Communication methods and devices

Abstract

A communication method and apparatus are provided. The method includes a first communication apparatus determining a first reference signal resource group from a plurality of reference signal resource groups, and the first communication apparatus determining a first reference signal resource based on the first reference signal resource group and a model. According to the method of this application, a reference signal resource group is selected from a plurality of reference signal resource groups, and model inference is performed using the selected reference signal resource group. Compared to methods that perform model inference using fixed patterns or random patterns, this method can improve the accuracy of the inference results.

Description

Technical Field

[0001] Cross-reference of related applications This application claims the priority of Chinese Patent Application No. 202310227057.4, titled "COMMUNICATION METHOD AND APPARATUS", filed with the China National Intellectual Property Administration on February 27, 2023, the entire content of which is incorporated herein by reference.

[0002] Technical field Embodiments of this application relate to the field of communication technologies, particularly to communication methods and apparatuses.

Background Art

[0003] In wireless communication networks, such as mobile communication networks, the services supported by the network are becoming increasingly diverse, and thus, it is necessary to meet increasingly diverse requirements. For example, the network needs to support ultra-high speed, ultra-low latency, and / or a large number of connections. This feature makes network planning, network configuration, and / or resource scheduling increasingly complex. In addition, as network functions become more powerful, such as supporting higher frequencies, supporting higher-order multiple-input multiple-output (MIMO), beamforming and / or beam management, and other new technologies, network energy saving has become a research topic of concern. These new requirements, scenarios, and features pose unprecedented challenges to network planning, operation and maintenance, and efficient operation. To address this issue, artificial intelligence technology can be introduced into wireless communication networks to implement network intelligence. Based on this, how to effectively implement artificial intelligence in the network, for example, how to use artificial intelligence in information transmission, is a research-worthy issue.

Summary of the Invention

[0004] Embodiments of this application provide a communication method and apparatus that integrate artificial intelligence technology into a wireless communication network and improve the accuracy of signal resources inferred based on a model. [Means for solving the problem]

[0005] According to a first embodiment, a communication method is provided which can be performed by a first communication device. The first communication device may be a terminal, or a chip or circuit within a terminal. The method includes the first communication device determining a first reference signal resource group from a plurality of reference signal resource groups, and the first communication device determining a first reference signal resource based on the first reference signal resource group and a model.

[0006] In the design described above, an example is used where the first communication device is a terminal. Each time the terminal predicts the optimal received beam, the terminal determines a reference signal resource group from a plurality of reference signal resource groups. The terminal performs model inference by using the determined reference signal resource group. In this way, each time the terminal predicts the optimal received beam, a different reference signal resource group may be selected as the input to the model inference. This is relatively flexible and solves the problem of the prior art in which the accuracy of the predicted received beam is low because the terminal uses beams at several fixed locations as input to the model inference each time it predicts the optimal beam. For example, when the terminal predicts the optimal received beam at a particular time, if beams at several fixed locations are fixedly used as input to the model inference, some of the beams at those fixed locations may be blocked. As a result, when the terminal measures multiple beams, it cannot extract the measurement results of the blocked beams. Therefore, the accuracy of predicting the optimal received beam is low. However, in the solution of the present invention, the terminal can use a different beam each time it makes a prediction, avoiding the possibility of beams being blocked when fixed beams are used for prediction, and as a result, the accuracy of predicting the optimal beam can be improved.

[0007] In one design, the determination of a first reference signal resource group by a first communication device from a plurality of reference signal resource groups includes the first communication device receiving first instruction information from a second communication device, wherein the first instruction information includes instruction information for a first reference signal resource group, and the first communication device determining a first reference signal resource group from a plurality of reference signal resource groups based on the first instruction information. The second communication device may be an access network device, or a chip or circuit within an access network device.

[0008] In the design described above, an example is used where the first communication device is a terminal and the second communication device is an access network device. When a terminal accesses a different access network device, the accessed access network device learns that the beams supported by the access network device block each other. Therefore, the access network device can select a reference signal resource group from multiple reference signal resource groups that corresponds to an unblocked beam, or a reference signal resource group that corresponds to a beam that is not blocked, and indicate the selected reference signal resource group to the terminal. In this way, the terminal can extract measurement results from all reference signals corresponding to the reference signal resource group indicated by the access network device, thereby improving the accuracy of predicting the optimal received beam. Furthermore, in this design, each group of reference signal resources may have a corresponding number, and the first indication information may include the number of the first reference signal resource group. Based on the number included in the first indication information, the terminal can determine the first reference signal resource group indicated by the access network device. Since multiple reference signal resource groups are numbered separately, and the first indication information includes the number of the first reference signal resource group, the overhead for indicating reference signal resource groups can be reduced.

[0009] In one design, the determination of a first reference signal resource group by a first communication device from a plurality of reference signal resource groups includes the first communication device receiving second instruction information from a second communication device, wherein the second instruction information includes instruction information for at least one reference signal resource included in the first reference signal resource group, and the first communication device determining the first reference signal resource group from a plurality of reference signal resource groups based on the second instruction information.

[0010] Unlike the design described above, this design does not require separate numbering for multiple reference signal resource groups. Each group of reference signal resources contains at least one reference signal resource. Each reference signal resource included in each group of reference signal resources has a global identifier, which may be a number corresponding to each reference signal resource in the universal set of reference signal resources. Since the reference signal resources included in all reference signal resource groups are not exactly the same, each reference signal resource included in each reference signal resource group can indicate the corresponding reference signal resource group. This design reduces processing steps because it does not require separate numbering for multiple reference signal resource groups.

[0011] In one design, the method further includes a first communication device receiving configuration information from a second communication device, wherein the configuration information includes configuration information for a plurality of reference signal resource groups, and the first communication device acquiring a plurality of reference signal resource groups based on the configuration information for the plurality of reference signal resource groups.

[0012] The above design uses an example where the first communication device is a terminal and the second communication device is an access network device. An access network device can define multiple reference signal resource groups. For example, when a terminal accesses the network, multiple reference signal resource groups are configured for the terminal. Compared to a terminal, an access network device has more powerful processing capabilities. Because the aforementioned design is used, an access network device can accurately define multiple reference signal resource groups.

[0013] In one design, the determination of a first reference signal resource by a first communication device based on a first reference signal resource group and a model includes the first communication device measuring the first reference signal resource group and obtaining the measurement results of the first reference signal resource group, the first communication device determining the inputs of a model based on the measurement results of the first reference signal resource group, the first communication device determining the outputs of a model based on the inputs of a model and a model, and the first communication device determining a first reference signal resource based on the outputs of a model.

[0014] According to a second embodiment, a communication method is provided which can be performed by a second communication device. The second communication device may be an access network device, or a chip or circuit within an access network device. The method includes the second communication device generating configuration information, the configuration information including configuration information for a plurality of reference signal resource groups, and the second communication device transmitting the configuration information to a first communication device.

[0015] In one design, the method further includes a second communication device transmitting first instruction information to the first communication device, wherein the first instruction information includes instruction information for a first reference signal resource group among a plurality of reference signal resource groups.

[0016] In one design, the method further includes a second communication device transmitting a second instruction information to a first communication device, the second instruction information including instruction information for at least one reference signal resource included in a first reference signal resource group, and the first reference signal resource group belonging to a plurality of reference signal resource groups.

[0017] According to a third aspect, an apparatus is provided. The apparatus includes a corresponding unit or module for performing the method according to the first or second aspect. The unit or module may be implemented by hardware circuitry, by software, or by a combination of hardware circuitry and software.

[0018] According to a fourth aspect, an apparatus is provided, comprising a processor and an interface circuit. The processor is configured to communicate with another apparatus via the interface circuit and to perform the method according to the first or second aspect. One or more processors are present.

[0019] According to the fifth aspect, an apparatus is provided, including a processor coupled to memory. The processor is configured to execute a program stored in memory and to perform the method according to the first or second aspect. The memory may be located inside or outside the apparatus. In addition, one or more processors may be present.

[0020] According to the sixth aspect, an apparatus is provided, comprising a processor and memory. The memory is configured to store computer instructions. When the apparatus is executed, the processor executes the computer instructions stored in memory, enabling the apparatus to perform the method according to the first or second aspect.

[0021] According to the seventh aspect, a chip system is provided which includes a processor or circuit and is configured to perform the method according to the first or second aspect.

