Model control and management for wireless communications

Abstract

Techniques are described for controlling and / or managing models applied in a wireless communication system. One example of a wireless communication method includes transmitting, by a first wireless device, a first message to a second device located in the network requesting model information, and receiving, by the first wireless device, a second message in response to transmitting the first message, the second message including model description information describing one or more characteristics of the model to be used by the first wireless device.

Description

[Technical Field]

[0001] Technical Field This document relates generally to digital wireless communications. [Background technology]

[0002] background Mobile communication technologies are moving the world towards an increasingly connected and networked society. Compared to existing wireless networks, next-generation systems and wireless communication technologies will need to support a much wider range of use case characteristics and provide a more complex and sophisticated range of access requirements and flexibility.

[0003] Long Term Evolution (LTE) is a wireless communication standard for mobile devices and data terminals developed by the 3rd Generation Partnership Project (3GPP®). LTE Advanced (LTE-A) is a wireless communication standard that extends the LTE standard. The fifth-generation wireless system, known as 5G, advances the LTE and LTE-A wireless standards and is dedicated to supporting higher data rates, a large number of connections, ultra-low latency, high reliability, and other emerging business needs. Summary of the Invention [Means for solving the problem]

[0004] overview Techniques are disclosed for controlling and / or managing models applied to wireless communication systems, where the models may relate to artificial intelligence (AI) and / or machine learning (ML).

[0005] A first example of a wireless communication method includes transmitting, by a first wireless device, a first message to a second device located within the network requesting model information, and receiving, by the first wireless device, a second message in response to transmitting the first message, the second message including model description information describing one or more characteristics of a model to be used by the first wireless device.

[0006] In some embodiments, the model description information includes an identifier of the model. In some embodiments, the first message includes a field identifying a purpose for requesting the model information. In some embodiments, the first message includes any one or more of the following information: a physical cell identifier (PCI), a transmitting / receiving point (TRP) identifier, an absolute radio frequency channel number (ARFCN), a bandwidth of the cell, a subcarrier spacing of the cell, a number of beams, a direction of each of the number of beams, a number of ports, a time periodicity of the reference signal, a subband granularity, a first expected number of time opportunities to be used for input to the model, a second expected number of time opportunities to be used for output to the model, and / or a resource mapping pattern in the frequency domain. In some embodiments, the method further includes transmitting, by the first wireless device to a second device in the network, a third message requesting model deployment information, the third message including the identifier of the model; and receiving, in response to transmitting the third message, a fourth message including model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model. In some embodiments, the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model.

[0007] In some embodiments, the method further includes receiving, by the first wireless device, a failure message in response to transmitting the first message, where the failure message indicates that a second device in the network does not have the model information. In some embodiments, the first wireless device is a base station. In some embodiments, the first wireless device is a communications device. In some embodiments, the first message is transmitted by the communications device and the second message is received by the communications device using a first Non-Access Stratum (NAS) message and a second NAS message, respectively. In some embodiments, the method further includes transmitting, by the communications device, a request to obtain assistance data to the base station, and receiving, in response to transmitting the request, assistance data including a boresight direction of a reference signal when the base station transmits the reference signal and / or a beamwidth of the reference signal when the base station transmits the reference signal.

[0008] In some embodiments, the method further includes transmitting, by the communications device to the base station, a request to obtain any one or more configurations from a physical cell identifier (PCI), a transmission / reception point (TRP) identifier, an absolute radio frequency channel number (ARFCN), a bandwidth of the cell, a subcarrier spacing of the cell, a number of beams, a direction of each of the number of beams, a number of ports, a time periodicity of the reference signal, a subband granularity, a first expected number of time opportunities to be used for input to the model, a second expected number of time opportunities to be used for output to the model, and / or a resource mapping pattern in the frequency domain, and receiving, in response to transmitting the request, any one or more configurations included in the request.

[0009] A second example of a wireless communication method includes receiving, by a second device located within the network, a first message from the first wireless device requesting model information, and transmitting, by the second device in response to receiving the first message, a second message, the second message including model description information describing one or more characteristics of the model to be used by the first wireless device.

