WO2024146303A2 - 一种通信方法、网络设备和终端设备 - Google Patents
一种通信方法、网络设备和终端设备 Download PDFInfo
- Publication number
- WO2024146303A2 WO2024146303A2 PCT/CN2023/134580 CN2023134580W WO2024146303A2 WO 2024146303 A2 WO2024146303 A2 WO 2024146303A2 CN 2023134580 W CN2023134580 W CN 2023134580W WO 2024146303 A2 WO2024146303 A2 WO 2024146303A2
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- model
- report
- terminal device
- moment
- resources
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/10—Scheduling measurement reports ; Arrangements for measurement reports
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- 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
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/10—Interfaces, programming languages or software development kits, e.g. for simulating neural networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L43/00—Arrangements for monitoring or testing data switching networks
- H04L43/06—Generation of reports
- H04L43/062—Generation of reports related to network traffic
-
- 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
Definitions
- the embodiments of the present application relate to the field of communications, and specifically to a communication method, a network device, and a terminal device.
- the communication method includes: a terminal device determines the number of AI resources required to simultaneously run a first artificial intelligence AI model and a second AI model at a first moment, the first AI model is used to determine a first report, and the second AI model is used to determine a second report; the terminal device determines that the first report is reported at a second moment based on the number of AI resources available at the first moment and the required number of AI resources; the terminal device reports the first report at the second moment, and reports the second report at a third moment after the second moment; or, the terminal device reports the first report at the second moment, and does not report the second report at the second moment; or, the terminal device reports the first report and the third report at the second moment, the third report is determined based on a non-AI model, and the third report is related to the second report.
- the network device can indicate the priorities of different AI models through the first indication information, so that the network device can To control the terminal device to prioritize the execution of the AI model that the network device needs to execute.
- the terminal device can report the number of AI resources required for different AI models and the number of AI resources required for AI model combinations to the network device through the first information, so that the network device and the terminal device can reach a consensus on the usage of the number of AI resources required for AI models and AI model combinations.
- the terminal device can dynamically adjust the total number of AI resources and send the adjusted total number of AI resources to the network device through the third information, so that the terminal device can reduce the total number of available AI resources.
- the network device when the network device receives the first report and the third report from the terminal device at a fourth moment, the network device determines that the terminal device cannot run the second AI model at the first moment to obtain the second report, including: the network device determines that the terminal device determines the third report based on the non-AI model at the first moment.
- the first AI model includes one or more AI models
- the second AI model includes one or more AI models
- the multiple AI models correspond to at least one AI function
- the method also includes: the terminal device sends fourth information to the network device, and the fourth information is used to report the number of AI resources required for each AI function in the at least one AI function.
- the method further includes: the network device receives fifth information from the terminal device, and the fifth information is used to report the updated number of AI resources required for the first AI function among the at least one AI function.
- the method further includes: the network device sending the multiple AI models to the terminal device.
- a terminal device is provided.
- the terminal device is used to execute the first aspect and any one of its implementation modes.
- the terminal device includes a processor and a memory, the memory is used to store a computer program; the processor is used to call and run the computer program from the memory, so that the first relay device executes the first aspect and any one of its implementation modes.
- a network device is provided.
- the network device is used to execute the second aspect and any one of the implementation modes thereof.
- the network device includes a processor and a memory, the memory is used to store a computer program; the processor is used to call and execute the computer program from the memory; A computer program enables the network device to execute the above-mentioned second aspect and any one of its implementation modes.
- a communication device is provided.
- the communication device is used to execute the method provided in the first aspect and the second aspect and any one of the embodiments thereof.
- the communication device may include a unit and/or module (e.g., a processing unit, a transceiver unit) for executing the method provided in the first aspect and the second aspect and any one of the embodiments thereof.
- the present application provides a processor for executing the methods provided in the first and second aspects above.
- a computer-readable storage medium stores a computer program, and when the computer program is executed on a communication device, the communication device executes a method in any one of the implementation modes of the first and second aspects.
- a chip comprising a processor and a communication interface, the processor reads instructions through the communication interface and executes the method provided by any one of the implementation modes of the first and second aspects.
- the chip also includes a memory, the memory stores a computer program or instructions, and the processor is used to execute the computer program or instructions stored in the memory.
- the processor is used to execute the method provided by any implementation method of the first and second aspects above.
- FIG1 is a schematic diagram of a communication system to which the present application is applicable.
- FIG2 is a schematic flow chart of a communication method provided in the present application.
- FIG4 is a schematic diagram of a resource conflict of an AI model provided in an embodiment of the present application.
- FIG5 is a schematic diagram of reporting the number of AI resources provided in an embodiment of the present application.
- FIG6 is a schematic flow chart of another communication method provided by the present application.
- FIG. 7 is a schematic flow chart of yet another communication method provided in the present application.
- FIG8 is a schematic flowchart of yet another communication method provided in the present application.
- FIG9 is a schematic flow chart of yet another communication method provided in the present application.
- FIG. 10 is a schematic block diagram of a communication device provided in an embodiment of the present application.
- FIG. 11 is a schematic diagram of another communication device provided in an embodiment of the present application.
- FIG. 12 is a schematic diagram of a chip system provided in an embodiment of the present application.
- the technical solutions of the embodiments of the present application can be applied to various communication systems.
- the fifth generation (5th generation, 5G) system or new radio (new radio, NR) long term evolution (long term evolution, LTE) system, LTE frequency division duplex (frequency division duplex, FDD) system, LTE time division duplex (time division duplex, TDD), etc.
- the technical solutions provided in the present application can also be applied to future communication systems, such as the sixth generation mobile communication system.
- the technical solutions of the embodiments of the present application can also be applied to device to device (D2D) communication, vehicle-to-everything (V2X) communication, machine to machine (M2M) communication, machine type communication (MTC), and Internet of Things (IoT) communication system or other communication systems.
- D2D device to device
- V2X vehicle-to-everything
- M2M machine to machine
- MTC machine type communication
- IoT Internet of Things
- the terminal equipment in the embodiments of the present application may refer to an access terminal, a user unit, a user station, a mobile station, a mobile station, a relay station, a remote station, a remote terminal, a mobile device, a user terminal, a user equipment (UE), a terminal, a wireless communication device, a user agent or a user device.
- the terminal equipment may also be a cellular phone, a cordless phone, a session initiation protocol (SIP) phone, a wireless local loop (WLL) station, a personal digital assistant (PDA), a handheld device with wireless communication function, a computing device or other processing device connected to a wireless modem, a vehicle-mounted device, a wearable device, a terminal device in a 5G network or a terminal device in a future public land mobile network (PLMN) or a terminal device in a future Internet of Vehicles, etc., and the embodiments of the present application are not limited to this.
- SIP session initiation protocol
- WLL wireless local loop
- PDA personal digital assistant
- PDA personal digital assistant
- wearable devices may also be referred to as wearable smart devices, which are a general term for wearable devices that are intelligently designed and developed using wearable technology for daily wear.
- wearable devices are portable devices that are worn directly on the body or integrated into the user's clothes or accessories.
- Wearable devices are not only hardware devices, but also powerful functions achieved through software support, data interaction, and cloud interaction.
- wearable smart devices include full-featured devices that can achieve complete or partial functions without relying on smartphones.
- smart watches or smart glasses it can also be a portable device that only focuses on a certain type of application function and needs to be used in conjunction with other devices such as smartphones.
- smart bracelets and smart jewelry for vital sign monitoring.
- the terminal device may also be a terminal device in an IoT system.
- IoT is an important part of the future development of information technology. Its main technical feature is to connect objects to the network through communication technology, thereby realizing an intelligent network of human-machine interconnection and object-to-object interconnection.
- IoT technology can achieve massive connections, deep coverage, and terminal power saving through, for example, narrowband (NB) technology.
- NB narrowband
- the terminal device may also include a sensor, whose main functions include collecting data (part of the terminal device), receiving control information and downlink data from the network device, and sending electromagnetic waves to transmit uplink data to the network device.
- the network device in the embodiment of the present application can be any communication device with wireless transceiver function for communicating with the terminal device.
- the device includes but is not limited to: evolved Node B (eNB), radio network controller (RNC), Node B (NB), home evolved Node B (HeNB, or home Node B, HNB), baseband unit (BBU), access point (AP) in wireless fidelity (WIFI) system, wireless relay node, wireless backhaul node, transmission point (TP) or transmission and reception point (TRP), etc.
- eNB evolved Node B
- RNC radio network controller
- NB Node B
- HeNB home evolved Node B
- BBU baseband unit
- AP access point
- WIFI wireless fidelity
- WIFI wireless relay node
- TP transmission point
- TRP transmission and reception point
- It can also be a 5G system, such as gNB in NR system, or transmission point (TRP or TP), one or a group of (including multiple antenna panels) antenna panels of a base station in 5G system, or it can also be a network node constituting gNB or transmission point, such as baseband unit (BBU), or distributed unit (DU), etc.
- gNB in NR system
- TRP or TP transmission point
- BBU baseband unit
- DU distributed unit
- the network equipment and terminal equipment can be deployed on land, including indoors or outdoors, handheld or vehicle-mounted; they can also be deployed on the water surface; they can also be deployed on aircraft, balloons and satellites in the air.
- the embodiments of the present application do not limit the scenarios in which the network equipment and terminal equipment are located.
- the terminal device or network device includes a hardware layer, an operating system layer running on the hardware layer, and an application layer running on the operating system layer.