[0022] According to the eighth aspect, a computer-readable storage medium is provided. The computer-readable storage medium stores instructions. When an instruction is executed by a communication device, the method according to the first or second aspect is performed.

[0023] According to the ninth aspect, a computer program product is provided. The computer program product includes a computer program or instructions. When the computer program or instructions are executed by the device, the method according to the first or second aspect is performed.

[0024] According to the tenth aspect, a system is provided, including a first communication device that executes the method according to the first aspect and a second communication device that executes the method according to the second aspect.

[0025] For the advantageous effects of the second aspect to the tenth aspect, please refer to the description of the first aspect.

Brief Description of the Drawings

[0026] [Figure 1] It is a diagram of a communication system according to an embodiment of the present application.

[0027] [Figure 2] It is a diagram of deploying an AI model according to an embodiment of the present application.

[0028] [Figure 3] It is another diagram of deploying an AI model according to an embodiment of the present application.

[0029] [Figure 4] It is a diagram of the architecture of an access network device according to an embodiment of the present application.

[0030] [Figure 5] It is another diagram of the architecture of an access network device according to an embodiment of the present application.

[0031] [Figure 6] It is a diagram of the application architecture of an AI model according to an embodiment of the present application.

[0032] [Figure 7] It is a diagram of a neuron according to an embodiment of the present application.

[0033] [Figure 8] It is a diagram of the hierarchical relationship of a neural network according to an embodiment of the present application.

[0034] [Figure 9] This is a flowchart of the communication method according to the embodiment of this application.

[0035] [Figure 10] This is another flowchart of the communication method according to the embodiment of this application.

[0036] [Figure 11] This is a diagram showing four reference signal resource groups according to an embodiment of the present application.

[0037] [Figure 12] This is yet another flowchart of the communication method according to the embodiment of this application.

[0038] [Figure 13] This is a diagram of the apparatus according to an embodiment of the present application.

[0039] [Figure 14] Another diagram of the apparatus according to the embodiment of this application. [Modes for carrying out the invention]

[0040] Figure 1 is a diagram of the architecture of a communication system 1000 to which embodiments of the present application can be applied. As shown in Figure 1, the communication system 1000 includes a wireless access network 100 and a core network 200. Optionally, the communication system 1000 may further include an internet 300.

[0041] The wireless access network 100 may include at least one access network device (e.g., 100a and 110b in Figure 1) and may further include at least one terminal (e.g., 120a to 120j in Figure 1). The terminal is wirelessly connected to the access network device, and the access network device is wirelessly or wiredly connected to the core network. The core network device and the access network device may be separate physical devices, or the functions of the core network device and the logical functions of the access network device may be integrated into the same physical device, or some of the functions of the core network device and some of the functions of the access network device may be integrated into one physical device. The terminals may be connected to each other by wired or wireless means, and the access network devices may be connected to each other by wired or wireless means. Figure 1 is for illustrative purposes only. The communication system 1000 may further include other network devices, for example, wireless relay devices and wireless backhaul devices not shown in Figure 1.

[0042] Access network devices may include base stations, evolved NodeBs (eNodeBs), transmission receiving points (TRPs), next-generation NodeBs (gNBs) in 5th generation (5G) mobile communication systems, access network devices in open radio access networks (O-RANs), next-generation base stations in 6th generation (6G) mobile communication systems, base stations in future mobile communication systems, access nodes in wireless fidelity (Wi-Fi) systems, etc., or they may be modules or units that complete part of the functions of a base station, such as central units (CUs), distributed units (DUs), central unit control plane (CU-CP) modules, and central unit user plane (CU-UP) modules. The access network device may be a macro base station (e.g., 110a in Figure 1), a micro base station or indoor base station (e.g., 110b in Figure 1), or a relay node or donor node, etc. The specific technologies and device forms used by the access network device are not limited to the embodiments of this application.

[0043] In embodiments of this application, the device configured to implement the functions of an access network device may be an access network device, or a device capable of supporting an access network device when implementing its functions, such as a chip system, hardware circuitry, software modules, or a combination of hardware circuitry and software modules. The device may be mounted on an access network device, or used in combination with an access network device. In embodiments of this application, the chip system may include a chip, or it may include a chip and other separate components. For convenience of explanation, the technical solutions provided in embodiments of this application will be described below using an example where the device configured to implement the functions of an access network device is an access network device.

[0044] Terminals are also sometimes called terminal devices, user equipment (UE), mobile stations, mobile terminals, etc. Terminals can be widely used for communication in a variety of scenarios, including, but are not limited to, one or more of the following scenarios: device-to-device (D2D), vehicle-to-everything (V2X), machine-type communication (MTC), internet of things (IoT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grids, smart furniture, smart offices, smart wearables, smart transportation, smart cities, etc. Terminals may also be mobile phones, tablet computers, computers with wireless transceiver functionality, wearable devices, vehicles, unmanned aerial vehicles, helicopters, airplanes, ships, robots, mechanical arms, smart home devices, etc. The specific technologies and specific device forms used by terminals are not limited to the embodiments of this application.

[0045] In embodiments of this application, the device configured to implement the functions of a terminal may be a terminal, or a device that can assist the terminal in implementing those functions, such as a chip system, hardware circuitry, software module, or a combination of hardware circuitry and software module. The device may be mounted on a terminal or used in combination with a terminal. For convenience of explanation, the technical solutions provided in embodiments of this application will be described below using an example where the device configured to implement the functions of a terminal is a terminal.

[0046] Access network devices and terminals may be in a fixed location or may be mobile. Access network devices and / or terminals may be deployed on land, including indoors or outdoors, may be handheld or mounted on a vehicle, may be deployed on water, or may be deployed in the air on an aircraft, balloon, or satellite. The application scenarios for access network devices and terminals are not limited to the embodiments of this application. Access network devices and terminals may be deployed in the same or different scenarios. For example, both access network devices and terminals may be deployed on land, or the access network device may be deployed on land and the terminal on water. Examples are not provided individually.

[0047] The roles of access network devices and terminals may be relative. For example, the helicopter or unmanned aerial vehicle 120i in Figure 1 may be configured as a mobile access network device. For terminal 120j, which accesses the wireless access network 100 via 120i, terminal 120i is an access network device. However, for access network device 110a, 120i is a terminal; specifically, 110a and 120i communicate with each other according to the wireless air interface protocol. Alternatively, 110a and 120i communicate with each other by using the interface protocol between access network devices. In this case, for 110a, 120i is also an access network device. Therefore, access network devices and terminals are sometimes collectively referred to as communication devices. 110a and 110b in Figure 1 may be called communication devices having the functionality of access network devices, and 120a to 120j in Figure 1 may be called communication devices having the functionality of terminals.

[0048] Communication between access network devices and terminals, between access network devices, or between terminals may be carried out using licensed spectrum, unlicensed spectrum, or both licensed and unlicensed spectrum, using spectrum below 6 gigahertz (GHz), using spectrum above 6 GHz, or using both spectrum below 6 GHz and spectrum above 6 GHz. The frequency resources used by wireless communication in embodiments of this application are not limited.

[0049] In embodiments of this application, an independent network element, sometimes called an artificial intelligence (AI) network element or AI node, can be introduced into the communication system shown in Figure 1 to implement AI-related operations. The AI ​​network element may be directly connected to an access network device within the communication system, or it may be indirectly connected to the access network device via a third-party network element. The third-party network element may be a core network element, such as an authentication management function (AMF) or a user plane function (UPF). Alternatively, an AI function, AI module, or AI entity may be built within another network element within the communication system to implement AI-related operations. The other network element may be an access network device, a core network device, a network management system, etc. In this case, the network element that performs the AI-related operations may be a network element with built-in AI functionality. Operation administration and maintenance (OAM) is used to perform operation, management, maintenance, etc., for the access network device and / or core network device.

[0050] As shown in Figure 2 or Figure 3, the AI ​​model may be deployed on at least one of the following: a core network device, an access network device, a terminal, an OAM, etc., and the corresponding functionality is implemented by using the AI ​​model. In embodiments of this application, the AI ​​models deployed on different nodes may be the same or different. Different models include at least one of the following differences: namely, at least one of different structural parameters of the model, e.g., differences in the number of layers and / or weights of the model, different input parameters of the model, or different output parameters of the model. Different input parameters and / or different output parameters of the model may be described as different functions of the model. Unlike Figure 2, in Figure 3, the functionality of the access network device is divided into CU and DU. Optionally, the CU and DU may be the CU and DU in an O-RAN architecture. One or more AI models may be deployed within the CU, and / or one or more AI models may be deployed within the DU. Furthermore, the CU in Figure 3 may be divided into CU-CP and CU-UP. Optionally, one or more AI models may be deployed to CU-CP and / or one or more AI models may be deployed to CU-UP. Optionally, in Figure 2 or Figure 3, the OAM for the access network device and the OAM for the core network device may be deployed separately and independently.