[0010] In some embodiments, the model description information includes an identifier of the model. In some embodiments, the first message includes a field identifying a purpose for requesting the model information. In some embodiments, the first message includes any one or more of the following information: a physical cell identifier (PCI), a transmitting / receiving point (TRP) identifier, an absolute radio frequency channel number (ARFCN), a bandwidth of the cell, a subcarrier spacing of the cell, a number of beams, a direction of each of the number of beams, a number of ports, a time periodicity of the reference signal, a subband granularity, a first expected number of time opportunities to be used for input to the model, a second expected number of time opportunities to be used for output to the model, and / or a resource mapping pattern in the frequency domain. In some embodiments, the method further includes receiving, by a second device in the network, from the first wireless device a third message requesting model deployment information, the third message including the identifier of the model; and transmitting, in response to receiving the third message, a fourth message including model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model.

[0011] In some embodiments, the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model. In some embodiments, the method further includes transmitting, by the second device, a failure message in response to receiving the first message, wherein the failure message indicates that the second device in the network does not have the model information. In some embodiments, the first wireless device is a base station. In some embodiments, the first wireless device is a communications device.

[0012] A third example of a wireless communication method includes transmitting, by a base station, a system information message including model information, the model information including a plurality of model description information associated with a corresponding plurality of models, each model description information describing one or more characteristics of a model to be used by a communication device.

[0013] In some embodiments, one of the plurality of model description information includes at least one identifier of one model. In some embodiments, the system information message is transmitted in a system information block (SIB) or a radio resource control (RRC) message when the communication device is in an RRC connected state. In some embodiments, the system information message includes a configuration of a plurality of resources to be used by the communication device, at least one of the plurality of resources being mapped to at least one model from the plurality of models. In some embodiments, the plurality of resources include a plurality of physical uplink control channel (PUCCH) resources or a plurality of random access channels (RACHs). In some embodiments, the method further includes transmitting assistance data to the communication device, the assistance data including a boresight direction of a reference signal when the base station transmits the reference signal and / or a beamwidth of the reference signal when the base station transmits the reference signal.

[0014] In some embodiments, the at least one identifier includes a first identifier associated with an encoder of the model, a second identifier associated with a decoder of the model, or any one of the first identifier and the second identifier. In some embodiments, the method further includes receiving, by the base station, a message from the communications device including an identifier associated with an encoder of the model used by the communications device.

[0015] A fourth example of a wireless communication method includes receiving, by a communication device, a system information message from a base station, the system information message including model information, the model information including a plurality of model description information associated with a corresponding plurality of models, each model description information describing one or more characteristics of a model to be used by the communication device.

[0016] In some embodiments, one of the plurality of model description information includes at least one identifier of one model. In some embodiments, the system information message is received in a system information block (SIB) or a radio resource control (RRC) message when the communication device is in an RRC connected state. In some embodiments, the system information message includes a configuration of a plurality of resources to be used by the communication device, at least one of the plurality of resources being mapped to at least one model from the plurality of models. In some embodiments, the plurality of resources include a plurality of physical uplink control channel (PUCCH) resources or a plurality of random access channels (RACH).

[0017] In some embodiments, the method further includes receiving assistance data, the assistance data including a boresight direction of the reference signal when the base station transmits the reference signal and / or a beamwidth of the reference signal when the base station transmits the reference signal. In some embodiments, the at least one identifier includes a first identifier associated with an encoder of the model, a second identifier associated with a decoder of the model, or any one of the first identifier and the second identifier. In some embodiments, the method further includes transmitting, by the communications device, a message to the base station including an identifier associated with the encoder of the model used by the communications device.

[0018] In yet another exemplary aspect, the foregoing methods are embodied in the form of processor-executable code and stored in a non-transitory computer-readable storage medium, the code contained in the computer-readable storage medium, when executed by a processor, causing the processor to perform the methods described in this patent document.

[0019] In yet another exemplary embodiment, a device configured or operable to perform the aforementioned method is disclosed.

