量满足第一信号质量条件；
第一设备的应用场景满足第一应用场景条件；
第一设备的资源满足第一资源条件；
第一设备服务的第二设备的资源满足第二资源条件；
第一设备所处的区域满足第一区域条件；
第一设备服务的第二设备所处的区域满足第二区域条件；
第一设备的能力满足第一能力条件；
第一设备当前使用的第二AI模型的性能满足第一性能条件。
可选的，第一信号质量条件包括：第一设备的信号质量属于第一信号质量范围；
第一应用场景条件包括：第一设备的应用场景属于第一应用场景范围；
第一资源条件包括：第一设备的资源属于第一资源范围；
第二资源条件包括：第二设备的资源属于第二资源范围；
第一区域条件包括：第一设备所处的区域属于第一区域范围；
第二区域条件包括：第二设备所处的区域属于第二区域范围；
第一能力条件包括：第一设备的能力属于第一能力范围；和/或
第一性能条件包括：第二AI模型的性能属于第一性能范围。
可选的，第一信息还用于指示至少一个退出条件和至少一个AI模型的对应关系。
可选的，第二条件为至少一个退出条件中的一个条件，第二条件包括以下至少一项：
第一设备的信号质量满足第二信号质量条件；
第一设备的应用场景满足第二应用场景条件；
第一设备的资源满足第三资源条件；
第一设备服务的第二设备的资源满足第四资源条件；
第一设备所处的区域满足第三区域条件；
第一设备服务的第二设备所处的区域满足第四区域条件；
第一设备的能力满足第二能力条件；
第一AI模型的性能满足第二性能条件。
可选的，第二信号质量条件包括：第一设备的信号质量属于第二信号质量范围；
第二应用场景条件包括：第一设备的应用场景属于第二应用场景范围；
第三资源条件包括：第一设备的资源属于第三资源范围；
第四资源条件包括：第二设备的资源属于第四资源范围；
第三区域条件包括：第一设备所处的区域属于第三区域范围；
第四区域条件包括：第二设备所处的区域属于第四区域范围；
第二能力条件包括：第一设备的能力属于第二能力范围；和/或
第二性能条件包括：第一AI模型的性能属于第二性能范围。
可选的，所述处理单元602具体用于：通过所述通信单元601接收来自第一设备的第二信息，第二信息用于指示第一设备使用的AI模型发生变动。
可选的，第二信息包括以下至少一项：
第一设备使用的AI模型发生变动的时间；
触发第一设备使用的AI模型发生变动的条件；
第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，第一设备的网络性能和/或AI模型性能；
第一设备使用的AI模型发生变动前和/或变化后，第一设备所使用的AI模型的指示信息。
可选的，第一设备为终端设备，第二设备为网络设备；或者
第一设备为网络设备，第二设备为终端设备。
可选的，当第二设备为网络设备时，所述处理单元602具体用于：第一设备从第二设备向第四设备切换时，通过所述通信单元601向第四设备发送第一信息。
可选的，所述处理单元602具体用于：获取第一信息之前，通过所述通信单元601接收来自第一设备的第三信息，第三信息用于指示第一设备的能力和/或第一设备的应用场景，第三信息用于确定第一信息。
在又一种实施方式中，所述通信装置600应用于图5所示的本申请实施例中的第一设备。第一设备可为网络设备(例如，图1中的gNB)或终端设备。下面对该实施方式中的所述处理单元602的具体功能进行介绍。
所述处理单元602用于通过所述通信单元601接收来自第二设备的第四信息，第四信息包括至少一个人工智能AI模型的配置信息；通过所述通信单元601接收来自第二设备的第五信息，第五信息包括第三AI模型的指示信息，第三AI模型为至少一个AI模型中的一个模型；使用第三AI模型。
可选的，第三AI模型的指示信息为第三AI模型的索引。
在再一种实施方式中，所述通信装置600应用于图5所示的本申请实施例中的第二设备。第二设备可为网络设备(例如，图1中的gNB)或终端设备。下面对该实施方式中的所述处理单元602的具体功能进行介绍。
所述处理单元602用于通过所述通信单元601向第一设备发送第四信息，第四信息包括至少一个人工智能AI模型的配置信息；通过所述通信单元601向第一设备发送第五信息，第五信息包括第三AI模型的指示信息，第三AI模型为至少一个AI模型中的一个模型，第五信息用于指示第一设备使用第三AI模型。
可选的，第三AI模型的指示信息为第三AI模型的索引。
需要说明的是，本申请以上实施例中对模块的划分是示意性的，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是单独物理存在，也可以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)或处理器(processor)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。
基于相同的技术构思，本申请实施例通过图7所示提供了一种通信装置，可用于执行上述方法实施例中相关的步骤。所述通信装置可以应用于网络设备(例如，图1中的gNB)或终端设备，可以实现以上本申请实施例以及实例提供的通信方法，具有图6所示的通信装置的功能。参阅图7所示，所述通信装置700包括：通信模块701、处理器702以及存储器703。其中，所述通信模块701、所述处理器702以及所述存储器703之间相互连接。
可选的，所述通信模块701、所述处理器702以及所述存储器703之间通过总线704相互连接。所述总线704可以是外设部件互连标准(peripheral component interconnect，PCI)总线或扩展工业标准结构(extended industry standard architecture，EISA)总线等。所述总线可以分为地址总线、数据总线、控 制总线等。为便于表示，图7中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
所述通信模块701，用于接收和发送数据，实现与其他设备之间的通信交互。例如，所述通信模块701可以通过物理接口、通信模块、通信接口、输入输出接口实现。