- the hardware layer includes hardware such as a central processing unit, a memory management unit (MMU), and a memory (also called main memory).
- MMU memory management unit
- the operating system can be any one or more computer operating systems that implement business processing through processes, for example, operating system, operating system, operating system, Operating system or Operating system, etc.
- the application layer includes applications such as browsers, address books, word processing software, and instant messaging software.
- various aspects or features of the present application can be implemented as methods, devices or products using standard programming and/or engineering techniques.
- product used in this application covers computer programs that can be accessed from any computer-readable device, carrier or medium.
- computer-readable media include, but are not limited to: magnetic storage devices (e.g., hard disks, floppy disks or tapes, etc.), optical disks (e.g., compact discs (CDs), digital versatile discs (DVDs), etc.), smart cards and flash memory devices (e.g., erasable programmable read-only memory (EPROM), cards, sticks or key drives, etc.).
- the various storage media described herein may represent one or more devices and/or other machine-readable media for storing information.
- machine-readable storage medium may include, but is not limited to, wireless channels and various other media capable of storing, containing and/or carrying instructions and/or data.
- the communication system 100 may include at least one network device 101 and at least one terminal device 102 to 107.
- the terminal devices 102 to 107 may be mobile or fixed.
- the network device 101 and one or more of the terminal devices 102 to 107 may communicate via a wireless link.
- Each network device may provide communication coverage for a specific geographic area and may communicate with other devices located within the coverage area. terminal equipment to communicate.
- the terminal devices may communicate directly with each other.
- direct communication between the terminal devices may be achieved using device to device (D2D) technology.
- D2D device to device
- the terminal devices 105 and 106 and the terminal devices 105 and 107 may communicate directly using D2D technology.
- the terminal devices 106 and 107 may communicate with the terminal device 105 individually or simultaneously.
- Terminal devices 105 to 107 may also communicate with network device 101 respectively. For example, they may communicate directly with network device 101, such as terminal devices 105 and 106 in the figure may communicate directly with network device 101. They may also communicate indirectly with network device 101, such as terminal device 107 in FIG. 1 communicates with network device 101 via terminal device 105.
- Each communication device may be configured with multiple antennas.
- the multiple antennas configured may include at least one transmitting antenna for sending signals and at least one receiving antenna for receiving signals. Therefore, the communication devices in the communication system 100 may communicate with each other through multi-antenna technology.
- the interface between the network device and the terminal device can be a Uu interface (or air interface).
- Uu interface or air interface
- the names of these interfaces may remain unchanged, or may be replaced by other names, which is not limited in this application.
- the communication between the network device and the terminal device follows a certain protocol layer structure, and the network layering is to send, forward, package or unpack data of the network nodes (such as network devices and terminal devices), load or unpack control information, etc., which are completed by different hardware and software modules respectively. In this way, the complex problem of communication and network interconnection can be made simpler.
- FIG1 is only a simplified schematic diagram for ease of understanding, and the communication system 100 may also include other network devices or other terminal devices (not shown in FIG1).
- the communication system 100 may also include a core network device.
- the access network device provides a wireless access connection for the terminal device, and can send data to the terminal device or receive data sent by the terminal device; on the other hand, the access network device and the core network device are also connected, and the data received from the terminal device can be forwarded to the core network, or the data that needs to be sent to the terminal device can be received from the core network.
- AI model It can also be called AI algorithm (or AI operator), which is a general term for mathematical algorithms built on the principles of artificial intelligence, and is also the basis for using AI to solve specific problems.
- AI algorithm or AI operator
- the type of AI model is not limited in the embodiments of the present application.
- the AI model can be a machine learning model, a deep learning model, a reinforcement learning model, or a federated learning model.
- machine learning is a method to achieve artificial intelligence.
- the goal of this method is to design and analyze some algorithms (also known as models) that allow computers to "learn" automatically.
- the designed algorithms are called machine learning models.
- Machine learning models are a type of algorithm that automatically analyzes patterns from data and uses the patterns to predict unknown data.
- Deep learning is a new technical field that emerged in the process of machine learning research. Specifically, deep learning is a method in machine learning based on deep representation learning of data. Deep learning interprets data by establishing a neural network that simulates the human brain for analysis and learning. Since in machine learning methods, almost all features need to be determined by industry experts and then encoded. However, deep learning algorithms try to learn features from data by themselves. Algorithms designed based on deep learning ideas are called deep learning models.
- Reinforcement learning is a special field in machine learning. It is a process of continuously learning the optimal strategy, making sequential decisions, and obtaining the maximum reward through the interaction between the agent and the environment. In layman's terms, reinforcement learning is learning "what to do (i.e. how to map the current situation into actions) to maximize the numerical benefit signal". The agent will not be told what action to take, but must try to find out which actions will produce the most lucrative benefits. Reinforcement learning is different from supervised learning and unsupervised learning in the field of machine learning. Supervised learning is the process of learning from labeled training data provided externally (task-driven), and unsupervised learning is the process of finding implicit structures in unlabeled data (data-driven). Reinforcement learning is the process of finding a better solution through "trials”. The agent must develop existing experience to obtain benefits, and also conduct trials so that it can obtain a better action selection space in the future (i.e. learn from mistakes). The algorithm designed based on reinforcement learning is called a reinforcement learning model.
- Federated learning also known as collaborative learning
- collaborative learning is a machine learning technique that trains algorithms on multiple decentralized edge devices or servers holding local data samples without exchanging them. This approach is in stark contrast to traditional centralized machine learning techniques, where all local datasets are uploaded to a single server for training.
- Federated learning enables multiple participants to build a common, robust machine learning model without sharing data, allowing key issues such as data privacy, data security, data access rights, and access to heterogeneous data to be addressed.
- AI model training refers to the process of using a specified initial model to calculate the training data, and adjusting the parameters in the initial model using a certain method based on the calculation results, so that the model gradually learns certain rules and has specific functions.
- AI model reasoning is the process of using a trained AI model to calculate the input data and obtain the predicted reasoning results (also called output data).
- the AI module is a module with AI learning and computing capabilities.
- the AI module can be located in the operation administration and maintenance (OAM) network element, or in the RAN (e.g., the separation architecture is located in the CU), in some UEs, or as a separate network element entity.
- OAM operation administration and maintenance
- the main function of the AI module in a wireless communication system is to perform a series of AI calculations such as model building, training approximation, and reinforcement learning based on input data (e.g., in a wireless communication system, the input data can be network operation data provided by the RAN side or monitored by OAM, such as network load, channel quality, etc.).
- the trained model provided by the AI module has the function of predicting network changes on the RAN side, and can usually be used for load prediction, UE trajectory prediction, etc.
- the AI module can also perform policy reasoning from the perspectives of network energy saving and mobility optimization based on the prediction results of the trained model on the RAN network performance, so as to obtain reasonable and efficient energy saving strategies, mobility optimization strategies, etc.
- the AI module When the AI module is located in OAM, its communication with the gNB on the RAN side can reuse the current northbound interface; when the AI module is located in the gNB or CU, the current F1, Xn, Uu and other interfaces can be reused; when the AI module becomes an independent network entity, it is necessary to re-establish the communication link to the OAM and RAN side, such as based on a wired link or a wireless link.
- Model transfer The model is transferred from the network side to the UE side, or from the UE side to the network side, regardless of whether the transfer is implemented through air interface definition or non-air interface definition. If it is implemented through air interface definition, it can be called model transmission.
- Model activation Enable (disable) an AI model for a specific AI application function.
- Model switching For a specific AI application function, deactivate a currently activated model and activate another different model.
- an AI function may include functional modules, such as AI CSI feedback is an AI function, and AI positioning is another AI function. It may also be at the configuration level, such as an AI CSI configuration is an AI function. It can also be at the usage scenario level, such as a specific usage scenario is an AI function.
- Single-end model refers to model reasoning only on the network side or terminal side. Typical single-end model scenarios include AI beam management or AI positioning.
- Two-end model refers to the model reasoning that requires the joint participation of the terminal and the network side, that is, the models on the terminal side and the network side constitute a whole model. After the terminal side reasoning is completed, the reasoning result is sent to the network side for further reasoning, so as to obtain the final output of the overall model.
- Typical two-end model scenarios include AI CSI feedback.
- Batch size Indicates the number of data (samples) that are passed to the AI model for training or reasoning at one time. For example, if the training set has 1,000 data, and batch_size is set to 100, the AI model will first use the first 100 parameters in the data set, that is, the 1st to 100th data, to train the AI model. When the training is completed, the weights are updated, and then the 101st to 200th data are used for training, and the training stops after the 1,000 data in the training set are used for the tenth time.
- AI resources The AI resources involved in this application are virtual resources. Corresponding to physical resources, they can be storage resources in an AI processor, or computing resources of an AI processor, etc. Actual physical resources can be mapped to virtual resources through normalization and other methods.
- the total AI resources of the AI processor can be mapped to 0.1 for an agreed nominal value (such as 1 gigabyte (GB)).
- the maximum computing power of an AI processor is 1 Giga Floating-point Operations Per Second (GFLOPS).
- GFLOPS Giga Floating-point Operations Per Second
- the total resources of the AI processor can be mapped to 1.
- CSI processing rules CSI calculation in NR consumes a lot of computing resources, including the need to complete channel measurement and channel estimation for the configured measurement resources, and calculate CSI based on the estimated results.