[0020] These and other aspects and their implementations are described in more detail in the figures, description, and claims. The present invention provides, for example, the following. (Item 1) 1. A wireless communication method, the method comprising: transmitting, by a first wireless device, a first message to a second device located within the network requesting model information; receiving, by the first wireless device, a second message in response to transmitting the first message; Including, The method, wherein the second message includes model description information that describes one or more characteristics of a model to be used by the first wireless device. (Item 2) Item 2. The method according to item 1, wherein the model description information includes an identifier of the model. (Item 3) Item 10. The method of item 1, wherein the first message includes a field identifying a purpose for requesting the model information. (Item 4) The first message contains the following information: Physical Cell Identifier (PCI), Transmitting / Receiving Point (TRP) identifier, Absolute Radio Frequency Channel Number (ARFCN), Cell bandwidth, cell subcarrier spacing, Number of beams, the direction of each of said number of beams; Number of ports, the time periodicity of the reference signal, Subband granularity, the number of first expected time occasions used for input to the model; the number of second expected time opportunities used for output to the model; and / or Resource mapping pattern in the frequency domain Item 1. The method according to item 1, comprising any one or more of the following: (Item 5) transmitting, by the first wireless device to the second device in the network, a third message requesting model deployment information, the third message including the identifier of the model; receiving a fourth message in response to transmitting the third message, the fourth message including the model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model; Item 1, the method of claim 1 further comprising: (Item 6) Item 10. The method of item 1, wherein the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model. (Item 7) Item 10. The method of item 1, further comprising receiving, by the first wireless device, a failure message in response to transmitting the first message, the failure message indicating that the second device in the network does not have the model information. (Item 8) Item 10. The method of item 1, wherein the first wireless device is a base station. (Item 9) Item 10. The method of item 1, wherein the first wireless device is a communication device. (Item 10) Item 10. The method of item 9, wherein the first message is transmitted by the communication device and the second message is received by the communication device using a first Non-Access Stratum (NAS) message and a second NAS message, respectively. (Item 11) transmitting, by said communications device, a request to obtain assistance data to a base station; receiving, in response to transmitting the request, the assistance data including a boresight direction of a reference signal when the base station transmits the reference signal and / or a beamwidth of the reference signal when the base station transmits the reference signal; Item 10. The method of item 9, further comprising: (Item 12) by the communication device to a base station, Physical Cell Identifier (PCI), Transmitting / Receiving Point (TRP) identifier, Absolute Radio Frequency Channel Number (ARFCN), Cell bandwidth, cell subcarrier spacing, Number of beams, the direction of each of said number of beams; Number of ports, the time periodicity of the reference signal, Subband granularity, the number of first expected time occasions used for input to the model; the number of second expected time opportunities used for output to the model; and / or Resource mapping pattern in the frequency domain transmitting a request to obtain any one or more configurations from receiving, in response to transmitting the request, any one or more of the configurations included in the request; Item 10. The method of item 9, further comprising: (Item 13) 1. A wireless communication method, the method comprising: receiving, by a second device located within the network, a first message from the first wireless device requesting model information; transmitting, by the second device, a second message in response to receiving the first message; Including, The method, wherein the second message includes model description information that describes one or more characteristics of a model to be used by the first wireless device. (Item 14) Item 14. The method of item 13, wherein the model description information includes an identifier for the model. (Item 15) Item 14. The method of item 13, wherein the first message includes a field identifying a purpose for requesting the model information. (Item 16) The first message contains the following information: Physical Cell Identifier (PCI), Transmitting / Receiving Point (TRP) identifier, Absolute Radio Frequency Channel Number (ARFCN), Cell bandwidth, cell subcarrier spacing, Number of beams, the direction of each of said number of beams; Number of ports, the time periodicity of the reference signal, Subband granularity, the number of first expected time occasions used for input to the model; the number of second expected time opportunities used for output to the model; and / or Resource mapping pattern in the frequency domain Item 14. The method according to item 13, comprising any one or more of the following: (Item 17) receiving, by the second device in the network, from the first wireless device, a third message requesting model deployment information, the third message including the identifier of the model; In response to receiving the third message, transmitting a fourth message including the model deployment information, the fourth message including one or more values ​​corresponding to one or more parameters to be used by the model. Item 14. The method of item 13, further comprising: (Item 18) Item 14. The method of item 13, wherein the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model. (Item 19) Item 14. The method of item 13, further comprising transmitting, by the second device, a failure message in response to receiving the first message, the failure message indicating that the second device in the network does not have the model information. (Item 20) Item 14. The method of item 13, wherein the first wireless device is a base station. (Item 21) Item 14. The method of item 13, wherein the first wireless device is a communication device. (Item 22) 1. A wireless communication method, the method comprising: transmitting, by a base station, a system information message including model information; the model information includes a plurality of model description information associated with a corresponding plurality of models; The method, wherein each model description information describes one or more characteristics of a model to be used by the communication device. (Item 23) 23. The method of claim 22, wherein one of the plurality of model description information includes at least one identifier of a model. (Item 24) 23. The method of claim 22, wherein the system information message is transmitted in a system information block (SIB) or a radio resource control (RRC) message when the communication device is in an RRC connected state. (Item 25) the system information message includes a configuration of a plurality of resources to be used by the communication device; at least one of the plurality of resources is mapped to at least one model from the plurality of models; Item 23. The method according to item 22. (Item 26) 26. The method of claim 25, wherein the plurality of resources includes a plurality of physical uplink control channel (PUCCH) resources or a plurality of random access channels (RACH). (Item 27) transmitting assistance data to the communication device; the assistance data includes a boresight direction of the reference signal when the base station transmits the reference signal, and / or a beamwidth of the reference signal when the base station transmits the reference signal; Item 23. The method according to item 22. (Item 28) The at least one identifier is: a first identifier associated with an encoder of said model; a second identifier associated with a decoder of said model; or the first identifier and the second identifier 23. The method according to item 22, comprising any one of the following: (Item 29) receiving, by the base station, a message from the communication device including an identifier associated with an encoder of the model used by the communication device; 23. The method of claim 22, further comprising: (Item 30) 1. A wireless communication method, the method comprising: receiving, by the communication device, a system information message from a base station, the system information message including model information; the model information includes a plurality of model description information associated with a corresponding plurality of models; A method, wherein each model description information describes one or more characteristics of a model to be used by the communication device. (Item 31) Item 31. The method of item 30, wherein one of the plurality of model description information includes at least one identifier of a model. (Item 32) 31. The method of claim 30, wherein the system information message is received in a system information block (SIB) or a radio resource control (RRC) message when the communication device is in an RRC connected state. (Item 33) the system information message includes a configuration of a plurality of resources to be used by the communication device; at least one of the plurality of resources is mapped to at least one model from the plurality of models; The method according to item 30. (Item 34) Item 34. The method of item 33, wherein the plurality of resources includes a plurality of physical uplink control channel (PUCCH) resources or a plurality of random access channels (RACH). (Item 35) receiving assistance data; the assistance data includes a boresight direction of the reference signal when the base station transmits the reference signal, and / or a beamwidth of the reference signal when the base station transmits the reference signal; The method according to item 30. (Item 36) The at least one identifier is: a first identifier associated with an encoder of said model; a second identifier associated with a decoder of said model; or the first identifier and the second identifier Item 31. The method according to item 30, comprising any one of the following: (Item 37) 31. The method of claim 30, further comprising transmitting, by the communication device, a message to the base station including an identifier associated with an encoder of the model used by the communication device. (Item 38) 38. A wireless communication device comprising a processor configured to perform the method described in one or more of items 1 to 37. (Item 39) 38. A non-transitory computer-readable program storage medium having stored thereon code that, when executed by a processor, causes the processor to perform a method according to one or more of items 1 to 37. [Brief explanation of the drawings]