所述处理器702可用于支持所述通信装置700执行上述方法实施例中的处理动作。当所述通信装置700用于实现上述方法实施例时，处理器702还可用于实现上述处理单元602的功能。所述处理器702可以是CPU，还可以是其它通用处理器、DSP、ASIC、FPGA或者其它可编程逻辑器件、晶体管逻辑器件，硬件部件或者其任意组合。通用处理器可以是微处理器，也可以是任何常规的处理器。
在一种实施方式中，所述通信装置700应用于图2或图3所示的本申请实施例中的第一设备。所述处理器702具体用于：通过所述通信模块701接收来自第二设备的第一信息，第一信息用于指示至少一个条件和至少一个AI模型的对应关系；当第一设备满足第一条件时，使用与第一条件对应的第一AI模型，第一条件为至少一个条件中的任一条件。
在另一种实施方式中，所述通信装置700应用于图2或图4所示的本申请实施例中的第二设备。所述处理器702具体用于：获取第一信息，第一信息用于指示至少一个条件和至少一个AI模型的对应关系；通过所述通信模块701向第一设备发送第一信息。
在又一种实施方式中，所述通信装置700应用于图5所示的本申请实施例中的第一设备。所述处理器702具体用于：通过所述通信模块701接收来自第二设备的第四信息，第四信息包括至少一个人工智能AI模型的配置信息；通过所述通信模块701接收来自第二设备的第五信息，第五信息包括第三AI模型的指示信息，第三AI模型为至少一个AI模型中的一个模型；使用第三AI模型。
在再一种实施方式中，所述通信装置700应用于图5所示的本申请实施例中的第二设备。所述处理器702具体用于：通过所述通信模块701向第一设备发送第四信息，第四信息包括至少一个人工智能AI模型的配置信息；通过所述通信模块701向第一设备发送第五信息，第五信息包括第三AI模型的指示信息，第三AI模型为至少一个AI模型中的一个模型，第五信息用于指示第一设备使用第三AI模型。
所述处理器702的具体功能可以参考以上本申请实施例以及实例提供的通信方法中的描述，以及图6所示本申请实施例中对所述通信装置600的具体功能描述，此处不再赘述。
所述存储器703，用于存放程序指令和数据等。具体地，程序指令可以包括程序代码，该程序代码包括计算机操作指令。存储器703可能包含RAM，也可能还包括非易失性存储器(non-volatile memory)，例如至少一个磁盘存储器。处理器702执行存储器703所存放的程序指令，并使用所述存储器703中存储的数据，实现上述功能，从而实现上述本申请实施例提供的通信方法。
可以理解，本申请图7中的存储器703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，本文描述的系统和方法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。
基于以上实施例，本申请实施例还提供了一种计算机程序，当所述计算机程序在计算机上运行时，使得所述计算机执行以上实施例提供的方法。
基于以上实施例，本申请实施例还提供了一种计算机可读存储介质，该计算机可读存储介质中存储有计算机程序，所述计算机程序被计算机执行时，使得计算机执行以上实施例提供的方法。
其中，存储介质可以是计算机能够存取的任何可用介质。以此为例但不限于：计算机可读介质可以包括RAM、ROM、EEPROM、CD-ROM或其他光盘存储、磁盘存储介质或者其他磁存储设备、或者能够用于携带或存储具有指令或数据结构形式的期望的程序代码并能够由计算机存取的任何其他介质。
基于以上实施例，本申请实施例还提供了一种芯片，所述芯片用于读取存储器中存储的计算机程序，实现以上实施例提供的方法。
基于以上实施例，本申请实施例提供了一种芯片系统，该芯片系统包括处理器，用于支持计算机装置实现以上实施例中各设备所涉及的功能。在一种可能的设计中，所述芯片系统还包括存储器，所述存储器用于保存该计算机装置必要的程序和数据。该芯片系统，可以由芯片构成，也可以包含芯片和其他分立器件。
综上所述，本申请实施例提供了一种通信方法及装置，在该方法中，第一设备可接收来自第二设备的第一信息。其中，第一信息可用于指示至少一个条件和至少一个AI模型的对应关系。当第一设备满足第一条件时，第一设备可使用与第一条件对应的第一AI模型，其中，第一条件为至少一个条件中的任一条件。通过该方法，第一设备可根据第二设备指示的至少一个条件和至少一个AI模型的对应关系，选择使用的AI模型，从而可提高第一设备确定和使用AI模型的效率，使得第一设备能够合理、有效地使用AI模型，进而能够提升设备之间的协作效率。
在本申请的各个实施例中，如果没有特殊说明以及逻辑冲突，不同的实施例之间的术语和/或描述具有一致性、且可以相互引用，不同的实施例中的技术特征根据其内在的逻辑关系可以组合形成新的实施例。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本申请是参照根据本申请的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器，使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中，使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品，该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上，使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理，从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
显然，本领域的技术人员可以对本申请进行各种改动和变型而不脱离本申请的范围。这样，倘若本申请的这些修改和变型属于本申请权利要求及其等同技术的范围之内，则本申请也意图包含这些改动和变型在内。