[0051] In embodiments of this application, the access network device may use an O-RAN architecture. An example of an O-RAN architecture is described below, which is not intended to limit the embodiments of this application.

[0052] Refer to Figure 4 for the first design. The access network devices include a quasi-real-time access network intelligent controller (RAN intelligent controller, RIC), CU, DU, RU, etc. The quasi-real-time RIC is used for model training and inference. For example, the quasi-real-time RIC may train an AI model and use the AI ​​model for inference. For example, the quasi-real-time RIC may acquire network-side or terminal-side information from one or more of the CU, DU, RU, terminal, etc. This information may be used as training data or inference data. For example, the information may be used as training data, and the quasi-real-time RIC may train an AI model using the collected training data. Alternatively, the information may be used as inference data, and the quasi-real-time RIC may perform model inference based on the collected inference data and the AI ​​model to determine the inference result. Optionally, the quasi-real-time RIC may transmit the inference result to one or more of the CU, DU, RU, terminal, etc. Optionally, the CU and DU may exchange inference results. For example, the quasi-real-time RIC transmits the inference result to the CU, and the CU forwards the inference result to the DU. Optionally, the DU and RU may exchange inference results. For example, a near-real-time RIC might send the inference results to the DU, or a near-real-time RIC might send the inference results to the CU, which then forwards the inference results to the DU. The DU then forwards the inference results to the RU.

[0053] In the first design, the quasi-real-time RIC is included in the access network device. Whether the non-real-time RIC is included outside the access network device is not limited. For example, the non-real-time RIC may be included outside the access network device, or it may not be included outside the access network device.

[0054] For the second design, please refer to Figure 4. The non-real-time RIC is located outside the access network device. For example, the non-real-time RIC may be located in the OAM or core network device, but is not limited to this. The non-real-time RIC may train an AI model and use the AI ​​model for inference. Optionally, the non-real-time RIC may collect network-side or terminal-side information from one or more of the CU, DU, RU, terminal, etc. This information may be used as training data or inference data. For example, the information may be used as training data, and the non-real-time RIC may train an AI model by using the training data. Alternatively, the information may be used as inference data, and the non-real-time RIC may use the inference data and the AI ​​model to determine the inference result. Optionally, the non-real-time RIC may transmit the inference result to one or more of the CU, DU, RU, terminal, etc. Optionally, the CU and DU may exchange inference results. The DU and RU may exchange inference results.

[0055] In the second design, the non-real-time RIC is located outside the access network device. Whether or not the quasi-real-time RIC is included in the access network device is not limited. For example, the quasi-real-time RIC may or may not be included in the access network device.

[0056] In the third design, please refer to Figure 4. The quasi-real-time RIC is contained within the access network device, and the non-real-time RIC is contained outside the access network device. Similar to the first design, the quasi-real-time RIC may perform model training and inference, and / or similar to the second design, the non-real-time RIC may perform model training and inference, and / or the non-real-time RIC may perform model training and the quasi-real-time RIC may perform model inference. For example, the non-real-time RIC may transmit the trained AI model to the quasi-real-time RIC, which then uses the AI ​​model for model inference. Optionally, the non-real-time RIC and / or the quasi-real-time RIC may collect network-side or terminal-side information from one or more of the CU, DU, RU, terminal, etc. This information may be used as training data or inference data. For example, the information may be used as training data, and the non-real-time RIC may train the AI ​​model using the training data. Using this information as inference data, the quasi-real-time RIC may use the AI ​​model and the inference data to determine the inference result. Optionally, the quasi-real-time RIC may transmit the inference result to one or more of the CU, DU, RU, terminal, etc. CU and DU may optionally exchange their inference results. DU and RU may also exchange their inference results.

[0057] Figure 5 shows another O-RAN architecture according to one embodiment of the present application. Compared to Figure 4, in Figure 5 the CU is separated into CU-CP and CU-UP.

[0058] In this embodiment of the present application, the first communication device may perform AI-related operations based on a first reference signal resource group. The AI ​​technology described below is not intended to limit this embodiment of the present application.

[0059] An AI model is a specific implementation of an AI function. An AI model represents a mapping relationship between the inputs and outputs of a model. An AI model may be a neural network, a linear regression model, a decision tree model, a support vector machine (SVM), a Bayesian network, a Q-learning model, another machine learning model, etc. In this embodiment of the present application, the AI ​​function may include one or more of the following: namely, data collection (collection of training data and / or inference data), data preprocessing, model training, model information publication (construction of model information), model validation, model inference, inference result publication, etc. In this embodiment of the present application, an AI model may be abbreviated as a model.

[0060] Figure 6 shows the application architecture of an AI model. The data source is configured to store training data and inference data. The model training node (model training host) analyzes or trains the training data provided by the data source to obtain an AI model and deploys the AI ​​model in the model inference node (model inference host). Optionally, the model training node may further update the AI ​​model deployed on the model inference node. The model inference node may further feed back relevant information about the deployed model to the model training node so that the model training node can optimize or update the deployed AI model.

[0061] The acquisition of an AI model through learning using a model training node is equivalent to the model training node acquiring a mapping relationship between the model's inputs and outputs through learning based on training data. A model inference node uses the AI ​​model to perform inference based on inference data provided by a data source and obtains an inference result. This method may also be described as follows: a model inference node inputs inference data into an AI model and obtains an output through the AI ​​model. The output is the inference result. The inference result may indicate configuration parameters used (executed) by actor objects, and / or operations performed by actor objects. The inference result may be centrally planned by an actor entity and sent to one or more actor objects (e.g., a network entity) for execution. Optionally, an actor entity or actor object may feed back parameters or measurements of measures collected by the actor entity or actor object to the data source. This process may be called performance feedback, and the fed-back parameters may be used as training data or inference data. Optionally, feedback information related to model performance may be further determined based on the inference results output by the model inference node, and this feedback information may be fed back to the model inference node. The model inference node may then feed back model performance information to the model training node based on this feedback information. As a result, the model training node performs optimization, updates, etc., on the deployed AI model. This process is sometimes called model feedback.

[0062] An AI model may be a neural network or another machine learning model. Let's take a neural network as an example. A neural network is a specific implementation of machine learning techniques. According to universal approximation theorems, a neural network can theoretically approximate any continuous function, and therefore has the ability to learn any mapping. Thus, a neural network can accurately perform abstract modeling for complex high-dimensional problems.

[0063] The idea of ​​neural networks originates from the neuronal structure of brain tissue. Each neuron performs a weighted sum operation on the input values ​​of the neuron and outputs the result of the weighted sum via an activation function. Figure 7 shows the structure of a neuron. The input to the neuron is x=[x0,x1,...,x n ] and the weights corresponding to the input are w=[w,w1,...,w n It is assumed that ] are the case for each, and the offset of the weighted sum is b. The form of the activation function can be diverse. If the activation function of a neuron is y=f(z)=max(0,z), then the output of the neuron is,

number

number

[0064] A neural network generally has a multilayer structure, and each layer may contain one or more neurons. Increasing the depth and / or width of a neural network can improve its expressive power and provide stronger information extraction and abstract modeling capabilities for complex systems. The depth of a neural network may refer to the number of layers contained in the neural network, and the number of neurons contained in each layer may be called the width of the layer. Figure 8 is a diagram of the hierarchical relationships of a neural network. In one implementation, a neural network includes an input layer and an output layer. After processing the received input with neurons, the input layer of the neural network transfers the result to the output layer, which obtains the output result of the neural network. In another implementation, a neural network includes an input layer, a hidden layer, and an output layer. After processing the received input with neurons, the input layer of the neural network transfers the result to an intermediate hidden layer, which then transfers the computation result to the output layer or an adjacent hidden layer, and finally the output layer obtains the output result of the neural network. A neural network may contain one hidden layer or multiple consecutively connected hidden layers. This is not limited to this. A loss function may be defined during the training of a neural network. The loss function describes the gap or difference between the output value of the neural network and an ideal target value. The specific form of the loss function is not limited to this embodiment of the present application. The neural network training process is the process of adjusting the parameters of the neural network, such as the number and width of the neural network's hierarchies, the weights of the neurons, and / or the parameters of the activation function of the neurons, so that the value of the loss function is below a threshold or satisfies the target condition.