[0021] [Figure 1] FIG. 1 shows a block diagram of a communication system comprising a plurality of communication devices.

[0022] [Figure 2] FIG. 2 shows an example of a flow chart for requesting model information.

[0023] [Figure 3] FIG. 3 shows an example of a flowchart for transmitting model information.

[0024] [Figure 4A] FIG. 4A shows another example of a flowchart for transmitting model information.

[0025] [Figure 4B] FIG. 4B shows an example of a flow chart for receiving model information.

[0026] [Figure 5] FIG. 5 shows an example block diagram of a hardware platform that may be part of a network or communication device.

[0027] [Figure 6] FIG. 6 illustrates an example of a wireless communication including a base station (BS) and user equipment (UE) according to some implementations of the disclosed technology. DETAILED DESCRIPTION OF THE INVENTION

[0028] Detailed Description I. Introduction

[0029] Artificial intelligence / machine learning has been studied and used in various fields. There has also been some research to improve the efficiency of wireless communication systems, especially the physical layer. For example, AI / ML models can be used to improve the accuracy of channel state information (CSI). Furthermore, AI / ML models can predict channel beam information in both the spatial and time domains. Positioning, channel estimation, power saving, and mobility management are some other use cases. However, this does not provide a clear solution for how to control and manage AI models within the current architecture and signaling design of 5G wireless communication systems. This patent document proposes several technical solutions for controlling and managing AI / ML models applied to wireless communication systems.

[0030] For ease of discussion, the following terms are given with some general explanation.

[0031] The example headings of the various sections below are used to facilitate understanding of the disclosed subject matter and are not intended to limit the scope of the claimed subject matter in any way. Accordingly, one or more features of one example section may be combined with one or more features of another example section. Furthermore, although 5G terminology is used for clarity of description, the technology disclosed herein is not limited to only 5G or 5G Advance technology and may be used in wireless systems implementing other protocols. [Table 1-1] [Table 1-2] Furthermore, the AI / ML model is an exemplary scenario, and the technical solutions described in this patent document may be generalizable or applicable to any model that determines the relationship between inputs and outputs.