- the concept of CSI processing unit CPU
- the CSI processing unit is used to represent the ability of the terminal device to process CSI.
- the CSI processing unit is a concept of virtual resources, not a specific physical resource concept. For example, a CPU is a computing core, and it can also be a computing thread, etc., which is not defined in the standard.
- CSI calculations include multiple categories, such as different CSI calculation amounts, or different configurations of the measured channels, such as different numbers of ports.
- the CPU will be different according to the different calculation contents, which is predefined for the protocol.
- the terminal device's capability information will carry a total number of CPUs and report it to the network device.
- the terminal device can calculate the current remaining CPU resources based on the number of currently occupied CPU resources L.
- the CPU needs to be occupied from a specific symbol to perform CSI calculation.
- the terminal device determines which CSI reports need to be updated based on the remaining CPU resources on the corresponding OFDM symbol.
- the CPU is occupied for calculation, or which CSI reports do not need to be updated, and the specific determination can be made according to the priority predefined in the protocol. For example, the terminal device determines the priority of the corresponding report according to the type of report (such as periodic or semi-static or non-periodic), the reporting content (such as reference signal received power (RSRP) or signal to interference plus noise ratio (SINR)), the cell, or the report ID.
- the type of report such as periodic or semi-static or non-periodic
- the reporting content such as reference signal received power (RSRP) or signal to interference plus noise ratio (SINR)
- RSRP reference signal received power
- SINR signal to interference plus noise ratio
- used for indication may include being used for direct indication and being used for indirect indication.
- indication information When describing that a certain indication information is used for indicating A, it may include that the indication information directly indicates A or indirectly indicates A, but it does not mean that A must be included in the indication information.
- the information indicated by the indication information is called the information to be indicated.
- the information to be indicated can be sent as a whole, or it can be divided into multiple sub-information and sent separately, and the sending period and/or sending time of these sub-information can be the same or different.
- the specific sending method is not limited in this application.
- the sending period and/or sending time of these sub-information can be pre-defined, for example, pre-defined according to the protocol, or it can be configured by the transmitting device by sending configuration information to the receiving device.
- the configuration information can be but is not limited to one or a combination of at least two of wireless resource control signaling, MAC layer signaling and physical layer signaling.
- wireless resource control signaling for example, includes RRC signaling
- MAC layer signaling for example, includes MAC CE
- physical layer signaling for example, includes DCI.
- the "storage” involved in the embodiments of the present application may refer to storage in one or more memories.
- the one or more memories may be separately set or integrated in an encoder or decoder, a processor, or a communication device.
- the one or more memories may also be partially separately set and partially integrated in a decoder, a processor, or a communication device.
- the type of memory may be any form of storage medium, which is not limited by the present application.
- the “protocol” involved in the embodiments of the present application may refer to a standard protocol in the communication field, for example, it may include an LTE protocol, an NR protocol, and related protocols used in future communication systems, and the present application does not limit this.
- RRC radio resource control
- the term "and/or" in this article is only a description of the association relationship of the associated objects, indicating that there can be three relationships.
- a and/or B can mean: A exists alone, A and B exist at the same time, and B exists alone.
- the character "/" in this article generally indicates that the associated objects before and after are in an "or" relationship.
- the training or reasoning of the AI model is a computationally intensive task, it can generally be implemented through devices with denser computing units, such as GPUs or NPUs, to achieve higher computing efficiency. Therefore, the UE-side AI algorithm used for communication in the future may also follow this model and be executed on AI-specific hardware.
- communication-related calculations may overlap in time, for example, for measurement resources, it is necessary to calculate the relevant reporting amount reported to the base station. Therefore, after the introduction of AI, the terminal may need to execute multiple AI models or AI functions at the same time.
- the AI model may be used to generate relevant reports (called Normal AI reasoning), for example, in CSI, based on the measurement quantities supported by the air interface, such as channels, the model is input and the corresponding AI CSI report is output.
- the reasoning of the AI model may also be used for model monitoring. Since the AI method is data-driven, the quality of the model generally depends on the quality of the collected data set. Therefore, compared with traditional algorithms, the performance of the AI algorithm may not be guaranteed when the scene or configuration changes, that is, the model performance may deteriorate due to substandard generalization performance. Therefore, AI model monitoring is very necessary.
- the network side or UE can get results and monitor the performance of the model by running the AI model periodically, semi-statically or non-periodically.
- the capabilities and resources of these devices on the terminal side may be limited (such as the number of AI accelerators, the number of accelerator cores, the memory size of the accelerator, etc.), when multiple AI models need to run simultaneously, it may not be possible to run all AI models. At this time, model conflicts occur, so it is necessary to solve the problem of which AI models are running and which AI models are not running among these conflicting models.
- the present application provides a communication method, which mainly defines the terminal device's operation strategy for multiple AI models when multiple AI models need to be run simultaneously on the terminal device side, but are limited by the insufficient capabilities of the terminal AI processor.
- the communication method provided in the embodiment of the present application can be applied to a system that communicates through a multi-antenna technology, for example, the communication system 100 shown in Figure 1.
- the communication system may include at least one network device and at least one terminal device.
- the network device and the terminal device can communicate through a multi-antenna technology.
- the embodiments shown below do not specifically limit the specific structure of the execution subject of the method provided in the embodiments of the present application, as long as it is possible to communicate according to the method provided in the embodiments of the present application by running a program that records the code of the method provided in the embodiments of the present application.
- the execution subject of the method provided in the embodiments of the present application may be a terminal device, or a functional module in the terminal device that can call and execute a program.
- FIG2 is a schematic flow chart of a communication method provided by the present application. The method comprises the following steps:
- the first AI model is used to determine the first report
- the second AI model is used to determine the second report.
- the reports involved in this embodiment include, but are not limited to: CSI measurement feedback report, positioning result report, beam management report, etc. It should be understood that the first report and the second report involved in this embodiment may be reports obtained by running the AI model, and examples are not given one by one here.
- the first report may be one or more reports, and the reports determined based on the first AI model may be referred to as the first report.
- the second report may be one or more reports, and the reports determined based on the second AI model may be referred to as the second report.
- the network device can determine the number of AI resources required for the terminal device to simultaneously run the first AI model and the second AI model at the first moment on the premise that: the network device knows the number of AI resources required for the terminal device to run the first AI model, and knows the number of AI resources required for the terminal device to run the second AI model. For example, the terminal device reports the number of AI resources required to run the first AI model and the number of AI resources required to run the second AI model to the network device, so that the network device can determine the number of AI resources required for the terminal device to simultaneously run the first AI model and the second AI model at the first moment.
- the terminal device may be triggered to run the first AI model and the second AI model simultaneously at the first moment through the following possible implementation methods:
- the network device triggers the terminal device non-periodically through trigger information to simultaneously run the first AI model and the second AI model at a first moment to obtain the first report and the second report, and report them.
- the network device sends trigger information to the terminal device at the first moment #1, and the trigger information is used to trigger the terminal device to measure the corresponding measurement resources, and report a first report and a second report, wherein the first report is determined based on the first AI model and the second report is determined based on the second AI model.
- the first moment is the starting point of the time slot where the trigger information is located, and the trigger information and the reference signal are in the same time slot, and the starting point may be the first symbol of the time slot.
- the terminal device determines the first report, which can be based on the corresponding measurement resources first, and then use the corresponding model based on the measurement results. Therefore, the first moment can refer to the moment where the measurement resource is located, such as the first symbol (or the starting time point of the first symbol) where the time domain resource position of the measurement resource is located, the first time slot (or the starting time point of the first time slot), the first subframe (or the starting time point of the first subframe), or the first radio frame (or the starting time point of the first radio frame), etc.
- FIG 3 is a schematic diagram of the first moment provided by an embodiment of the present application. It can be seen from Figure 3 that the first moment refers to the first symbol where the required AI resource is located.
- the trigger information carried by the physical downlink control channel (PDCCH) is sent in the first OFDM symbol.
- the first moment is the first OFDM symbol where the measurement resource (such as the channel state information-reference signal (CSIRS) shown in Figure 3) is located.
- CSIRS channel state information-reference signal
- the measurement resources can be the corresponding reference signals.
- the measurement resources are CSI-RS resources; for example, in beam management, the measurement resources are CSI-RS resources or synchronization signal block SSB resources; for example, in positioning, the measurement resources are positioning reference signal PRS resources, etc.
- the network device first configures two AI CSI reporting configurations for the terminal device in the CSI measurement configuration (e.g., CSI-MeasConfig), for example, configures two CSI-ReportConfigs, and configures the corresponding CSI report as an AI report in the CSI-ReportConfig, and the AI model used by the CSI report can also be indicated in the relevant report configuration.
- CSI-MeasConfig CSI-MeasConfig
- the two CSI report configurations and the corresponding measurement resources are associated therein (the measurement resources associated with the two reports may be the same or different), so that the network device side can dynamically trigger the aperiodic CSI-RS measurement and reporting corresponding to the two CSI reports at the first moment #1 through DCI indication.
- the first moment can be the symbol, time slot, subframe, or start time of the radio frame where the DCI or aperiodic CSI-RS is located.
- the measurement resources are also sent in the time slot where the first moment #1 is located, and the terminal device also starts measuring the corresponding CSI-RS resources in the time slot where the first moment #1 is located.