[0065] As 5G communication systems develop, the spectrum used gradually expands to higher frequency bands. Due to physical transmission characteristics, free-space transmission loss and penetration loss in high-frequency bands are significantly higher than those in low-frequency bands. To compensate for these drawbacks and resist the attenuation of received signals caused by losses, large antenna arrays are typically used in high-frequency bands. Antenna beamforming is used to concentrate energy into narrow beams, which can effectively improve network coverage and enhance the user experience.

[0066] Narrow beams have a similar effect to spotlights, concentrating limited transmission energy in a narrow direction, which can significantly improve the coverage area of ​​access network devices. However, while narrow beams improve coverage, they also introduce enormous beam management overhead. Access network devices require narrower beams to cover all available space. To select the most suitable beam for a terminal, the terminal must traverse a large number of candidate beams. This results in enormous beam management overhead.

[0067] For example, the process of a terminal selecting the beam best suited to it, also known as beam alignment, involves three phases. The first phase is coarse alignment. For example, a wide beam is configured by an access network device. The terminal sweeps through all wide beams to obtain the corresponding measurement results and find the optimal wide beam. In the second phase, a narrow beam is configured within the coverage area of ​​the wide beam or a corresponding range around the wide beam. The terminal sweeps through the narrow beam to obtain the optimal narrow beam. The third phase mainly concerns beam maintenance and beam recovery processes. This is not relevant to this embodiment of the present application and will not be described in detail.

[0068] To significantly reduce beam sweep overhead, AI technology is introduced. For example, an access network device transmits a reference signal using a transmission beam, and a terminal receives the reference signal using a different beam, measuring the received reference signal using each beam to determine the measurement result. The measurement result may be reference signal received power (RSRP), reference signal received quality (RSRQ), signal-to-interference plus noise ratio (SINR), etc. In one design, the AI ​​model may be pre-trained, and the AI ​​model may predict the optimal receiving beam. In one implementation, one or more beams at fixed positions are selected from all beams, and the measurement results of one or more beams at fixed positions are used as input to the AI ​​model. The optimal receiving beam may be determined based on the output of the AI ​​model. This solution will be referred to as the fixed-pattern solution below. In another implementation, several beams are randomly selected from all beams, and the measurement results of the randomly selected beams are used as input to the AI ​​model. The optimal receiving beam may be determined based on the output of the AI ​​model. This solution will be referred to as the random pattern solution below.

[0069] In the case of a fixed pattern solution, the beam is blocked during beam sweeping. When a terminal performs a beam sweep on a beam at a fixed position, the beam is blocked, and it is not possible to extract measurement results for the blocked beam. As a result, the accuracy of the optimal beam obtained by inference is low. Furthermore, a fixed pattern is configured for the terminal by the access network device. When a terminal performs a cell handover, for example, if the source cell belongs to access network device 1 and the target cell belongs to access network device 2, access network device 1 and access network device 2 may each configure a fixed pattern for the terminal. The fixed patterns configured by access network device 1 and access network device 2 may be different. As a result, the AI ​​model may not match the fixed pattern, affecting the inference performance of the AI ​​model. For example, during model training, the AI ​​model is trained using fixed pattern 1. The AI ​​model may implement good inference capabilities for fixed pattern 1. If access network device 2 configures a fixed pattern 2, which is different from fixed pattern 1, for the terminal during a cell handover, the terminal will perform AI estimation using fixed pattern 2, which may result in lower accuracy of the optimal received beam obtained by estimation.

[0070] In the case of random pattern solutions, the convergence of the AI ​​model during training depends on feature extraction. The AI ​​model establishes a correspondence between inputs and outputs. Random and irregular patterns can make it difficult for the AI ​​model to converge. In other words, the AI ​​model cannot achieve ideal inference performance.

[0071] From this perspective, embodiments of the present application provide a solution for improving the accuracy of predicting the optimal received beam. This solution includes: a first communication device determines a first reference signal resource group from a plurality of reference signal resource groups; the first communication device determines a first reference signal resource based on the first reference signal resource group and a model; the first communication device may be a terminal, or a chip, circuit, etc. used in the terminal; however, it is not limited to this. In these embodiments of the present application, there is a correspondence between the reference signal resource and the beam. Therefore, the reference signal resource may be used to represent the beam. In other words, in the following description, the beam and the reference signal resource may represent the same concept and may be used interchangeably. In other words, in these embodiments of the present application, the terminal determines a beam group from a plurality of beam groups, where the beam group may be called the first beam group, and can perform model inference based on the first beam group to determine a beam that satisfies the conditions. The beam that satisfies the conditions may be the optimal received beam. Compared to existing solutions where model inference is performed based on beams at several fixed locations to predict the optimal received beam, this solution allows the appropriate beam group to be selected from multiple beam groups. For example, an access network device can select beam groups that do not block each other and instruct the terminal to use the selected beam group, thereby avoiding situations where the terminal cannot extract beam measurement results and improving the accuracy of the optimal received beam obtained by inference.

[0072] Furthermore, an AI model used as a model for inferring / predicting the optimal received beam may be pre-trained using multiple beam groups. In this way, the AI ​​model can implement accurate inference performance for any one of the multiple beam groups. For example, in the aforementioned solution that predicts or infers the optimal received beam based on a fixed position, the model is trained based on the fixed position. Access network devices constitute fixed positions for terminals. If the fixed positions configured by the access network devices differ from the fixed positions used for model training, a problem may arise, for example, in which the accuracy of the optimal received beam obtained through inference or prediction is lower. In the solution of this application, the model is trained by using multiple beam groups and can implement accurate inference or prediction of the optimal received beam regardless of which beam group is selected from the multiple beam groups. In addition, in this application, since the AI ​​model is trained by using multiple beam groups, the convergence performance of the trained AI model can be better, and the AI ​​model can achieve ideal inference capabilities.

[0073] For ease of understanding, the concept of a beam is described and / or explained below. This description and / or explanation is not intended to limit this embodiment of the present application.

[0074] A beam is a special directional transmission and reception effect formed by the transmitter or receiver of an access network device or terminal using an antenna array, similar to a flashlight that focuses light in a certain direction to form a beam of light. By transmitting and receiving signals in the form of a beam, the signal transmission distance can be effectively increased.

[0075] The beam may be a wide beam, a narrow beam, or another type of beam. The technique for forming the beam may be beamforming technology or another technique. Specifically, the beamforming technology may be digital beamforming technology, analog beamforming technology, hybrid digital / analog beamforming technology, etc.

[0076] Beams generally correspond to resources. For example, during beam measurement, an access network device measures different beams by using different resources, the terminal provides feedback on the measured resource quality, and the access network device learns the corresponding beam quality. During data transmission, beam information is also indicated by the resources corresponding to the beam information. For example, an access network device indicates the transmission configuration indicator (TCI, TCI-state) by using the transmission configuration indicator (TCI) in the down link control information (DCI), and the terminal decides to use the beam corresponding to the reference resource based on the reference signal resource included in the TCI state.

[0077] In communication protocols, beams may be specifically represented as reference signal resources, digital beams, analog beams, spatial domain filters, spatial filters, spatial parameters, TCI, TCI states, etc. The beam used to transmit a signal may be called a transmit beam (or Tx beam), spatial domain transmission filter, spatial transmission filter, spatial domain transmission parameter, spatial transmission parameter, etc. The beam used to receive a signal may be called a receive beam (or Rx beam), spatial domain reception filter, spatial reception filter, spatial domain reception parameter, spatial reception parameter, etc. In the following description of embodiments of this application, beams are described as reference signal resources.

[0078] For example, the first communication device is a terminal, and the second communication device is an access network device. As shown in Figure 9, a procedure is provided, which includes the following steps.

[0079] Step 901: The terminal determines a first reference signal resource group from among multiple reference signal resource groups.

[0080] In one design, multiple reference signal resource groups include a first reference signal resource group and further include reference signal resource groups other than the first reference signal resource group. Each of the multiple reference signal resource groups includes at least one reference signal resource. The number of reference signal resources included in different reference signal resource groups may be the same or different. This is not limited. For example, the number of reference signal resources included in different reference signal resource groups may be the same. For example, multiple reference signal resource groups include a first reference signal resource group and a second reference signal resource group. The first reference signal resource group and the second reference signal resource group each include 16 reference signal resources. Alternatively, the first reference signal resource group includes 16 reference signal resources and the second reference signal resource group includes 32 reference signal resources.