[0032] II. Exemplary Technical Solutions

[0033] First, the AI ​​model information may include or reference AI model description information and / or AI model deployment information. The AI ​​model description information may include information describing the characteristics of the AI ​​model used, and the characteristics of the AI ​​model included in the AI ​​model description information may include any one or more of the following: Model function / purpose (e.g., this could be one of CSI compression, beam prediction in the spatial domain, or UE positioning) ● Model Identifier (ID) and / or Version ID For two-sided models, the model ID can contain one of the following: A first ID (or generative model ID) associated with the encoder of the AI ​​model, ○ A second ID (or reconstruction model ID) associated with the decoder of the AI ​​model, ○ Primary ID and secondary ID. ●Two AI models with the same model ID may have different version IDs, which may mean at least one of the following: Two AI models with the same AI model structure but different AI model parameters ○ Because one AI model is updated / fine-tuned from another AI model, the two AI models may have some common parts of the AI ​​model structure, and the two AI models may have some common parts of the AI ​​model structure and AI model parameters. Preprocessing of AI model input For example, how to normalize data used as input for AI models For example, data should be normalized to the maximum value of the data before being fed into the AI ​​model input. ●AI model input type For example, channel measurements and / or assistance data o For example, the channel measurements can be channel impulse response in time domain / channel frequency response in frequency domain / Reference Signal Received Power (RSRP). For example, the assistance data may be the boresight direction of the reference signal when the gNB transmits the reference signal and / or the beamwidth (e.g., 3 dB beamwidth) of the reference signal when the gNB transmits the reference signal. ●AI model input shape For example, the number of dimensions of the data used for AI model input and the size of each dimension ●AI model input order For example, the order of different dimensions in the data used to input an AI model. ●The following examples explain how to understand the AI ​​model input type, AI model input shape, and AI input order. The data includes channel measurements (i.e., AI model input types) that have three dimensions (i.e., the number of dimensions and the size of each dimension). Here, the data is a three-dimensional matrix. Therefore, the AI ​​model description information should include how to construct (or order) the three-dimensional matrix. One example may be that the spatial domain, frequency domain, and time domain correspond to the first, second, and third dimensions of the data, respectively (i.e., the ordering of the different dimensions). The number (or size) of ports in the volume ● The number (or size) of frequency units in the frequency domain ● The number (or size) of time opportunities in the time domain • For example, quantization methods for AI model inputs, using a limited number of bits to represent each element of data. ●AI model output type • For example, UE location, different types of measurements (e.g., RSRP, timing information), etc. ●AI model output shape For example, the number of dimensions of the AI ​​model output and the size of each dimension ●AI model output order For example, the order in which different dimensions are included in the AI ​​model output. Post-processing of AI model output For example, a method for quantizing an AI model output, where a limited number of bits are used to represent each element of the AI ​​model output. ●AI model inference latency For example, the time / latency it takes for an AI model inference entity to perform AI model inference. ●Applicable scenarios For example, the AI ​​model can be used to indicate which physical cells and / or carrier frequencies it can only be applied to. The AI ​​model deployment information may include one or more values ​​corresponding to one or more parameters used by the AI ​​model. The one or more values ​​of the AI ​​model deployment information may relate to the model structure and / or the model parameters. The AI ​​model deployment information includes any one or more of the following: An AI model structure, which may include: Number of layers ●The network for each layer (e.g., a fully connected neural network or a convolutional neural network) ●Number of neurons in each layer ●Activation functions used in each layer ● Layer order AI model parameters (e.g., neuron values / weights within an AI model) For example, the AI ​​model deployment information may be a compiled file (e.g., a runtime binary image) that can be executed by an AI model inference entity (e.g., a UE or a network). In another example, the AI ​​model deployment information may be an interpretable file that can be further compiled by an AI model inference entity for execution, and the interpretable file may use a unified representation format that can be interpreted by different AI model inference entities. The AI ​​model reasoning entity can update the AI ​​model structure and / or the AI ​​model parameters of the AI ​​model. The AI ​​model inference entity can receive updates to the AI ​​model structure and / or AI model parameters to the legacy AI model.