- the network device first configures two beam measurement sets or reference signal sets SetB#1 and SetB#2 for the terminal device in the beam management measurement configuration (such as Beam-MeasConfig or CSI-MeasConfig), wherein SetB#1 is used for beam prediction in the spatial domain, and is used for the terminal to use AI model #1 to predict the best beam in the spatial domain beam set SetA#1, and SetB#2 is used for beam prediction in the time domain, and is used for the terminal to use AI model #2 to predict the best beam in the time domain beam set SetA#2.
- the beam management measurement configuration such as Beam-MeasConfig or CSI-MeasConfig
- the two beam management report configurations and corresponding measurement resources are associated therein (the measurement configuration associated with one report is SetB#1, and the measurement configuration associated with the other report is SetB#2), so that the network device side can dynamically trigger the two non-periodic beam measurements and reports at the first moment #1 through DCI indication.
- the network device first configures the AI positioning measurement report for line-of-sight (LOS) or non-line-of-sight (NLOS) judgment and the AI positioning measurement report for obtaining the time of arrival (TOA) for the terminal device in the positioning-related configuration.
- the two measurement reports are obtained using AI model #1 and AI model #2 respectively, and the two AI positioning reports are associated with the same PRS measurement resource.
- the network device associates the two positioning reports in the same trigger state, so that the network device side can dynamically trigger the two non-periodic positioning measurements and reports at the first moment #1 through DCI indication.
- the network device configures the terminal device to periodically determine the first report and the second report, wherein the moments for periodically or semi-statically determining the first report and the second report include the above-mentioned first moment, so that the terminal device determines to simultaneously run the first AI model and the second AI model at the first moment.
- the first report can be a non-periodic report
- the second report can be a periodic report (or the first report is a periodic report, and the second report is a non-periodic report)
- the starting time of the measurement resources corresponding to the first report and the starting time of the measurement resources corresponding to the second report are both the first time.
- the first report and the second report can be any combination of an AI CSI measurement feedback report, an AI beam management report, or an AI positioning measurement report.
- the first report can be an AI CSI measurement feedback report
- the second report can be an AI beam management report.
- determining the number of AI resources required for the terminal device to simultaneously run the first AI model and the second AI model at the first moment can be understood as: determining the number of AI resources required for the first AI model and the second AI model to occupy each other starting from the first moment.
- the number of AI resources required for the AI model to be occupied may refer to the process from the terminal device measuring the corresponding measurement resources to the end of reporting the corresponding report, or the process from the terminal device measuring the corresponding measurement resources to obtaining the corresponding report, or the process from the terminal device inputting the measurement results into the model to obtaining the corresponding report, etc.
- This is not limited in this embodiment.
- there may be multiple definitions of the time period for running the AI model to occupy AI resources, which is not limited in this embodiment.
- the time period for running the AI model to occupy the number of AI resources is from the first symbol or the first time slot where the time domain resource location of the measurement resource is located, etc., until the end time point of the last symbol of the physical uplink shared channel (PUSCH) or physical uplink control channel (PUCCH) carrying the report.
- PUSCH physical uplink shared channel
- PUCCH physical uplink control channel
- the number of AI resources occupied by the first AI model starting at the first moment is number #1
- the number of AI resources occupied by the second AI model starting at the first moment is number #2
- the number of AI resources required for the terminal device to simultaneously run the first AI model and the second AI model at the first moment is number #1 + number #2.
- the number of AI resources required for multiple AI models simultaneously running on the terminal device at the first moment can be determined.
- the example of determining the number of AI resources required for the terminal device to simultaneously run the first AI model and the second AI model at the first moment is used for illustration.
- the number of AI resources required for the first AI model, the second AI model, and the third AI model simultaneously running on the terminal device at the first moment can also be determined, which is not repeated here.
- the terminal device after determining the number of AI resources required for the terminal device to simultaneously run the first AI model and the second AI model at the first moment, it can be determined whether the terminal device can obtain the first report and the second report based on the first AI model and the second AI model respectively at the first moment based on the number of AI resources available to the terminal device at the first moment and the required number of AI resources obtained by the aforementioned determination.
- the method flow shown in FIG. 2 also includes:
- the terminal device and the network device determine whether the terminal device can obtain the first report and the second report.
- This embodiment mainly considers the situation that the terminal device needs to run multiple AI models at the same time, but the terminal device AI processor capacity is insufficient. Take the example that the terminal device needs to run the first AI model and the second AI model at the same time, and the terminal device AI processor capacity cannot support the operation of the first AI model and the second AI model at the same time, such as the terminal device supports the operation of the first AI model at the first moment, but does not support the operation of the second AI model.
- FIG4 is used to illustrate the situation where the network device triggers the terminal device to report the first report and the second report, but the terminal device cannot support the operation of the first AI model and the second AI model at the same time.
- FIG4 is a schematic diagram of a resource conflict of an AI model provided in an embodiment of the present application.
- the total number of AI resources of the terminal device is 2 (such as AI resource #1' and AI resource #2' shown in FIG4), and the number of AI resources that each AI model needs to occupy is 1 (such as the occupied AI resource #1' or AI resource #2' shown in FIG4).
- the network device triggers the terminal device to non-periodically report CSI report #1 and CSI report #2 through DCI #1 at the first moment #1 (e.g., the AI resource required for determining CSI report #1 as shown in Figure 4 is AP-CSIRS #1, and the AI resource required for determining CSI report #2 is AP-CSIRS #2.
- the AP-report #1 shown in Figure 4 is CSI report #1, and AP-report #2 is CSI report #2), and triggers the terminal device to non-periodically report positioning report #1 through DCI #2 at the first moment #2 (e.g., the AI resource required for determining positioning report #1 as shown in Figure 4 is AP-PRS #1, and the AP-pos #1 shown in Figure 4 is positioning report #1).
- the terminal device determines CSI report #1 and CSI report #2 based on the AI model
- the number of AI resources available to the terminal device is 1 (because AI resource #1’ is occupied by the periodic CSI report, and the AI resource required for determining the periodic CSI report as shown in Figure 4 is P-CSIRS, and the P-report shown in Figure 4 is the periodic CSI report).
- AI resource #2’ is used to support the determination of CSI report #1, so the terminal device does not support the determination of CSI report #2 based on the AI model; when the terminal device determines the positioning report #1 based on the AI model, among the number of AI resources available to the terminal device, AI resource #1’ is released, and AI resource #2’ is occupied by the periodic CSI report.
- the terminal device can occupy AI resource #1’ and determine the positioning report #1 based on the AI model.
- the terminal device and the network device determine the number of AI resources available to the terminal device at the first moment and ... Determine the amount of AI resources required to run the first AI model and the second AI model, and determine that the terminal device can report the first report at the second moment, but cannot report the second report at the second moment.
- the terminal device can run the first AI model at the first moment to obtain a first report, and report it at the second moment; and the terminal device cannot run the second AI model at the first moment to obtain a second report.
- the terminal device can determine how to allocate the available AI resources based on the amount of AI resources required by different AI models.
- the first AI model is one AI model
- the second AI model is also one AI model.
- the first AI model is AI model #1
- the second AI model is AI model #2
- the number of AI resources required to run AI model #1 and AI model #2 at the first moment is AI resource number #1.
- the terminal device determines, based on the number of AI resources available at the first moment and the number of AI resources #1, that the first report is reported at the second moment, and determines that the second report cannot be reported at the second moment, including:
- the terminal device determines, based on the number of AI resources available at the first moment and the number of AI resources #1, that the number of AI resources available at the first moment is less than the number of AI resources #1;
- the terminal device determines that the number of AI resources #3 required by AI model #1 is less than the number of AI resources #2 required by AI model #2, and allocates the number of AI resources #3 from the number of AI resources available at the first moment to AI model #1;
- the terminal device determines that the first report corresponding to AI model #1 allocated to AI resource quantity #3 is reported at the second moment, and determines that the second report corresponding to AI model #2 not allocated to AI resource quantity #2 cannot be reported at the second moment.
- the terminal device can also allocate AI resources to more than two AI models based on the AI resources required by each of the two or more AI models.
- the first AI model is a plurality of AI models
- the second AI model is also a plurality of AI models.
- the first AI model includes AI model #1_1 and AI model #1_2, and AI model #1_1 and AI model #1_2 can be referred to as AI model combination #1;
- the second AI model includes AI model #2_1 and AI model #2_2, and AI model #2_1 and AI model #2_2 can be referred to as AI model combination #2.
- the terminal device determines, based on the number of AI resources available at the first moment and the number of AI resources #1, that the first report is reported at the second moment, and determines that the second report cannot be reported at the second moment, including:
- the terminal device determines, based on the number of AI resources available at the first moment and the number of AI resources #1, that the number of AI resources available at the first moment is less than the number of AI resources #1;
- the terminal device determines that the number of AI resources #1_1 required by the AI model combination #1 is less than the number of AI resources #2_1 required by the AI model #2, and allocates the number of AI resources #1_1 of the available AI resources at the first moment to the AI model combination #1;
- the terminal device determines that the first report corresponding to the AI model combination #1 allocated to the AI resource quantity #1_1 is reported at the second moment, and determines that the second report corresponding to the AI model combination #2 not allocated to the AI resource quantity #2_1 cannot be reported at the second moment.
- the terminal device can also allocate AI resources to more than two AI model combinations based on the AI resources required by each of the more than two AI model combinations.
- the first AI model may be a set including AI models and AI model combinations.
- the second AI model may be a set including AI models and AI model combinations.