[0081] In a single design, multiple reference signal resource groups belong to a subset of the reference signal resource universal set. For example, the reference signal resource universal set contains 64 reference signal resources. Each of the multiple reference signal resource groups may contain 16 reference signal resources. These 16 reference signal resources belong to the 64 reference signal resources. For example, if the numbers of the 64 reference signal resources are 0 to 63, a reference signal resource group may contain any 16 of the 64 reference signal resources. In this explanation, each reference signal resource group contains a portion of the reference signal resources within the reference signal resource universal set; in other words, just as a beam group contains a portion of the beams within the beam universal set, reference signal resource groups are also called sparse beam groups, sparse beam patterns, etc.

[0082] Multiple reference signal resource groups may be specified and preset in the protocol, configured by the access network device for a terminal, and even determined and configured by the terminal for the access network device. Let us use an example in which the access network device configures multiple reference signal resource groups for a terminal. Prior to step 901, the procedure may further include the following steps: Step 900: The access network device sends configuration information to the terminal, and the terminal receives configuration information from the access network device, where the configuration information includes configuration information for multiple reference signal resource groups. Furthermore, the terminal determines multiple reference signal resource groups based on the configuration information for multiple reference signal resource groups.

[0083] Step 902: The terminal determines the first reference signal resource based on the first reference signal resource group and model.

[0084] For example, the first reference signal resource may be any reference signal resource that satisfies the conditions. Optionally, the reference signal resource that satisfies the conditions may also be the reference signal resource corresponding to the terminal's optimal received beam. The model is configured to predict the reference signal resource that satisfies the conditions. The model may be called an AI model, the first model, etc., but is not limited to these. The model may be specified in the protocol, preset, trained by another device for the terminal, trained by the terminal, etc. Other devices may include access network devices, core network devices, OAMs, etc.

[0085] In one design, the terminal measures a first reference signal resource group and obtains the measurement results for the first reference signal resource group. Based on the measurement results for the first reference signal resource group, the terminal determines the input to the model. Based on the model's input and the model, the terminal determines the model's output. Based on the model's output, the terminal determines the first reference signal resource. For example, if the first reference signal resource group contains 16 reference signal resources, the terminal measures the reference signals corresponding to the 16 reference signal resources included in the first reference signal resource group and obtains the measurement results for the 16 reference signal resources. Optionally, the reference signals may be a synchronization signal block (SSB), a channel state information-reference signal (CSI-RS), etc., and the measurement results may be RSRP, RSRQ, SINR, etc. The terminal inputs the measurement results of the 16 reference signals into the model and determines the first reference signal resource based on the model's output. For example, the model's output may be the first reference signal resource directly, or the model's output may be processed to obtain the first reference signal resource.

[0086] For example, when training a model using a classification training method, the model's output is the probability that each reference signal resource will be a reference signal resource that satisfies the conditions. For example, if the universal set of reference signal resources contains 64 reference signal resources, the model's output is the probability that each of the 64 reference signal resources will be a reference signal resource that satisfies the conditions. The terminal selects the reference signal resource with the highest probability as the reference signal resource that satisfies the conditions, in other words, as the first reference signal resource. Alternatively, when training a model using a regression training method, the model's output is the predicted measurement result corresponding to each reference signal resource. The terminal selects the reference signal resource with the largest measurement result value as the reference signal resource that satisfies the conditions, in other words, as the first reference signal resource.

[0087] In one design, the model's inputs include N input ports, and the first reference signal resource group includes M reference signal resources. In this case, the terminal obtains the measurement results of M reference signals. When the value of N is equal to the value of M, the terminal inputs the M measurement results into the model's N input ports. For example, if each of several reference signal resource groups includes the same number of reference signal resources, e.g., 16 reference signal resources, the model may include 16 input ports, and both the values ​​of N and M are 16. Alternatively, when the value of N is greater than the value of M, the terminal may input the M measurement results into the corresponding positions of the N input ports, and input preset values ​​into the remaining NM input ports. The preset values ​​may be zero, or they may be non-zero measurement values, etc. For example, if the first reference signal resource group includes 16 reference signal resources, the terminal obtains the measurement results of 16 reference signals. In this case, the value of M is 16. If the reference signal resource universal set includes 64 reference signal resources, the model includes 64 input ports. In this case, the value of N is 64. In this case, the terminal inputs the measurement results of 16 reference signals into 16 input ports within the model, and the remaining 48 input ports into preset values.

[0088] According to the design described above, the terminal determines a first reference signal resource group from a plurality of reference signal resource groups. The terminal performs model inference using the first reference signal resource group, solving problems such as low accuracy of inference results due to failures in extracting measurement results of blocked beams in fixed pattern designs, and further solving problems such as low accuracy of inference results due to poor convergence of models trained using random patterns. In other words, the solution according to the embodiments of this application can improve the accuracy of inference results.

[0089] [Embodiment 1] In one design, the determination of a first reference signal resource group by a terminal from a plurality of reference signal resource groups includes: the terminal receiving first instruction information from an access network device, wherein the first instruction information includes instruction information for the first reference signal resource group. Based on the first instruction information, the terminal determines the first reference signal resource group from a plurality of reference signal resource groups.

[0090] For example, multiple reference signal resource groups include at least two reference signal resource groups, and the two reference signal resource groups may use various encoding schemes. For example, multiple reference signal resource groups may use encoding schemes such as decimal, binary, another number system (e.g., octal or hexadecimal), Greek letters (e.g., α, β, γ, or δ), and case-sensitive characters (e.g., a, A, b, or B). Decimal encoding is used as an example. The encoding of the multiple reference signal resource groups may be sequential, such as 1, 2, 3, 4, etc. The first instruction information may indicate any one of the multiple reference signal resource groups, and the reference signal resource group indicated by the first instruction information is called the first reference signal resource group. The first instruction information may explicitly indicate the first reference signal resource group. For example, the first instruction information includes number information for the first reference signal group. Alternatively, the first instruction information may implicitly indicate the first reference signal resource group. For example, if there is a correspondence between multiple reference signal resources and other information, the first instruction information may include the other information. The terminal may obtain the first reference signal resource group that has a correspondence with the other information through inference based on the other information.

[0091] For example, multiple reference signal resource groups include four reference signal resource groups. The four reference signal resource groups may be specified by the protocol, preset, configured by the access network device for the terminal, or even configured by the terminal for the access network device. This is not limited to these. An example is used in which the access network device configures four reference signal resource groups for the terminal. The access network device may pre-configure four reference signal resource groups for the terminal. The access network device may then send first instruction information to the terminal, which may indicate any one of the four reference signal resource groups. For example, the numbers of the four reference signal resource groups are 1, 2, 3, and 4. In this case, when the first instruction information includes instruction information "1", the terminal may determine that the access network device is indicating the first reference signal resource group, in other words, the first reference signal resource group may be determined to be the first reference signal resource group.

[0092] As shown in Figure 10, a procedure is provided, mainly comprising a training phase and an inference phase. In the training phase (including steps 1000 to 1002 below), the access network device configures four reference signal resource groups for the terminal device, and the terminal trains an AI model based on the four reference signal resource groups configured by the access network device. In the inference phase (including steps 1003 and 1004 below), the access network device directs the terminal to one of several reference signal resource groups, which is the first reference signal resource group described above. The terminal performs model inference using the first reference signal resource group and the trained model. Based on the inference results, the terminal determines the reference signal resources that satisfy the conditions. The procedure in Figure 10 includes the following steps.

[0093] Step 1000: The access network device sends configuration information to the terminal, which is used to configure four reference signal resource groups for the terminal.