Claims (29)

  1. 一种通信方法，应用于第一设备，其特征在于，包括：
    接收来自第二设备的第一信息，所述第一信息用于指示至少一个条件和至少一个人工智能AI模型的对应关系；
    当所述第一设备满足第一条件时，使用与所述第一条件对应的第一AI模型，所述第一条件为所述至少一个条件中的任一条件。
  2. 如权利要求1所述的方法，其特征在于，所述第一条件包括以下至少一项：
    所述第一设备的信号质量满足第一信号质量条件；
    所述第一设备的应用场景满足第一应用场景条件；
    所述第一设备的资源满足第一资源条件；
    所述第一设备服务的所述第二设备的资源满足第二资源条件；
    所述第一设备所处的区域满足第一区域条件；
    所述第一设备服务的所述第二设备所处的区域满足第二区域条件；
    所述第一设备的能力满足第一能力条件；
    所述第一设备当前使用的第二AI模型的性能满足第一性能条件。
  3. 如权利要求2所述的方法，其特征在于，所述第一信号质量条件包括：所述第一设备的信号质量属于第一信号质量范围；
    所述第一应用场景条件包括：所述第一设备的应用场景属于第一应用场景范围；
    所述第一资源条件包括：所述第一设备的资源属于第一资源范围；
    所述第二资源条件包括：所述第二设备的资源属于第二资源范围；
    所述第一区域条件包括：所述第一设备所处的区域属于第一区域范围；
    所述第二区域条件包括：所述第二设备所处的区域属于第二区域范围；
    所述第一能力条件包括：所述第一设备的能力属于第一能力范围；和/或
    所述第一性能条件包括：所述第二AI模型的性能属于第一性能范围。
  4. 如权利要求1～3任一项所述的方法，其特征在于，所述第一信息还用于指示至少一个退出条件和所述至少一个AI模型的对应关系，使用与所述第一条件对应的第一AI模型之后，还包括：
    当所述第一设备满足第二条件时，停止使用所述第一AI模型，所述第二条件为所述至少一个退出条件中与所述第一AI模型对应的条件。
  5. 如权利要求4所述的方法，其特征在于，所述第二条件包括以下至少一项：
    所述第一设备的信号质量满足第二信号质量条件；
    所述第一设备的应用场景满足第二应用场景条件；
    所述第一设备的资源满足第三资源条件；
    所述第一设备服务的所述第二设备的资源满足第四资源条件；
    所述第一设备所处的区域满足第三区域条件；
    所述第一设备服务的所述第二设备所处的区域满足第四区域条件；
    所述第一设备的能力满足第二能力条件；
    所述第一AI模型的性能满足第二性能条件。
  6. 如权利要求5所述的方法，其特征在于，所述第二信号质量条件包括：所述第一设备的信号质量属于第二信号质量范围；
    所述第二应用场景条件包括：所述第一设备的应用场景属于第二应用场景范围；
    所述第三资源条件包括：所述第一设备的资源属于第三资源范围；
    所述第四资源条件包括：所述第二设备的资源属于第四资源范围；
    所述第三区域条件包括：所述第一设备所处的区域属于第三区域范围；
    所述第四区域条件包括：所述第二设备所处的区域属于第四区域范围；
    所述第二能力条件包括：所述第一设备的能力属于第二能力范围；和/或
    所述第二性能条件包括：所述第一AI模型的性能属于第二性能范围。
  7. 如权利要求1～6任一项所述的方法，其特征在于，还包括：
    当所述第一设备使用的AI模型发生变动时，向所述第二设备发送第二信息，所述第二信息用于指示所述第一设备使用的AI模型发生变动。
  8. 如权利要求7所述的方法，其特征在于，所述第二信息包括以下至少一项：
    所述第一设备使用的AI模型发生变动的时间；