- the terminal device may determine the number of AI resources required for the first AI model based on the AI models and AI model combinations included in the first AI model, and determine the number of AI resources required for the second AI model based on the AI models and AI model combinations included in the second AI model. After determining the number of AI resources required for the first AI model and the second AI model, the available number of AI resources are allocated according to the above-mentioned method, which will not be repeated here.
- the terminal device can determine how to allocate available AI resources based on the priorities of different AI models. Number of sources.
- the first AI model is one AI model
- the second AI model is also one AI model.
- the first AI model is AI model #1
- the second AI model is AI model #2
- the number of AI resources required to run AI model #1 and AI model #2 at the first moment is AI resource number #1.
- the terminal device determines, based on the number of AI resources available at the first moment and the number of AI resources #1, that the first report is reported at the second moment, and determines that the second report cannot be reported at the second moment, including:
- the terminal device determines, based on the number of AI resources available at the first moment and the number of AI resources #1, that the number of AI resources available at the first moment is less than the number of AI resources #1;
- the terminal device determines that the priority of AI model #1 is higher than the priority of AI model #2, and allocates AI resource number #3 of the AI resource number available at the first moment to AI model #1;
- the terminal device determines that the first report corresponding to AI model #1 allocated to AI resource quantity #3 is reported at the second moment, and determines that the second report corresponding to AI model #2 not allocated to AI resource quantity #2 cannot be reported at the second moment.
- the terminal device allocates AI resources to different AI models using the example of the terminal device allocating AI resources to AI model #1 and AI model #2 based on the priorities of AI model #1 and AI model #2 does not constitute any limitation on the scope of protection of the present application.
- the terminal device can also allocate AI resources to more than two AI models based on the priorities of more than two AI models.
- the terminal device determines that multiple AI models need to be run simultaneously at the first moment, and determines the AI models that prioritize the use of AI resources in turn according to the priorities of the multiple AI models, until the number of AI resources available at the first moment is fully occupied. For example, among the number of AI resources available at the first moment, the AI model corresponding to the priority P1 is run first to occupy the AI resources. If there are remaining AI resources after the AI model corresponding to P1 occupies the number of AI resources, the AI model with the priority P2 is run to occupy the AI resources, where the priority of P1 is higher than the priority of P2.
- the number of AI resources is allocated according to the number of AI resources required by each of them. For example, if the first AI model and the second AI model have the same priority, the number of AI resources is allocated to the first AI model and the second AI model according to the first AI resource number required by the first AI model and the second AI resource number required by the second AI model.
- the above-mentioned first AI model is a plurality of AI models
- the second AI model is also a plurality of AI models.
- the first AI model includes AI model #1_1 and AI model #1_2, and AI model #1_1 and AI model #1_2 can be referred to as AI model combination #1;
- the second AI model includes AI model #2_1 and AI model #2_2, and AI model #2_1 and AI model #2_2 can be referred to as AI model combination #2.
- the priority of the AI model combination can be defined in a predetermined manner, such as pre-defining the priority of the AI model combination as the priority corresponding to a higher or lower AI model.
- the priority of the AI model combination can also be defined directly, such as directly defining the priority of model combination #1 and model combination #2, without obtaining the priority of the model combination according to the priority of the AI models in the AI model combination.
- the terminal device can determine the priorities of different AI models or AI model combinations by itself; alternatively, the network device can indicate the priorities of different AI models or AI model combinations through first indication information.
- the network device can indicate the priorities of different AI models or AI model combinations through first indication information.
- the terminal device sends a first report and a second report to the network device.
- the terminal device and the network device determine that the terminal device cannot run the second AI model at the first moment to obtain the second AI model.
- the second report includes: the terminal device and the network device determine that the terminal device can run the second AI model after the first moment to obtain the second report, and report the second report at a third moment after the second moment. For example, the terminal device can run the second AI model to obtain the second report after some AI resources are released.
- Method 2 The terminal device reports the first report at the second moment, and does not report the second report at the second moment. This can be understood as an agreement between the terminal device and the network device that the report that cannot be obtained by running the AI model at the first moment can be not reported.
- the network device may receive the first report of the terminal device at the fourth moment (if the air interface delay is not considered, the fourth moment is the second moment mentioned above), and the second report is not received.
- the method flow shown in FIG. 2 also includes:
- the terminal device and the network device determine that the terminal device cannot run the second AI model at the first moment to obtain the second report, including: the terminal device and the network device determine that the terminal device does not report the second report at the second moment.
- Method 3 The terminal device reports the first report and the third report at the second moment, the third report is determined based on the non-AI model, and the third report is related to the second report. It can be understood that the terminal device and the network device agree that the report that cannot be obtained by running the AI model at the first moment can be determined and reported by the non-AI model.
- the method determined by the non-AI model may be to determine the report based on the legacy method, such as the PMI report determined by the codebook-based method of NR R15/R16/R17, wherein the legacy method may be to execute an algorithm on a general-purpose processor to determine the third report.
- the legacy method may be to execute an algorithm on a general-purpose processor to determine the third report.
- the traditional computing resources such as DSP resources
- the third report may be calculated and reported using the legacy method.
- the network device can associate the relevant configuration of the third report in the configuration related to the second report, such as the reporting information or reporting amount of the third report, the physical resources carrying the third report, etc.
- the terminal device uses the configuration of the associated third report to report the third report.
- the terminal device can send a request message #1 to the network device, where the request message #1 is used to request the network device to agree to determine and report based on the traditional method. If the network device agrees to determine and report based on the traditional method, the terminal device reports a third report.
- the terminal device carries the third report and the identifier of the third report in a pre-allocated reporting air interface channel, and the identifier of the third report indicates that the third report is a traditional report rather than a report determined based on the AI model.
- the third report and the identifier of the third report may be carried in RRC signaling.
- the network device may receive the first report and the third report of the terminal device at the fourth moment (if the air interface delay is not considered, the fourth moment is the second moment mentioned above), then in the case shown in the third mode, the method flow shown in FIG. 2 further includes:
- the above-mentioned methods 1 to 3 list several processing strategies implemented by the terminal device and the network device when the terminal device cannot start running the second AI model at the first moment to obtain the second report. It should be noted that this embodiment does not limit how the terminal device specifically handles the conflict situation in which the second report cannot be obtained based on the second AI model.
- the processing method for the conflict situation can be agreed upon in advance with the network device.
- the first AI model can also be understood as a first type of AI model.
- the second AI model can also be understood as a second type of AI model.
- the number of AI resources available on the terminal device cannot support the simultaneous operation of the first type of AI model and the second type of AI model.
- AI model for example, at a first moment, the AI resources available to the terminal device support the operation of a first type of AI model, then multiple AI models of the first type can start running at the same time at the first moment and occupy AI resources.
- the AI model on the terminal device side may be received from the network device side.
- the following will describe in detail how the network device sends the AI model to the terminal device in conjunction with Figure 6, which will not be described in detail here.
- the AI model on the terminal device side may also be determined by the terminal device itself.
- the AI model required to be used on the terminal device side (whether it is the terminal device side AI model in the dual-end model or only the single-end model) is developed separately for the terminal device or chip manufacturer. In this embodiment, there is no restriction on how the terminal device obtains the AI model.
- the terminal device sends first information to the network device, or the network device receives first information from the terminal device.
- the first information is used to indicate the number of AI resources required for each AI model in at least one AI model of the terminal device, and the number of AI resources required for different AI model combinations composed of the at least one AI model.
- the resources required for the AI model combination can be understood as the number of AI resources occupied when multiple AI models on the terminal device side are running at the same time. For example, the number of AI resources occupied when AI model #n and AI model #m on the terminal device side are running at the same time.
- Two AI models forming an AI model combination is just an example, and it may not be limited to two in actual implementation.
- the terminal device can report the number of AI resources required by the second AI model through the first information. If the second AI model includes multiple AI models #2, the first information is used to report the number of AI resources required for each AI model #2 in the multiple AI models #2, and the number of AI resources required for each AI model combination #2 in at least one AI model combination #2 composed of multiple AI models #2, and AI model combination #2 includes multiple AI models #2.
- the terminal device can consider the number of AI resources required by different AI models and allocate the number of AI resources.
- the number of AI resources required by different AI models or AI model combinations shown in Table 1 is combined with specific examples to illustrate how the terminal device allocates the number of AI resources.
- Example 1 The number of AI resources available to the terminal device at time #1 is 0.4. Based on the number of AI resources required for different AI models, the terminal device can determine that it supports running Model 1 but not Model 2 at time #1.
- Example 2 The number of AI resources available to the terminal device at time #1 is 0.6. Since Model 1 and Model 2 have been jointly optimized to ⁇ Model 1, Model 2 ⁇ , the number of AI resources required is 0.5. Therefore, Model 1 and Model 2 can be supported at the same time.
- Example 3 The terminal device can determine the number of AI resources actually required by the AI model based on the input sample size (i.e. batch_size) corresponding to different AI models. For example, if Model 1 uses two corresponding CSI reports, two samples can be input into Model 1 at one time, and the resource usage of Model 1 is 0.4.
- the process of the terminal device sending the first information to the network device can be understood as the terminal device capability reporting stage, that is, the reported content can be carried in the terminal device capability information (UECapabilityInformation), for example, the first information is sent to the network device in UECapabilityInformation.
- the terminal device capability reporting stage that is, the reported content can be carried in the terminal device capability information (UECapabilityInformation)
- UECapabilityInformation terminal device capability information
- the terminal device may also send the first information to the network device through other signaling.