[0094] For example, the universal reference signal resource set contains 64 reference signal resources, numbered sequentially from 0 to 63. Each of the four reference signal resource groups contains 16 reference signal resources, and these 16 resources belong to the 64 reference signal resources. An example is given below: the reference signal resource numbers in the first reference signal resource group are [0, 7, 9, 14, 18, 21, 27, 28, 35, 36, 42, 45, 49, 54, 56, 63], and the reference signal resource numbers in the second reference signal resource group are [3, 4, 10, 13, 16, 23, 27, 28, 35, 36, 40, 47, 50, 53, 59]. The reference signal resource numbers included in the third reference signal resource group are [3,5,9,14,19,21,25,30,35,37,41,46,51,53,57,62], and the reference signal resource numbers included in the fourth reference signal resource group are [0,7,10,13,16,23,26,29,32,39,42,45,48,55,58,61]. Furthermore, as shown in Figure 11, another example is given where the reference signal resource numbers in the first reference signal resource group are [0,7,9,14,18,21,27,28,35,36,42,45,49,54,56,63], and the reference signal resource numbers in the second reference signal resource group are [0,3,4,7,10,13,16,18,21,23,27,28,35,36,40,45,47,50,53,56,59,60,6 3] is the reference signal resource number included in the third reference signal resource group, and the reference signal resource numbers included in the fourth reference signal resource group are [0, 3, 5, 7, 9, 14, 18, 19, 21, 25, 30, 35, 37, 41, 42, 45, 46, 51, 53, 56, 57, 62, 63], and the reference signal resource numbers included in the fourth reference signal resource group are [0, 7, 10, 13, 16, 18, 21, 23, 26, 29, 32, 39, 42, 45, 48, 55, 56, 58, 61, 63].

[0095] In one design, multiple reference signal resource groups are defined and configured on the access network side, and each group of reference signal resources is assigned an independent number. After a terminal accesses the network, the access network device configures multiple reference signal resource groups for the terminal. For example, the access network device may configure each reference signal resource group for the terminal by using a bitmap. The bitmap may contain 64 bits. The 64 bits are 0 through 63 bits, and each of the 64 bits represents one of the 64 reference signal resources in the reference signal resource universal set. A bit being 1 indicates that a corresponding reference signal resource exists for that bit. A bit being 0 indicates that a corresponding reference signal resource does not exist for that bit. Of course, the reverse is also possible. A bit being 1 indicates that a corresponding reference signal resource does not exist for that bit. A bit being 0 indicates that a corresponding reference signal resource exists for that bit. Alternatively, the terminal may define multiple reference signal resource groups and notify the access network device of these multiple reference signal resource groups, and as a result, when supporting inference, the access network device may indicate the corresponding first reference signal resource group based on the multiple reference signal resource groups. Similar to the design described above, the terminal may notify the access network device of each reference signal resource group by using a bitmap. Indeed, the procedure in Figure 10 uses an example in which the access network device configures four reference signal resource groups for the terminal.

[0096] Step 1001: The access network device sends a reference signal to the terminal.

[0097] Optionally, the reference signal may be a reference signal within the universal set. For example, the reference signal may include 64 reference signals corresponding to the 64 reference signal resources mentioned above. The terminal sweeps (or measures) the 64 reference signals and determines the measurement results for the 64 reference signals. This process is sometimes called a full codebook sweep.

[0098] Step 1002: The terminal trains the model based on the measurement results of the four reference signal resource groups.

[0099] For example, a terminal can acquire measurement results for each of 64 reference signals. Furthermore, based on the measurement results of the 64 reference signals, measurement results corresponding to each of the four reference signal resource groups are determined. The model is trained using the measurement results of the four reference signal resource groups. The model training process includes at least one round. For example, during the first round of model training, the terminal trains an initial model using the measurement results of the four reference signal resource groups. The initial model may be specified and preset in the protocol, configured by another device for the terminal, etc., but is not limited to this. For example, during training, the measurement results of the four reference signal resource groups may be input separately to the initial model. The output of the initial model is compared with a label, which is the ideal output of the model. If the difference between the output of the initial model and the label is greater than or equal to a threshold, the parameters of the initial model are adjusted. Let's take a neural network as an example. The parameters of the initial model to be adjusted include at least one of the following, namely, the number and width of layers in the neural network, the weights of neurons, the parameters of the activation function of neurons, etc. In one design, a loss function may be defined, and the difference between the model's output and the label may be determined using the loss function. During the second round of model training, the model training may continue on the model obtained through the first round of model training by using the measurement results corresponding to the four groups of reference signal resources. The second round of model training is the same as the first round of model training and will not be described again. During each round of model training, the terminal may determine whether the difference between the model's output and the label is less than a threshold. If the difference is less than the threshold, the model training stops and is completed. If the difference is greater than or equal to the threshold, the next round of model training continues.

[0100] In one design, the terminal trains the model by using a classification training method. For example, the model's inputs are measurement results corresponding to four reference signal resource groups, and the model's output is the probability that each of the predicted reference signal resources is a reference signal resource that satisfies the condition. The labels are the reference signal resources that actually satisfy the condition. During each round of model training, the terminal compares whether the model's output matches the labels. For example, if the preset reference signal resource with the highest probability of satisfying the model's condition is the same as the reference signal resource in the labels, model training stops. Otherwise, the next round of model training continues.

[0101] In one design, the terminal trains the model using a regression training method. For example, the model's inputs are measurement results corresponding to four reference signal resource groups, and the model's outputs are predicted measurement results corresponding to each reference signal resource out of all the reference signal resources. For example, the measurement results may be RSRP. The labels may be the actual measurement results corresponding to each reference signal resource. The terminal compares the predicted measurement results with the actual measurement results in the labels. If the difference between the predicted measurement results and the actual measurement results is less than a threshold, model training stops. Otherwise, model training continues.

[0102] During model training, the terminal may require multiple rounds of training to ensure the model meets the inference requirements. The conditions for meeting the inference requirements may be as follows: the model can achieve ideal inference performance for each of the four reference signal resource groups. For example, for each reference signal resource group, the difference between the predictive inference result and the label is less than a threshold.

[0103] Indeed, the above explanation uses an example where the terminal trains the model for illustrative purposes. Alternatively, the model may be trained and configured for the terminal by another device. This other device may include an access network device, a core network device, an OAM, etc. The process of training the model by another device is similar to the process of training the model by the terminal. The difference is that the terminal needs to report the measurement results corresponding to each measured reference signal resource group to the other device, and this other device performs model training based on the measurement results corresponding to each reference signal resource group. In one design, the access network device may train the model and send the trained model to the terminal. In the design, the main consideration is that the terminal may have limited computing power and may not be able to support model training.

[0104] Step 1003: The access network device transmits first instruction information to a terminal, where the first instruction information includes instruction information for a first reference signal resource group.

[0105] In one design, the access network device transmits information to the terminal indicating one of four reference signal resource groups. This reference signal resource group is referred to as the first reference signal resource group. Alternatively, the access network device determines a reference signal resource group from the four and indicates the determined reference signal resource group to the terminal. This determined reference signal resource group is referred to as the first reference signal resource group. For example, beams corresponding to reference signal resources included in the reference signal resource group determined by the access network device are not blocked, or the reference signal resource group may be determined based on other conditions, but is not limited to these. For example, the access network device may indicate the first reference signal resource group to the terminal device by using a CSI report configuration (CSI-ReportConfig), a CSI resource configuration (CSI-resourceConfig), etc. For example, CSI-ReportConfig and CSI-resourceConfig may include information indicating the first reference signal resource group.

[0106] Step 1004: The terminal determines a reference signal resource that satisfies the conditions based on the measurement results and model of the first reference signal resource group.

[0107] In one design, the terminal performs a beam sweep to determine the measurement results for each reference signal resource included in the first reference signal resource group, which is sometimes abbreviated as the measurement results for the first reference signal resource group. The terminal inputs the measurement results for the first reference signal resource group into the model. For example, the first reference signal resource group may contain 16 reference signal resources, and the terminal may determine the measurement results for these 16 reference signal resources. The model may contain 64 input ports, and the terminal inputs the 16 measurement results for the first reference signal resource group into the 16 input ports corresponding to the model. Preset values ​​may be input for the remaining 48 output ports of the model. This process is sometimes called interpolation completion. Based on the output of the model, the terminal may determine the first reference signal resource that satisfies the conditions. For example, in a classification training method, the output of the model is the probability that each reference signal resource is a reference signal resource that satisfies the conditions. The reference signal resource with the highest probability may be selected from the output of the model, and the reference signal resource with the highest probability is the first reference signal resource that satisfies the conditions. Alternatively, in regression training, the model output is the measurement result for each reference signal resource. The reference signal resource with the largest measurement result value may be selected from the model output, and the reference signal resource with the largest value is the first reference signal resource that satisfies the condition. In this description, a correspondence exists between reference signal resources and beams, and the terminal's determination of the first reference signal resource that satisfies the condition includes the terminal determining the beam that satisfies the condition. Optionally, the beam that satisfies the condition may be the terminal's optimal receiving beam. For example, when an access network device transmits downlink information to a terminal, the terminal receives the downlink information by using the optimal receiving beam.