[0034] The following sections describe and suggest how to perform model control and management according to which entity performs AI model inference. Network-side model ○UE side model ○2 side model

[0035] FIG. 1 shows a block diagram of a communication system including multiple communication devices. As shown in FIG. 1, a user equipment (UE) communicates with a base station (e.g., a gNB or an NG-RAN node) and a network device (e.g., a core network entity or a cloud server). The network side device (e.g., a core network entity) can also communicate with the base station. For ease of explanation, a model control entity is defined that is responsible for model storage, transfer of AI model information, data collection, etc. The model control entity may reside in the gNB, the core network entity, or the cloud server.

[0036] III. Case 1: Network-side model

[0037] In this case, the gNB / NG-RAN node determines the AI ​​model inference, and the model control entity resides in the core network entity. In this case, the gNB / NG-RAN node can be considered as the first wireless device, and the second device can be a device residing in the core network entity.

[0038] The gNB may send a request for AI model information to the model control entity. In some embodiments, the request may include information indicating the function / purpose of the requested AI model information. For example, in the case of beam prediction, CSI prediction or LoS / NLoS identification. ●In some embodiments, the request may further include any one or more of the following information (which can be used by the model control entity to determine what type of AI model the gNB requires): Physical Cell ID (PCI) ○TRP (transmission / reception point) ID ARFCN (Absolute Radio Frequency Channel Number) Cell bandwidth ○ Cell subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ○Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities to be used for the AI ​​model input (can be used for beam prediction in the time domain or CSI prediction in the time domain). The expected number of time opportunities to be used for the AI ​​model output (can be used for beam prediction in the time domain or CSI prediction in the time domain). o Resource mapping patterns in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain)

[0039] The model control entity sends a response in response to a request from the gNB. The response may indicate a failure to the request (e.g., the model control entity does not have the requested AI model information). In some embodiments, the response may include only the AI ​​model description information of the AI ​​models, and each AI model may be uniquely identified by an AI model ID. In some embodiments, in response to the gNB receiving the AI ​​model description information, the gNB may send a request to the model control entity for AI model deployment information for the AI ​​model, where the request may include the AI ​​model ID. ● In some embodiments, the response may include both AI model description information and AI deployment information for the AI ​​model, and each AI model may be uniquely identified by an AI model ID.

[0040] 2 shows an example of a flowchart for requesting model information. Operation 202 includes transmitting, by a first wireless device, a first message to a second device located within the network requesting model information. Operation 204 includes receiving, by the first wireless device, a second message in response to transmitting the first message, the second message including model description information describing one or more characteristics of a model to be used by the first wireless device.

[0041] In some embodiments, the model description information includes an identifier of the model. In some embodiments, the first message includes a field identifying a purpose for requesting the model information. In some embodiments, the first message includes any one or more of the following information: a physical cell identifier (PCI), a transmitting / receiving point (TRP) identifier, an absolute radio frequency channel number (ARFCN), a bandwidth of the cell, a subcarrier spacing of the cell, a number of beams, a direction of each of the number of beams, a number of ports, a time periodicity of the reference signal, a subband granularity, a first expected number of time opportunities to be used for input to the model, a second expected number of time opportunities to be used for output to the model, and / or a resource mapping pattern in the frequency domain. In some embodiments, the method further includes transmitting, by the first wireless device to a second device in the network, a third message requesting model deployment information, the third message including the identifier of the model; and receiving, in response to transmitting the third message, a fourth message including model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model. In some embodiments, the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model.

[0042] In some embodiments, the method further includes receiving, by the first wireless device, a failure message in response to transmitting the first message, where the failure message indicates that a second device in the network does not have the model information. In some embodiments, the first wireless device is a base station. In some embodiments, the first wireless device is a communications device. In some embodiments, the first message is transmitted by the communications device and the second message is received by the communications device using a first Non-Access Stratum (NAS) message and a second NAS message, respectively. In some embodiments, the method further includes transmitting, by the communications device, a request to obtain assistance data to the base station, and receiving, in response to transmitting the request, assistance data including a boresight direction of a reference signal when the base station transmits the reference signal and / or a beamwidth of the reference signal when the base station transmits the reference signal.

[0043] In some embodiments, the method further includes transmitting, by the communications device to the base station, a request to obtain any one or more configurations from a physical cell identifier (PCI), a transmission / reception point (TRP) identifier, an absolute radio frequency channel number (ARFCN), a bandwidth of the cell, a subcarrier spacing of the cell, a number of beams, a direction of each of the number of beams, a number of ports, a time periodicity of the reference signal, a subband granularity, a first expected number of time opportunities to be used for input to the model, a second expected number of time opportunities to be used for output to the model, and / or a resource mapping pattern in the frequency domain, and receiving, in response to transmitting the request, any one or more configurations included in the request.