    触发所述第一设备使用的AI模型发生变动的条件；
    所述第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，所述第一设备的网络性能和/或AI模型性能；
    所述第一设备使用的AI模型发生变动前和/或变化后，所述第一设备所使用的AI模型的指示信息。
  9. 如权利要求1～8任一项所述的方法，其特征在于，所述第一设备为终端设备，所述第二设备为网络设备；或者
    所述第一设备为网络设备，所述第二设备为终端设备。
  10. 如权利要求9所述的方法，其特征在于，当所述第一设备为网络设备时，还包括：
    所述第二设备从所述第一设备向第三设备切换时，向所述第三设备发送所述第一信息。
  11. 如权利要求1～10任一项所述的方法，其特征在于，接收来自第二设备的第一信息之前，还包括：
    向所述第二设备发送第三信息，所述第三信息用于指示所述第一设备的能力和/或所述第一设备的应用场景，所述第三信息用于确定所述第一信息。
  12. 一种通信方法，应用于第二设备，其特征在于，包括：
    获取第一信息，所述第一信息用于指示至少一个条件和至少一个人工智能AI模型的对应关系；
    向第一设备发送所述第一信息。
  13. 如权利要求12所述的方法，其特征在于，所述至少一个条件中的任一条件包括以下至少一项：
    所述第一设备的信号质量满足第一信号质量条件；
    所述第一设备的应用场景满足第一应用场景条件；
    所述第一设备的资源满足第一资源条件；
    所述第一设备服务的所述第二设备的资源满足第二资源条件；
    所述第一设备所处的区域满足第一区域条件；
    所述第一设备服务的所述第二设备所处的区域满足第二区域条件；
    所述第一设备的能力满足第一能力条件；
    所述第一设备当前使用的第二AI模型的性能满足第一性能条件。
  14. 如权利要求13所述的方法，其特征在于，所述第一信号质量条件包括：所述第一设备的信号质量属于第一信号质量范围；
    所述第一应用场景条件包括：所述第一设备的应用场景属于第一应用场景范围；
    所述第一资源条件包括：所述第一设备的资源属于第一资源范围；
    所述第二资源条件包括：所述第二设备的资源属于第二资源范围；
    所述第一区域条件包括：所述第一设备所处的区域属于第一区域范围；
    所述第二区域条件包括：所述第二设备所处的区域属于第二区域范围；
    所述第一能力条件包括：所述第一设备的能力属于第一能力范围；和/或
    所述第一性能条件包括：所述第二AI模型的性能属于第一性能范围。
  15. 如权利要求12～14任一项所述的方法，其特征在于，所述第一信息还用于指示至少一个退出条件和所述至少一个AI模型的对应关系。
  16. 如权利要求15所述的方法，其特征在于，第二条件为所述至少一个退出条件中的一个条件，所述第二条件包括以下至少一项：
    所述第一设备的信号质量满足第二信号质量条件；
    所述第一设备的应用场景满足第二应用场景条件；
    所述第一设备的资源满足第三资源条件；
    所述第一设备服务的所述第二设备的资源满足第四资源条件；
    所述第一设备所处的区域满足第三区域条件；
    所述第一设备服务的所述第二设备所处的区域满足第四区域条件；
    所述第一设备的能力满足第二能力条件；
    所述第一AI模型的性能满足第二性能条件。
  17. 如权利要求16所述的方法，其特征在于，所述第二信号质量条件包括：所述第一设备的信号质量属于第二信号质量范围；
    所述第二应用场景条件包括：所述第一设备的应用场景属于第二应用场景范围；
    所述第三资源条件包括：所述第一设备的资源属于第三资源范围；