[0108] It can be understood that, in different model inference processes, the access network device may instruct the terminal to use the same or different groups of reference signal resources. This is not limited to this. For example, during the first model estimation, the access network device instructs the terminal to use the first of four reference signal resource groups. During the second model estimation, the access network device may instruct the terminal to use the second of the four reference signal resource groups, or it may instruct the terminal to use the first reference signal resource group again. The procedure in Figure 10 uses an example in which the access network device instructs the terminal to use the first reference signal resource group during the first model estimation and then instructs the terminal to use the third reference signal resource group during the second model estimation.

[0109] According to the design described above, during model training, one model is trained on multiple reference signal resource groups, and the model converges on multiple reference signal resource groups. During model inference, the access network device instructs the terminal to select one of the multiple reference signal resource groups, and the terminal performs model inference based on the instructed reference signal resource group to determine a reference signal resource that satisfies the conditions. According to the design described above, the accuracy of model estimation can be improved.

[0110] [Embodiment 2] In one design, a terminal determining a first reference signal resource group from a plurality of reference signal resource groups includes: the terminal receiving second instruction information from an access network device, wherein the second instruction information includes instruction information for at least one reference signal resource included in the first reference signal resource group. Based on the second instruction information, the terminal determines the first reference signal resource group from the plurality of reference signal resource groups.

[0111] The difference from Embodiment 1 is that, during model estimation, the access network device instructs the terminal on the reference signal resources included in the first reference signal resource group. Since each reference signal resource group contains different reference signal resources, the terminal can determine the first reference signal resource group instructed by the access network device based on the reference signal resources instructed by the access network device. For example, the four reference signal resource groups are [0,7,9,14,18,21,27,28,35,36,42,45,49,54,56,63], [3,4,10,13,16,23,27,28,35,36,40,47,50,53,59,60], [3,5,9,14,19,21,25,30,35,37,41,46,51,53,57,62] and [0,7,10,13,16,23,26,29,32,39,42,45,48,55,58,61]. If the reference signal resources indicated to the terminal by the access network device are [0, 7, 9, 14, 18, 21, 27, 28, 35, 36, 42, 45, 49, 54, 56, 63], the terminal may determine that the access network device is indicating the first reference signal resource group. The terminal may then perform model inference using the first reference signal resource group.

[0112] As shown in Figure 12, a procedure is provided. This procedure uses an example in which an access network device configures four reference signal resource groups for a terminal, and the terminal trains a model. The procedure in Figure 12 includes a training phase and an inference phase. The training phase includes steps 1200 to 1202 below, and the inference phase includes steps 1203 and 1204 below. The procedure in Figure 12 includes the following steps.

[0113] Step 1200: The access network device sends configuration information to the terminal, which is used to configure four reference signal resource groups.

[0114] In one design, the access network device defines four reference signal resource groups. When a terminal accesses the network, the access network device may configure four reference signal resource groups for the terminal. Optionally, the four reference signal resource groups do not have corresponding numbers. When configuring four reference signal resource groups for a terminal, the access network device may configure a bitmap for each reference signal resource group. The terminal can use the bitmap to determine the reference signal resources specifically included in each reference signal resource group. Alternatively, the terminal defines four reference signal resource groups. When the terminal accesses the network, the terminal informs the access network device of the four reference signal resource groups, and as a result, the access network device instructs the terminal to use the first reference signal resource group during the inference phase. Indeed, the procedure in Figure 12 uses an example where the access network device configures four reference signal resource groups for a terminal for illustrative purposes.

[0115] In Embodiment 2, the identifiers of the reference signal resources included in each reference signal resource group may be global identifiers, and it can be understood that global identifiers may be considered identifiers of reference signal resources in the Reference Signal Resource Universal Set. For example, a reference signal resource group includes 16 reference signal resources. In the Reference Signal Resource Universal Set, the identifiers of the 16 reference signal resources are [0, 7, 9, 14, 18, 21, 27, 28, 35, 36, 42, 45, 49, 54, 56, 63], which indicates that the reference signal resource group specifically includes reference signal resources numbered 0, 7, 9, 14, 18, 21, 27, 28, 35, 36, 42, 45, 49, 54, 56, 63 in the Reference Signal Resource Universal Set. In Embodiment 1, the identifiers of the reference signal resources included in each reference signal resource group may be global identifiers, or they may be other identifiers, which may be identifiers independent of the Reference Signal Resource Universal Set. For example, in Embodiment 1, the reference signal resource group includes 16 reference signal resources, and the identifiers of the 16 reference signal resources are 0 to 15.

[0116] Step 1201: The access network device sends a reference signal to the terminal.

[0117] Optionally, the reference signal may be a reference signal within the universal set. For example, if the universal set contains 64 reference signals, the access network device may transmit the 64 reference signals separately to the terminals.

[0118] Step 1202: The terminal trains the model based on the measurement results of the four reference signal resource groups.

[0119] Step 1202 may also be replaced as follows: The access network device trains a model based on measurements from four reference signal resource groups. When a terminal accesses the network, the access network device sends the trained model to the terminal.

[0120] Step 1203: The access network device transmits a second instruction information to the terminal, wherein the second instruction information includes instruction information for at least one reference signal resource included in the first reference signal resource group.

[0121] For example, the second instruction information may include a global identifier for at least one reference signal resource included in the first reference signal resource group. Since each reference signal resource group includes a different reference signal resource, the terminal may determine a group of reference signal resources that includes the reference signal resource indicated by the access network device, based on the reference signal resource indicated by the access network device, where the reference signal resource group is the first reference signal resource group.

[0122] Step 1204: The terminal determines a reference signal resource that satisfies the conditions based on the measurement results and model of the first reference signal resource group.

[0123] Optionally, in different model inference processes, the access network device directs the terminal to reference signal resources from the same group or different groups.

[0124] According to the design described above, during the training phase, the terminal trains the model by using four reference signal resource groups. In the inference phase, the access network device sends instruction information for the reference signal resource groups to the terminal device, and the terminal device performs model inference using the instruction reference signal resource groups and the model to determine the reference signal resources that satisfy the conditions, thereby achieving a balance between AI inference performance and generalizability.

[0125] To implement the functions of the embodiments described above, it can be understood that access network devices and terminals include corresponding hardware structures and / or software modules to perform said functions. Those skilled in the art will readily recognize that, in this application, the units and steps of the methods in the examples described in relation to the embodiments disclosed herein can be implemented by hardware, or by a combination of hardware and computer software. Whether the functions are performed by hardware or by hardware driven by computer software depends on the specific application scenario and the design constraints of the technical solution.

[0126] Figures 13 and 14 show the structure of a communication device according to an embodiment of the present application. These communication devices can be configured to implement the functions of a terminal device or access network device in the embodiment of the method described above. Therefore, the effects of the embodiment of the method described above can also be achieved. In the embodiment of the present application, the communication device may be one of the terminals 120a to 120j shown in Figure 1, an access network device 110a or 110b shown in Figure 1, or a module (e.g., a chip) used within the terminal or access network device.

[0127] As shown in Figure 13, the communication device 1300 includes a processing unit 1310 and a transmitting / receiving unit 1320. The communication device 1300 is configured to implement the functions of a terminal or access network device in the embodiment of the method shown in Figure 9.

[0128] When the communication device 1300 is configured to implement the functions of a terminal in the embodiment of the method shown in Figure 9, the processing unit 1310 is configured to determine a first reference signal resource group from a plurality of reference signal resource groups and to determine a first reference signal resource based on the first reference signal resource group and the model.

[0129] When the communication device 1300 is configured to implement the functions of an access network device in an embodiment of the method shown in Figure 9, the processing unit 1310 is configured to generate configuration information, which includes configuration information for a plurality of reference signal resource groups, and the transmitting / receiving unit 1320 is configured to transmit the configuration information to the first communication device.

[0130] For a more detailed description of the processing unit 1310 and the transmitting / receiving unit 1320, please refer directly to the relevant descriptions in the embodiments of the method shown in Figures 9, 10, and 12. Further details will not be explained here.

[0131] As shown in Figure 14, the communication device 1400 includes a processor 1410 and an interface circuit 1420. The processor 1410 and the interface circuit 1420 are coupled to each other. It can be understood that the interface circuit 1420 may be a transceiver or an input / output interface. Optionally, the communication device 1400 may further include a memory 1430 configured to store instructions executed by the processor 1410, input data for the processor 1410 to execute instructions, or data generated after the processor 1410 has executed instructions.