[0044] 3 shows an example of a flowchart for transmitting model information. Operation 302 includes receiving, by a second device located within the network, a first message from the first wireless device requesting model information. Operation 304 includes transmitting, by the second device, a second message in response to receiving the first message, the second message including model description information describing one or more characteristics of the model to be used by the first wireless device.

[0045] In some embodiments, the model description information includes an identifier of the model. In some embodiments, the first message includes a field identifying a purpose for requesting the model information. In some embodiments, the first message includes any one or more of the following information: a physical cell identifier (PCI), a transmitting / receiving point (TRP) identifier, an absolute radio frequency channel number (ARFCN), a bandwidth of the cell, a subcarrier spacing of the cell, a number of beams, a direction of each of the number of beams, a number of ports, a time periodicity of the reference signal, a subband granularity, a first expected number of time opportunities to be used for input to the model, a second expected number of time opportunities to be used for output to the model, and / or a resource mapping pattern in the frequency domain. In some embodiments, the method further includes receiving, by a second device in the network, from the first wireless device a third message requesting model deployment information, the third message including the identifier of the model; and transmitting, in response to receiving the third message, a fourth message including model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model.

[0046] In some embodiments, the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model. In some embodiments, the method further includes transmitting, by the second device, a failure message in response to receiving the first message, wherein the failure message indicates that the second device in the network does not have the model information. In some embodiments, the first wireless device is a base station. In some embodiments, the first wireless device is a communications device.

[0047] IV. Case 2: UE-side model

[0048] In this case, the UE can be considered as the first radio device, and the second device can be a device residing in a gNB / NG-RAN node or a core network entity.

[0049] Case 2-1: The UE can provide some AI model information to the gNB / NG-RAN node

[0050] In some embodiments, the UE can download the AI ​​model from a cloud server, but the gNB does not have AI model information for the AI ​​model before any AI model information is provided by the UE.

[0051] In some embodiments, some AI model information includes only a portion of the AI ​​model description information of the corresponding AI model.

[0052] In some embodiments, the UE does not need to provide AI deployment information to the gNB.

[0053] In some embodiments, the UE may send a request to the gNB requesting assistance data (which may be useful for AI model inference at the UE), and the assistance data may include any one or more of the following: The boresight direction of the reference signal when the gNB transmits the reference signal The beam width of the reference signal when the gNB transmits the reference signal (for example, 3 dB beam width)

[0054] In some embodiments, the UE may send a request to the gNB to provide a preferred configuration (measurement / assistance data based on the preferred configuration may be used as AI model input), and the preferred configuration indicated in the request may include any one or more of the following: Physical Cell ID (PCI) ●TRP (transmission / reception point) ID ●ARFCN (Absolute Radio Frequency Channel Number) Cell bandwidth ● Cell subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ●Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities used as AI model input (which can be used for time-domain beam prediction or time-domain CSI prediction) The expected number of time opportunities used for the AI ​​model output (which can be used for beam prediction in the time domain or CSI prediction in the time domain) A resource mapping pattern in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain)

[0055] Case 2-2: The model control entity resides in a gNB / NG-RAN node

[0056] In some embodiments, the gNB may transmit the AI ​​model information in a system information message. ● System information messages may be transmitted by SIBs (System Information Blocks) in the broadcast channel. ● System information messages may be transmitted to the UE in RRC messages when the UE is in an RRC connected state. ●The system information message may only include the AI ​​model description information of the AI ​​model information. Multiple AI models may be included in the system information message, each with its corresponding AI model description information. Each AI model may be uniquely identified by an AI model ID. ● The system information message may also include a configuration of resources that can be used by the UE to request or indicate AI model information. For example, the base station may store a mapping between multiple resources and multiple AI models so that when the UE transmits data on one of the resources, the base station can determine the AI ​​model that the UE is requested to use based on the resource being used by the UE to transmit the data. The resource can be a PUCCH resource. o The resource can be a random access channel (RACH). A resource can be associated with at least one AI model. The UE may send a PUCCH / RACH to request the gNB to provide AI model information for the associated AI model.