    所述第四资源条件包括：所述第二设备的资源属于第四资源范围；
    所述第三区域条件包括：所述第一设备所处的区域属于第三区域范围；
    所述第四区域条件包括：所述第二设备所处的区域属于第四区域范围；
    所述第二能力条件包括：所述第一设备的能力属于第二能力范围；和/或
    所述第二性能条件包括：所述第一AI模型的性能属于第二性能范围。
  18. 如权利要求12～17任一项所述的方法，其特征在于，还包括：
    接收来自所述第一设备的第二信息，所述第二信息用于指示所述第一设备使用的AI模型发生变动。
  19. 如权利要求18所述的方法，其特征在于，所述第二信息包括以下至少一项：
    所述第一设备使用的AI模型发生变动的时间；
    触发所述第一设备使用的AI模型发生变动的条件；
    所述第一设备使用的AI模型发生变动前第一时长和/或变化后第二时长内，所述第一设备的网络性能和/或AI模型性能；
    所述第一设备使用的AI模型发生变动前和/或变化后，所述第一设备所使用的AI模型的指示信息。
  20. 如权利要求12～19任一项所述的方法，其特征在于，所述第一设备为终端设备，所述第二设备为网络设备；或者
    所述第一设备为网络设备，所述第二设备为终端设备。
  21. 如权利要求20所述的方法，其特征在于，当所述第二设备为网络设备时，还包括：
    所述第一设备从所述第二设备向第四设备切换时，向所述第四设备发送所述第一信息。
  22. 如权利要求12～21任一项所述的方法，其特征在于，获取第一信息之前，还包括：
    接收来自所述第一设备的第三信息，所述第三信息用于指示所述第一设备的能力和/或所述第一设备的应用场景，所述第三信息用于确定所述第一信息。
  23. 一种通信方法，应用于第一设备，其特征在于，包括：
    接收来自第二设备的第四信息，所述第四信息包括至少一个人工智能AI模型的配置信息；
    接收来自所述第二设备的第五信息，所述第五信息包括第三AI模型的指示信息，所述第三AI模型为所述至少一个AI模型中的一个模型；
    使用所述第三AI模型。
  24. 如权利要求23所述的方法，其特征在于，所述第三AI模型的指示信息为第三AI模型的索引。
  25. 一种通信方法，应用于第二设备，其特征在于，包括：
    向第一设备发送第四信息，所述第四信息包括至少一个人工智能AI模型的配置信息；
    向所述第一设备发送第五信息，所述第五信息包括第三AI模型的指示信息，所述第三AI模型为所述至少一个AI模型中的一个模型，所述第五信息用于指示所述第一设备使用所述第三AI模型。
  26. 如权利要求25所述的方法，其特征在于，所述第三AI模型的指示信息为第三AI模型的索引。
  27. 一种通信装置，其特征在于，包括：
    通信单元，用于接收和发送数据；
    处理单元，用于通过所述通信单元，执行如权利要求1～26任一项所述的方法。
  28. 一种计算机可读存储介质，其特征在于，所述计算机可读存储介质中存储有计算机程序，当所述计算机程序在计算机上运行时，使得所述计算机执行如权利要求1～26任一项所述的方法。
  29. 一种芯片，其特征在于，所述芯片与存储器耦合，所述芯片读取所述存储器中存储的计算机程序，执行如权利要求1～26任一项所述的方法。
PCT/CN2023/126265 2022-11-04 2023-10-24 一种通信方法及装置 WO2024093739A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211379923.3 2022-11-04
CN202211379923.3A CN117997769A (zh) 2022-11-04 2022-11-04 一种通信方法及装置