- the terminal device sends the number of AI resources required by the local AI model to the network device through RRC signaling or MAC CE or UCI.
- ModelResourceInfo which carries the model ID or identifier (such as the above-mentioned ModelId field) and the number of AI resources corresponding to the AI model (such as the above-mentioned Resource field).
- the number of AI resources can be one of a set of discrete values.
- FIG5 is used to illustrate how to send the number of AI resources, which is a schematic diagram of reporting the number of AI resources provided in an embodiment of the present application.
- a 1-byte MAC-CE signaling is used to describe the AI model and the number of AI resources required for the AI model, where the number of AI resources can be represented by n1 (e.g., 6) bits, which are used to represent a value in a discrete value set (i.e., a possible set of AI resource quantities).
- the network device and the terminal device agree in advance on the corresponding UCI format, for example, first reporting the number of AI models that need to report resources, and then reporting the identification ID of each AI model and the number of AI resources required.
- the information in the UCI can be encoded independently or jointly. It can be carried in the PUCCH channel or the PUSCH channel.
- the network device and the terminal device may have a default value for the number of AI resources required for the AI model or AI model combination, such as a default value predetermined by the protocol, such as 0.1. If the corresponding resource occupancy value is not configured for a certain AI model or AI model combination in the above configuration, it is considered that the default value is used.
- the terminal device sends the second information to the network device, or the network device receives the second information from the terminal device.
- the second information is used to report the number of AI resources required after an update of any AI model among the multiple AI models and the number of AI resources required after an update of the AI model combination including the any AI model.
- the terminal device notifies the network device through the second information that the number of AI resources required for AI model #1 changes from O#1 to O#2 (for example, the second information indicates O#2 or the second information indicates the difference between O#1 and O#2, etc.).
- AI model #1 is updated, and the AI model combination containing AI model #1 includes AI model combination #1 and AI model combination #2, then the terminal device indicates the number of AI resources required by AI model combination #1 and AI model combination #2 through second information, and the number of AI resources required after the update.
- the number of AI resources required by AI model combination #1 is O#1′
- the number of AI resources required by the updated AI model combination #1 is O#2′.
- the structure of AI model #1 changes, or the structure of AI model #1 remains unchanged but only the parameters of AI model #1 change.
- the number of parameters may change, or the operator used by AI model #1 may change, which may cause the memory occupied by AI model #1 in the AI processor or the computing resources consumed when executing the model to change, thereby causing the number of AI resources required by AI model #1 to change; if the structure of AI model #1 remains unchanged but the parameters of AI model #1 change, a large number of parameters may be 0, which may also cause the memory occupied by AI model #1 to change, thereby causing the number of AI resources required by AI model #1 to change.
- the second information is related signaling for updating capability information of the terminal device.
- the implementation method of MAC-CE of the first information can be similar, and each AI model uses 1 byte of MAC-CE to indicate the required number of AI resources. Regardless of whether an update occurs, the MAC-CE corresponding to all AI models and AI model combinations will be sent.
- FIG7 is a schematic flow chart of another communication method provided by the present application, comprising the following steps:
- the fourth information is used to report the number of AI resources required by each AI function in the at least one AI function.
- Example 4 The number of AI resources available to the terminal device at time #1 is 0.4. Based on the number of AI resources required for different AI functions, the terminal device can determine to support AI CSI feedback but not AI beam management at time #1.
- the terminal device reports the updated number of AI resources required for the AI function to the network device.
- the network device can trigger the terminal device to perform an uplink report based on the AI function.
- an AI resource occupancy conflict occurs during the uplink report based on the AI function of the terminal device (for example, multiple AI functions need to be run simultaneously on the terminal device side, but the AI processor capacity of the terminal device is insufficient)
- the reporting process can refer to the communication method shown in Figure 2, which will not be repeated here; if no AI resource occupancy conflict occurs during the uplink report based on the AI function of the terminal device, the terminal device can obtain the corresponding report based on the AI function and report it.
- the terminal can dynamically update the number of AI resources required for its AI function and report it to the network device, so that the network device and the terminal device can reach a consensus on the use rules and usage of the AI function. And it can be combined with the communication method shown in FIG2 to determine which AI functions need to run first when multiple AI functions need to run at the same time, and can reach a consensus on the processing method of the report corresponding to the unusable AI function.
- FIG8 is a schematic flow chart of another communication method provided by the present application, comprising the following steps:
- the first indication information is used to indicate the priority of each AI model in at least one local AI model of the terminal device.
- the network device determines the priority of different AI models based on its own situation and the AI model usage of the terminal device, and sends the first indication information indicating the priority of different AI models to the terminal device.
- the use priority of the AI model can be lowered without shutting down or deactivating the AI model. If there are not sufficient AI resources, the AI model will not be used.
- the inconsistency can be handled in a predetermined manner, for example, the priority of the AI model combination can be predefined as the priority corresponding to the higher or lower AI model.
- the network device may send one or more AI models to the terminal device in any of the following ways:
- FIG10 is a schematic block diagram of a communication device 10 provided in an embodiment of the present application.
- the device 10 includes a transceiver module 11 and a processing module 12.
- the transceiver module 11 can implement corresponding communication functions, and the processing module 12 is used to perform data processing, or in other words, the transceiver module 11 is used to perform operations related to receiving and sending, and the processing module 12 is used to perform other operations besides receiving and sending.
- the transceiver module 11 can also be called a communication interface or a communication unit.
- the transceiver module 11 can be used to execute the steps of sending and receiving information in the method, such as steps S231, S232 and S233; the processing module 12 can be used to execute the processing steps in the method, such as steps S210 and S220.
- the transceiver module 11 may be used to execute the steps of sending and receiving information in the method, such as step S510 ; and the processing module 12 may be used to execute the processing steps in the method.
- module here may refer to an application specific integrated circuit (ASIC), an electronic circuit, a processor (such as a shared processor, a dedicated processor or a group processor, etc.) and a memory for executing one or more software or firmware programs, a merged logic circuit and/or other suitable components that support the described functions.
- ASIC application specific integrated circuit
- processor such as a shared processor, a dedicated processor or a group processor, etc.
- memory for executing one or more software or firmware programs, a merged logic circuit and/or other suitable components that support the described functions.
- the device 10 may be specifically a mobile management network element in the above-mentioned embodiment, and may be used to execute the various processes and/or steps corresponding to the mobile management network element in the above-mentioned method embodiments; or, the device 10 may be specifically a terminal device in the above-mentioned embodiment, and may be used to execute the various processes and/or steps corresponding to the terminal device in the above-mentioned method embodiments. To avoid repetition, it will not be repeated here.
- the device 20 further includes a transceiver 23, and the transceiver 23 is used for receiving and/or sending signals.
- the processor 21 is used to control the transceiver 23 to receive and/or send signals.
- the processor is a general-purpose processor, DSP, ASIC, FPGA or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, the memory (storage module) can be integrated into the processor.
- memory described herein is intended to include, but is not limited to, these and any other suitable types of memory.
- FIG12 is a schematic diagram of a chip system 30 provided in an embodiment of the present application.
- the chip system 30 (or also referred to as a processing system) includes a logic circuit 31 and an input/output interface 32.
- the logic circuit 31 can be a processing circuit in the chip system 30.
- the logic circuit 31 can be coupled to the storage unit and call the instructions in the storage unit so that the chip system 30 can implement the methods and functions of each embodiment of the present application.
- the input/output interface 32 can be an input/output circuit in the chip system 30, outputting information processed by the chip system 30, or inputting data or signaling information to be processed into the chip system 30 for processing.
- the chip system 30 is used to implement the operations performed by the terminal device in the above various method embodiments.
- the logic circuit 31 is used to implement the processing-related operations performed by the terminal device in the above method embodiment
- the input/output interface 32 is used to implement the sending and/or receiving-related operations performed by the terminal device in the above method embodiment.
- An embodiment of the present application also provides a computer-readable storage medium on which computer instructions for implementing the methods executed by the device in the above-mentioned method embodiments are stored.
- An embodiment of the present application also provides a computer program product, comprising instructions, which, when executed by a computer, implement the methods performed by a terminal device or a network device in the above-mentioned method embodiments.
- An embodiment of the present application also provides a communication system, including the aforementioned terminal device and network device.
- the disclosed devices and methods can be implemented in other ways.
- the device embodiments described above are only schematic.
- the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed.
- the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.
- the computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.
- the computer can be a personal computer, a server, or a network device, etc.
- the computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from a website site, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center.
- the computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media integrations.
- the available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a magnetic tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state disk (SSD)).