[0132] When the communication device 1400 is configured to implement the method shown in Figure 9, the processor 1410 is configured to implement the functions of the processing unit 1310, and the interface circuit 1420 is configured to implement the functions of the transmitting / receiving unit 1320.

[0133] When the communication device is a chip used within a terminal, the chip within the terminal implements the functions of the terminal in the embodiment of the method described above. The chip within the terminal receives information from another module within the terminal (e.g., a radio frequency module or an antenna), which is then transmitted to the terminal by the access network device. Alternatively, the chip within the terminal transmits information to another module within the terminal (e.g., a radio frequency module or an antenna), which is then transmitted to the access network device by the terminal.

[0134] When the communication device is a module used within an access network device, the module in the access network device implements the functions of the access network device in the embodiments of the method described above. The module in the access network device receives information from another module in the access network device (e.g., a radio frequency module or an antenna), which is then transmitted to the access network device by a terminal. Alternatively, the module in the access network device transmits information to another module in the access network device (e.g., a radio frequency module or an antenna), which is then transmitted to a terminal by the access network device. The module in the access network device may be the baseband chip of the access network device, or it may be a DU or another module. In this specification, a DU may be a DU in an open radio access network (O-RAN) architecture.

[0135] It can be understood that the processor in the embodiments of this application may be a central processing unit (CPU), or another general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), or another programmable logic device, a transistor logic device, a hardware component, or any combination thereof. The general-purpose processor may be a microprocessor or any conventional processor.

[0136] The steps of the method in the embodiments of this application may be implemented in hardware form or by executing software instructions by a processor. The software instructions may include corresponding software modules. The software modules may be stored in random access memory, flash memory, read-only memory, programmable read-only memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, registers, hard disk drives, removable hard disk drives, CD-ROMs, or any other form of storage medium known in the art. For example, by coupling the storage medium to the processor, the processor can read information from and write information to the storage medium. Of course, the storage medium may be a component of the processor. The processor and the storage medium may be located within an ASIC. In addition, the ASIC may be located in an access network device or terminal. Of course, the processor and the storage medium may exist as separate components within the access network device or terminal.

[0137] All or part of the embodiments described above may be implemented using software, hardware, firmware, or any combination thereof. When an embodiment is implemented using software, all or part of the embodiment may be implemented in the form of a computer program product. A computer program product includes one or more computer programs or instructions. When a computer program or instruction is loaded into a computer and executed, all or part of the procedures or functions in the embodiments of this application are performed. The computer may be a general-purpose computer, a dedicated computer, a computer network, a network device, a user device, or other programmable device. The computer program or instruction may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, a computer program or instruction may be transmitted by wire or wirelessly from one website, computer, server, or data center to another website, computer, server, or data center. The computer-readable storage medium may be any available medium accessible by a computer or a data storage device that integrates one or more available media, such as a server or data center. The usable media may be magnetic media, such as floppy disks, hard disk drives, or magnetic tapes; optical media, such as digital video discs; or semiconductor media, such as solid-state drives. The computer-readable storage medium may be volatile or non-volatile storage medium, or may include two types of storage media, namely volatile and non-volatile storage media.

[0138] In the various embodiments of this application, unless otherwise stated or unless there is a logical inconsistency, the terminology and / or descriptions in different embodiments are consistent and may be referenced to one another, and the technical features in different embodiments may be combined based on their internal logical relationships to form a new embodiment.

[0139] In this application, “at least one” means one or more, and “multiple” means two or more. “And / or” describes a relationship between related objects and indicates that three relationships may exist. For example, A and / or B may indicate the following three cases: that only A exists, that both A and B exist, and that only B exists, where A and B may be singular or plural. In text descriptions in this application, the letter “ / ” generally indicates an “or” relationship between related objects. In formulas in this application, the letter “ / ” indicates a “division” relationship between related objects. “Containing at least one of A, B and C” may mean containing A, containing B, containing C, containing A and B, containing A and C, containing B and C, or containing A, B and C.

[0140] It should be understood that the various numbers in the embodiments of this application are used merely for illustrative purposes and for distinction, and are not used to limit the scope of the embodiments of this application. The sequence numbers of the processes described above do not imply an execution order, and the execution order of the processes should be determined based on the function and internal logic of the processes.

Claims

1. A method of communication, The first communication device determines a first reference signal resource group from a plurality of reference signal resource groups, The first communication device determines a first reference signal resource based on the first reference signal resource group and model, A communication method that includes this.

2. The step of determining the first reference signal resource group from the plurality of reference signal resource groups using the first communication device is: A step of receiving first instruction information from a second communication device using the first communication device, wherein the first instruction information includes instruction information for the first reference signal resource group. The first communication device determines the first reference signal resource group from the plurality of reference signal resource groups based on the first instruction information, The method according to claim 1, including the method described in claim 1.

3. The step of determining the first reference signal resource group from the plurality of reference signal resource groups using the first communication device is: The first communication device receives second instruction information from a second communication device, wherein the second instruction information includes instruction information for at least one reference signal resource included in the first reference signal resource group. The first communication device determines the first reference signal resource group from the plurality of reference signal resource groups based on the second instruction information, The method according to claim 1, including the method described in claim 1.

4. The first communication device receives configuration information from a second communication device, wherein the configuration information includes configuration information for the plurality of reference signal resource groups. The first communication device acquires the plurality of reference signal resource groups based on the configuration information of the plurality of reference signal resource groups, The method according to any one of claims 1 to 3, further comprising:

5. The step of determining the first reference signal resource based on the first reference signal resource group and the model using the first communication device is: The first communication device measures the first reference signal resource group and obtains the measurement result of the first reference signal resource group. The steps include determining the input to the model based on the measurement results of the first reference signal resource group, The first communication device determines the output of the model based on the input of the model and the model, The first communication device determines the first reference signal resource based on the output of the model, The method according to any one of claims 1 to 4, including the method described in any one of claims 1 to 4.

6. A method of communication, A second communication device generates configuration information, wherein the configuration information includes configuration information for a plurality of reference signal resource groups. The second communication device transmits the configuration information to the first communication device, A communication method that includes this.

7. The process further includes the step of transmitting first instruction information to the first communication device by the second communication device, wherein the first instruction information includes instruction information for a first reference signal resource group among the plurality of reference signal resource groups. The method according to claim 6.

8. The process further includes the step of transmitting second instruction information to the first communication device by the second communication device, wherein the second instruction information includes instruction information for at least one reference signal resource included in the first reference signal resource group, and the first reference signal resource group includes a plurality of reference signal resource groups. The method according to claim 6.

9. A communication device comprising a unit configured to perform the method described in any one of claims 1 to 5.

10. A communication device comprising a processor, wherein the processor is configured to execute instructions so that the communication device performs the method according to any one of claims 1 to 5.

11. The apparatus according to claim 10, further comprising an interface circuit, wherein the interface circuit is configured to receive a signal from another communication device other than the communication device and transmit it to the processor, or to transmit a signal from the processor to another communication device other than the communication device.

12. A communication device comprising a unit configured to perform the method described in any one of claims 6 to 8.

13. A communication device comprising a processor, wherein the processor is configured to execute instructions so that the communication device performs the method described in any one of claims 6 to 8.

14. The apparatus according to claim 13, further comprising an interface circuit, wherein the interface circuit is configured to receive a signal from another communication device other than the communication device and transmit it to the processor, or to transmit a signal from the processor to another communication device other than the communication device.

15. A computer-readable storage medium, wherein the storage medium stores a computer program or instruction, and when the computer program or instruction is executed by a communication device, the method according to any one of claims 1 to 5 or the method according to any one of claims 6 to 8 is implemented.

16. A computer program product comprising a computer program or instruction, wherein when the computer program or instruction is executed by a device, the method described in any one of claims 1 to 5 is executed, or the method described in any one of claims 6 to 8 is executed.

17. A processor, wherein the processor is coupled to a memory and executes a computer program or instruction stored in the memory, and is configured to implement the method according to any one of claims 1 to 5 or the method according to any one of claims 6 to 8.

18. A communication system comprising a first communication device and a second communication device, The first communication device is configured to implement the method described in any one of claims 1 to 5, The second communication device is configured to implement the method described in any one of claims 6 to 8. Communication system.

Images