Publications (1)

Publication Number Publication Date
WO2024093739A1 true WO2024093739A1 (zh) 2024-05-10

Family

ID=90898016

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/126265 WO2024093739A1 (zh) 2022-11-04 2023-10-24 一种通信方法及装置

Country Status (2)

Country Link
CN (1) CN117997769A (zh)
WO (1) WO2024093739A1 (zh)

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114143799A (zh) * 2020-09-03 2022-03-04 华为技术有限公司 通信方法及装置
WO2022206411A1 (zh) * 2021-04-02 2022-10-06 华为技术有限公司 一种小区切换方法及装置
US20220342713A1 (en) * 2020-01-14 2022-10-27 Guangdong Oppo Mobile Telecommunications Corp., Ltd. Information reporting method, apparatus and device, and storage medium

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220342713A1 (en) * 2020-01-14 2022-10-27 Guangdong Oppo Mobile Telecommunications Corp., Ltd. Information reporting method, apparatus and device, and storage medium
CN114143799A (zh) * 2020-09-03 2022-03-04 华为技术有限公司 通信方法及装置
WO2022206411A1 (zh) * 2021-04-02 2022-10-06 华为技术有限公司 一种小区切换方法及装置

Also Published As

Publication number Publication date
CN117997769A (zh) 2024-05-07

Similar Documents

Publication Publication Date Title
WO2021243619A1 (zh) 信息传输方法、装置、通信设备及存储介质
US9681464B2 (en) Cooperative transmission within heterogeneous stations
CN112911691B (zh) 一种小区节能方法、设备及存储介质
US20220394570A1 (en) Message sending method and apparatus, message receiving method and apparatus, and device and storage medium
US20210234648A1 (en) Method and apparatus for distribution and synchronization of radio resource assignments in a wireless communication system
WO2022041285A1 (zh) 一种模型数据传输方法及通信装置
WO2021092937A1 (zh) 一种策略确定的方法、系统及装置
WO2022226713A1 (zh) 策略确定的方法和装置
US20230189057A1 (en) Service traffic steering method and apparatus
US20230376358A1 (en) Method and apparatus for managing load of network node
CN111491340B (zh) 一种通信系统切换方法、网络侧设备、装置和介质
US20240056945A1 (en) Systems and methods for machine learning based access class barring
CN111245878B (zh) 一种基于混合云计算和雾计算的计算卸载通信网络的方法
Wang et al. A new cloud-based network framework for 5G massive Internet of Things connections
WO2024093739A1 (zh) 一种通信方法及装置
CN108810977B (zh) 一种通信方法、通信设备及具有存储功能的设备
Gao et al. A load balancing scheme for supporting safety applications in heterogeneous software defined LTE-V networks
US20240163741A1 (en) Ran node, ue, and method
WO2022165721A1 (zh) Ran域的模型共享方法及装置
WO2021114192A1 (zh) 一种网络参数调整方法及网络管理设备
CN111432458B (zh) 一种基于双连接的通信方法及装置
WO2016201860A1 (zh) 数据传输方法及装置
CN106231656B (zh) Rrc内高带宽许可请求技术
WO2023011655A1 (zh) 一种通信方法及装置
WO2023237107A1 (zh) 寻呼方法与装置、终端设备、网络设备和芯片