- the aforementioned available medium includes, but is not limited to, various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- General Physics & Mathematics (AREA)
- Biophysics (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Mobile Radio Communication Systems (AREA)
- Data Exchanges In Wide-Area Networks (AREA)
Abstract
Description
Claims (35)
- 一种通信方法,其特征在于,包括:终端设备确定在第一时刻同时运行第一人工智能AI模型和第二AI模型所需的AI资源数量,所述第一AI模型用于确定第一报告,所述第二AI模型用于确定第二报告;所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述第一报告在第二时刻上报;所述终端设备在所述第二时刻上报所述第一报告,以及在所述第二时刻之后的第三时刻上报所述第二报告;或者,所述终端设备在所述第二时刻上报所述第一报告,以及在所述第二时刻不上报所述第二报告;或者,所述终端设备在第二时刻上报所述第一报告和第三报告,所述第三报告基于非AI模型确定,且所述第三报告与所述第二报告相关。
- 根据权利要求1所述的方法,其特征在于,所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述第一报告在第二时刻上报,包括:所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述第一报告在所述第二时刻上报,以及确定在所述第二时刻无法上报所述第二报告。
- 根据权利要求2所述的方法,其特征在于,在所述终端设备在第二时刻上报所述第一报告,以及在所述第二时刻之后的第三时刻上报所述第二报告的情况下,所述终端设备确定在所述第二时刻无法上报所述第二报告,包括:所述终端设备确定在所述第二时刻之后的第三时刻上报所述第二报告。
- 根据权利要求2所述的方法,其特征在于,在所述终端设备在第二时刻上报所述第一报告,以及在所述第二时刻不上报所述第二报告的情况下,所述终端设备确定在所述第二时刻无法上报所述第二报告,包括:所述终端设备确定在所述第二时刻不上报所述第二报告。
- 根据权利要求2所述的方法,其特征在于,所述终端设备在第二时刻上报所述第一报告和所述第三报告的情况下,所述终端设备确定在所述第二时刻无法上报所述第二报告,包括:所述终端设备确定在所述第一时刻基于非AI模型确定所述第三报告,并在所述第二时刻上报所述第三报告。
- 根据权利要求2至5中任一项所述的方法,其特征在于,所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述第一报告在所述第二时刻上报,以及确定在所述第二时刻无法上报所述第二报告,包括:所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述第一时刻可用的AI资源数量少于所述所需的AI资源数量,所述所需的AI资源数量为所述第一AI模型所需的第一AI资源数量和所述第二AI模型所需的第二AI资源数量之和;所述终端设备确定所述第一AI资源数量少于所述第二AI资源数量,并将所述第一时刻可用的AI资源数量中的所述第一AI资源数量分配给所述第一AI模型;所述终端设备确定分配到所述第一AI资源数量的第一AI模型对应的第一报告在所述第二时刻上报,以及确定未分配到所述第二AI资源数量的第二AI模型对应的第二报告无法在所述第二时刻上报。
- 根据权利要求2至5中任一项所述的方法,其特征在于,所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述第一报告在所述第二时刻上报,以及确定在所述第二时刻无法上报所述第二报告,包括:所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述第一时刻可用的AI资源数量少于所述所需的AI资源数量,所述所需的AI资源数量为所述第一AI模型所需的第一AI资源数量和所述第二AI模型所需的第二AI资源数量之和;所述终端设备确定所述第一AI模型的优先级高于所述第二AI模型的优先级,并将所述第一时刻可用的AI资源数量中的所述第一AI资源数量分配给所述第一AI模型;所述终端设备确定分配到所述第一AI资源数量的第一AI模型对应的第一报告在所述第二时刻上报,以及确定未分配到所述第二AI资源数量的第二AI模型对应的第二报告无法在所述第二时刻上报。
- 根据权利要求7所述的方法,其特征在于,在所述终端设备确定所述第一AI模型的优先级高于所述第二AI模型的优先级之前,所述方法还包括:所述终端设备接收来自所述网络设备的第一指示信息,所述第一指示信息用于指示所述第一AI模型的优先级高于所述第二AI模型的优先级。
- 根据权利要求1至8中任一项所述的方法,其特征在于,所述第一AI模型包括一个或者多个AI模型,和/或,所述第二AI模型包括一个或者多个AI模型。
- 根据权利要求9所述的方法,其特征在于,在所述第一AI模型包括多个AI模型的情况下,所述方法还包括:所述终端设备向所述网络设备发送第一信息,所述第一信息用于上报所述多个AI模型中每个AI模型所需的AI资源数量,以及所述多个AI模型组成的至少一个AI模型组合中每个AI模型组合所需的AI资源数量,所述AI模型组合包括多个AI模型。
- 根据权利要求10所述的方法,其特征在于,所述方法还包括:所述终端设备向网络设备发送第二信息,所述第二信息用于上报所述多个AI模型中的任一AI模型更新后的所需AI资源数量和包括所述任一AI模型的AI模型组合更新后的所需AI资源数量。
- 根据权利要求10或11所述的方法,其特征在于,所述方法还包括:所述终端设备向所述网络设备发送第三信息,所述第三信息用于指示调整可用的AI资源总数。
- 根据权利要求10至12中任一项所述的方法,其特征在于,所述多个AI模型对应至少一个AI功能,所述方法还包括:所述终端设备向网络设备发送第四信息,所述第四信息用于上报所述至少一个AI功能中每个AI功能所需的AI资源数量。
- 根据权利要求13所述的方法,其特征在于,所述方法还包括:所述终端设备向网络设备发送第五信息,所述第五信息用于上报所述至少一个AI功能中的第一AI功能更新后的所需AI资源数量。
- 根据权利要求10至14中任一项所述的方法,其特征在于,在所述终端设备向所述网络设备发送第一信息之前,所述方法还包括:所述终端设备接收来自所述网络设备的所述多个AI模型;所述终端设备确定所述多个AI模型中每个AI模型所需的AI资源数量。
- 根据权利要求1至15中任一项所述的方法,其特征在于,所述AI资源包括:AI处理器中的存储资源和/或所述AI处理器的算力资源。
- 一种通信方法,其特征在于,包括:网络设备触发终端设备在第一时刻同时运行第一人工智能AI模型和第二AI模型,所述第一AI模型用于确定第一报告,所述第二AI模型用于确定第二报告;所述网络设备根据所述第一时刻所述终端设备可用的AI资源数量和在所述第一时刻同时运行所述第一AI模型和所述第二AI模型所需的AI资源数量,确定所述终端设备能够在所述第一时刻运行所述第一AI模型获得所述第一报告;所述网络设备在第四时刻接收来自所述终端设备的所述第一报告,以及在所述第四时刻之后的第五时刻接收来自所述终端设备的所述第二报告;或者,所述网络设备在第四时刻接收来自所述终端设备的所述第一报告,以及未接收到来自所述终端设备的所述第二报告;或者,所述网络设备在第四时刻接收来自所述终端设备的所述第一报告和第三报告,所述第三报告基于非AI模型确定,且所述第三报告与所述第二报告相关。
- 根据权利要求17所述的方法,其特征在于,所述网络设备根据所述第一时刻所述终端设备可用的AI资源数量和所述所需的AI资源数量,确定所述终端设备能够在所述第一时刻运行所述第一AI模型获得所述第一报告,包括:所述终端设备根据所述第一时刻可用的AI资源数量和所述所需的AI资源数量,确定所述终端设备能够在所述第一时刻运行所述第一AI模型获得所述第一报告,以及确定所述终端设备无法在所述第一时刻运行所述第二AI模型获得所述第二报告。
- 根据权利要求18所述的方法,其特征在于,在所述网络设备在第四时刻接收来自所述终端设备 的所述第一报告,以及在所述第四时刻之后的第五时刻接收来自所述终端设备的所述第二报告的情况下,所述网络设备确定所述终端设备无法在所述第一时刻运行所述第二AI模型获得所述第二报告,包括:所述网络设备确定所述终端设备在所述第一时刻之后运行所述第二AI模型获得所述第二报告。
- 根据权利要求18所述的方法,其特征在于,在所述网络设备在第四时刻接收来自所述终端设备的所述第一报告,以及未接收到来自所述终端设备的所述第二报告的情况下,所述网络设备确定所述终端设备无法在所述第一时刻运行所述第二AI模型获得所述第二报告,包括:所述网络设备确定所述终端设备不在所述第一时刻运行所述第二AI模型。
- 根据权利要求18所述的方法,其特征在于,在所述网络设备在第四时刻接收来自所述终端设备的所述第一报告和第三报告的情况下,所述网络设备确定所述终端设备无法在所述第一时刻运行所述第二AI模型获得所述第二报告,包括:所述网络设备确定所述终端设备在所述第一时刻基于非AI模型确定所述第三报告。
- 根据权利要求17至21中任一项所述的方法,其特征在于,所述方法还包括:所述网络设备向所述终端设备发送第一指示信息,所述第一指示信息用于指示所述第一AI模型的优先级高于所述第二AI模型的优先级。
- 根据权利要求17至22中任一项所述的方法,其特征在于,所述第一AI模型包括一个或者多个AI模型,所述第二AI模型包括一个或者多个AI模型。
- 根据权利要求23所述的方法,其特征在于,在所述第一AI模型包括多个AI模型的情况下,所述方法还包括:所述网络设备接收来自所述终端设备的第一信息,所述第一信息用于上报所述多个AI模型中每个AI模型所需的AI资源数量,以及所述多个AI模型组成的至少一个AI模型组合中每个AI模型组合所需的AI资源数量,所述AI模型组合包括多个AI模型。
- 根据权利要求24所述的方法,其特征在于,所述方法还包括:所述网络设备接收来自所述终端设备的第二信息,所述第二信息用于上报所述多个AI模型中的任一AI模型更新后的所需AI资源数量和包括所述任一AI模型的AI模型组合更新后的所需AI资源数量。
- 根据权利要求24或25所述的方法,其特征在于,所述方法还包括:所述网络设备接收来自所述终端设备的第三信息,所述第三信息用于指示调整可用的AI资源总数。
- 根据权利要求24至26中任一项所述的方法,其特征在于,所述多个AI模型对应至少一个AI功能,所述方法还包括:所述终端设备向网络设备发送第四信息,所述第四信息用于上报所述至少一个AI功能中每个AI功能所需的AI资源数量。
- 根据权利要求27所述的方法,其特征在于,所述方法还包括:所述网络设备接收来自所述终端设备的第五信息,所述第五信息用于上报所述至少一个AI功能中的第一AI功能更新后的所需AI资源数量。
- 根据权利要求24至28中任一项所述的方法,其特征在于,所述方法还包括:所述网络设备向所述终端设备发送所述多个AI模型。
- 一种终端设备,其特征在于,所述终端设备包括处理器和存储器,所述处理器和所述存储器相耦合,所述存储器用于存储计算机程序,当所述处理器运行所述计算机程序时,使得所述终端设备执行如权利要求1-16中任意一项所述的方法。
- 一种网络设备,其特征在于,所述网络设备包括处理器和存储器,所述处理器和所述存储器耦合,所述存储器用于存储计算机程序,当所述处理器运行所述计算机程序时,使得所述网络设备执行如权利要求17-29中任意一项所述的方法。
- 一种计算机可读存储介质,其特征在于,所述计算机可读存储介质存储有计算机指令,当所述计算机指令在终端设备上运行时,使得所述终端设备执行如权利要求1-16中任一项所述的方法;或者,当所述计算机指令在网络设备上运行时,使得所述网络设备执行如权利要求17-29中任一项所述的方法。
- 一种计算机程序产品,其特征在于,包含指令,当所述计算机指令在终端设备上运行时,使得所述终端设备执行如权利要求1-16中任一项所述的方法;或者,当所述计算机指令在网络设备上运行时,使得所述网络设备执行如权利要求17-29中任一项所述的方法。
- 一种芯片,其特征在于,所述芯片包括处理器与通信接口,所述处理器通过所述通信接口读取指 令并运行,当所述芯片安装在终端设备中,使得所述终端设备执行如权利要求1-16中任一项所述的方法;或者,当所述芯片安装在网络设备中,使得所述网络设备执行如权利要求17-29中任一项所述的方法。
- 一种通信系统,其特征在于,所述通信系统包括终端设备和网络设备;其中,所述终端设备用于执行如权利要求1-16中任一项所述的方法,所述网络设备用于执行如权利要求17-29中任一项所述的方法。
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP23914437.1A EP4642089A4 (en) | 2023-01-06 | 2023-11-28 | COMMUNICATION METHOD, NETWORK DEVICE AND TERMINAL DEVICE |
| US19/261,598 US20250337659A1 (en) | 2023-01-06 | 2025-07-07 | Communication method, network device, and terminal device |
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310020588 | 2023-01-06 | ||
| CN202310020588.6 | 2023-01-06 | ||
| CN202310149561.7A CN118317356A (zh) | 2023-01-06 | 2023-02-14 | 一种通信方法、网络设备和终端设备 |
| CN202310149561.7 | 2023-02-14 |
Related Child Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| US19/261,598 Continuation US20250337659A1 (en) | 2023-01-06 | 2025-07-07 | Communication method, network device, and terminal device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| WO2024146303A2 true WO2024146303A2 (zh) | 2024-07-11 |
| WO2024146303A3 WO2024146303A3 (zh) | 2024-08-29 |
Family
ID=91721119
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2023/134580 Ceased WO2024146303A2 (zh) | 2023-01-06 | 2023-11-28 | 一种通信方法、网络设备和终端设备 |
Country Status (4)
| Country | Link |
|---|---|
| US (1) | US20250337659A1 (zh) |
| EP (1) | EP4642089A4 (zh) |
| CN (1) | CN118317356A (zh) |
| WO (1) | WO2024146303A2 (zh) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120499706A (zh) * | 2025-07-03 | 2025-08-15 | 荣耀终端股份有限公司 | 波束管理的信道状态信息报告传输方法、设备及存储介质 |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN121487009A (zh) * | 2024-08-05 | 2026-02-06 | 上海推络通信科技合伙企业(有限合伙) | 被用于无线通信的方法和装置 |
| WO2026034460A1 (ja) * | 2024-08-08 | 2026-02-12 | 京セラ株式会社 | 通信制御方法及びユーザ装置 |
| CN121645506A (zh) * | 2024-08-14 | 2026-03-10 | 上海科邸斯科技有限公司 | 被用于无线通信的方法和装置 |
| CN118735229B (zh) * | 2024-09-03 | 2024-12-17 | 南通泰盈网络科技有限公司 | 一种pos机数据远程管理方法及管理系统 |
| CN121586021A (zh) * | 2026-01-27 | 2026-02-27 | 荣耀终端股份有限公司 | 通信方法和通信装置 |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021248423A1 (zh) * | 2020-06-12 | 2021-12-16 | 华为技术有限公司 | 人工智能资源的调度方法、装置、存储介质和芯片 |
| US20220294666A1 (en) * | 2021-03-05 | 2022-09-15 | Samsung Electronics Co., Ltd. | Method for support of artificial intelligence or machine learning techniques for channel estimation and mobility enhancements |
| US20220391776A1 (en) * | 2021-06-08 | 2022-12-08 | Texas Instruments Incorporated | Orchestration of multi-core machine learning processors |
| CN113922936B (zh) * | 2021-08-31 | 2023-04-28 | 中国信息通信研究院 | 一种ai技术信道状态信息反馈方法和设备 |
| CN115280833A (zh) * | 2022-06-16 | 2022-11-01 | 北京小米移动软件有限公司 | 信道状态信息的处理方法、装置及通信设备 |
-
2023
- 2023-02-14 CN CN202310149561.7A patent/CN118317356A/zh active Pending
- 2023-11-28 EP EP23914437.1A patent/EP4642089A4/en active Pending
- 2023-11-28 WO PCT/CN2023/134580 patent/WO2024146303A2/zh not_active Ceased
-
2025
- 2025-07-07 US US19/261,598 patent/US20250337659A1/en active Pending
Non-Patent Citations (1)
| Title |
|---|
| See also references of EP4642089A2 |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN120499706A (zh) * | 2025-07-03 | 2025-08-15 | 荣耀终端股份有限公司 | 波束管理的信道状态信息报告传输方法、设备及存储介质 |
Also Published As
| Publication number | Publication date |
|---|---|
| EP4642089A2 (en) | 2025-10-29 |
| EP4642089A4 (en) | 2026-04-01 |
| US20250337659A1 (en) | 2025-10-30 |
| CN118317356A (zh) | 2024-07-09 |
| WO2024146303A3 (zh) | 2024-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20250337659A1 (en) | Communication method, network device, and terminal device | |
| JP7563895B2 (ja) | 端末及び基地局 | |
| US20210345134A1 (en) | Handling of machine learning to improve performance of a wireless communications network | |
| US20250048342A1 (en) | Network node in communication system and method performed by the same | |
| WO2024229857A1 (en) | Machine learning model processing method, wireless communication device, and system | |
| US20250184798A1 (en) | Method and apparatus for measurement reporting in communication system | |
| KR20250144481A (ko) | 통신 방법, 장치 및 시스템 | |
| CN117561778A (zh) | 通信方法、装置、设备及存储介质 | |
| EP4068838B1 (en) | Communication method and communication apparatus | |
| US20250159523A1 (en) | Method and device for transmitting or receiving quantization-based channel state information in wireless communication system | |
| WO2021258316A1 (en) | Handling configurations for reporting channel state information | |
| WO2024067248A1 (zh) | 一种获取训练数据集的方法和装置 | |
| WO2023193659A1 (zh) | 网络节能方法及通信装置 | |
| CN115484676A (zh) | 数据传输方法及装置 | |
| US20260136172A1 (en) | Machine learning processing capability management | |
| US20260113130A1 (en) | Sidelink synchronization signal transmission | |
| WO2026097307A1 (en) | Reporting inactive channel state information report configurations associated with beam prediction | |
| WO2026073390A1 (en) | Prediction accuracy for channel state information reports | |
| WO2026097430A1 (en) | Feature group or subgroup for receiving associated identifiers | |
| WO2025166773A1 (en) | A method for life-cycle management of artificial intelligence and machine learning models in wireless networks | |
| CN120786690A (zh) | 通信方法及通信装置 | |
| WO2025209206A1 (zh) | 通信方法及通信装置 | |
| WO2025241972A1 (zh) | 通信处理的方法、装置及终端 | |
| WO2026067180A1 (zh) | 通信方法及通信装置 | |
| WO2026067252A1 (zh) | 通信方法和装置 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23914437 Country of ref document: EP Kind code of ref document: A2 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 202527063655 Country of ref document: IN |
|
| REG | Reference to national code |
Ref country code: BR Ref legal event code: B01A Ref document number: 112025014024 Country of ref document: BR |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2023914437 Country of ref document: EP |
|
| WWP | Wipo information: published in national office |
Ref document number: 202527063655 Country of ref document: IN |
|
| ENP | Entry into the national phase |
Ref document number: 2023914437 Country of ref document: EP Effective date: 20250723 |
|
| ENP | Entry into the national phase |
Ref document number: 2023914437 Country of ref document: EP Effective date: 20250723 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| ENP | Entry into the national phase |
Ref document number: 2023914437 Country of ref document: EP Effective date: 20250723 |
|
| WWP | Wipo information: published in national office |
Ref document number: 2023914437 Country of ref document: EP |
|
| ENP | Entry into the national phase |
Ref document number: 112025014024 Country of ref document: BR Kind code of ref document: A2 Effective date: 20250707 |