[0057] In some embodiments, the gNB may provide assistance data to the UE (which may aid in AI model inference in the UE), where the assistance data may include any one or more of the following: The boresight direction of the reference signal when the gNB transmits the reference signal The beam width of the reference signal when the gNB transmits the reference signal (for example, 3 dB beam width)

[0058] 4A is an example flowchart for transmitting model information. Operation 402 includes transmitting, by a base station, a system information message including model information, the model information including a plurality of model description information associated with a corresponding plurality of models, each model description information describing one or more characteristics of a model to be used by a communication device.

[0059] In some embodiments, one of the plurality of model description information includes at least one identifier of one model. In some embodiments, the system information message is transmitted in a system information block (SIB) or a radio resource control (RRC) message when the communication device is in an RRC connected state. In some embodiments, the system information message includes a configuration of a plurality of resources to be used by the communication device, at least one of the plurality of resources being mapped to at least one model from the plurality of models. In some embodiments, the plurality of resources include a plurality of physical uplink control channel (PUCCH) resources or a plurality of random access channels (RACHs). In some embodiments, the method further includes transmitting assistance data to the communication device, the assistance data including a boresight direction of a reference signal when the base station transmits the reference signal and / or a beamwidth of the reference signal when the base station transmits the reference signal.

[0060] In some embodiments, the at least one identifier includes a first identifier associated with an encoder of the model, a second identifier associated with a decoder of the model, or any one of the first identifier and the second identifier. In some embodiments, the method further includes receiving, by the base station, a message from the communications device including an identifier associated with an encoder of the model used by the communications device.

[0061] 4B shows an example of a flowchart for receiving model information. Operation 452 includes receiving, by the communications device, a system information message from a base station including model information, the model information including a plurality of model description information associated with a corresponding plurality of models, each model description information describing one or more characteristics of a model to be used by the communications device.

[0062] In some embodiments, one of the plurality of model description information includes at least one identifier of one model. In some embodiments, the system information message is received in a system information block (SIB) or a radio resource control (RRC) message when the communication device is in an RRC connected state. In some embodiments, the system information message includes a configuration of a plurality of resources to be used by the communication device, at least one of the plurality of resources being mapped to at least one model from the plurality of models. In some embodiments, the plurality of resources include a plurality of physical uplink control channel (PUCCH) resources or a plurality of random access channels (RACH).

[0063] In some embodiments, the method further includes receiving assistance data, the assistance data including a boresight direction of the reference signal when the base station transmits the reference signal and / or a beamwidth of the reference signal when the base station transmits the reference signal. In some embodiments, the at least one identifier includes a first identifier associated with an encoder of the model, a second identifier associated with a decoder of the model, or any one of the first identifier and the second identifier. In some embodiments, the method further includes transmitting, by the communications device, a message to the base station including an identifier associated with the encoder of the model used by the communications device.

[0064] Case 2-3: Model control entity resides in core network entity

[0065] Case 2-3-1: AI model information is opaque to gNB / NG-RAN nodes

[0066] In this case, the UE determines / performs the AI ​​model inference, and the model control entity resides in the core network entity. The AI ​​model information is opaque to the gNB / NG-RAN node, which means that the gNB has both the AI ​​model description information and the AI ​​model deployment information of the AI ​​model. Transfer of AI model information between the gNB and the model control entity, which is similar to the description of Case 1 below (the difference is that the AI ​​model inference is performed entirely in the UE). The gNB may send a request for AI model information to the model control entity. In some embodiments, the request may include information indicating the purpose of the AI ​​model, e.g., for beam prediction, CSI prediction, or LoS / NLoS discrimination. o In some embodiments, the request may further include any one or more of the following information (which the model control entity can use to determine which AI models are required by the gNB): Physical Cell ID (PCI) ●TRP (transmission / reception point) ID ●ARFCN (Absolute Radio Frequency Channel Number) Cell bandwidth ● Cell subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ●Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities used as AI model input (which can be used for time-domain beam prediction or time-domain CSI prediction) The expected number of time opportunities used for the AI ​​model output (which can be used for beam prediction in the time domain or CSI prediction in the time domain) A resource mapping pattern in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain) The model control entity sends a response in response to a request from the gNB. The response may indicate a failure to the request (e.g., the model control entity does not have the requested AI model information). In some embodiments, the response may include only the AI ​​model description information of the AI ​​model, and each AI model may be uniquely identified by a model ID. The gNB may send a request to the model control entity for AI model deployment information for the AI ​​model, where the request may include the AI ​​model ID. ● In some embodiments, the response may include both AI model description information and AI deployment information for the AI ​​model, and each AI model may be uniquely identified by a model ID. The gNB may send a request to the model control entity for AI model deployment information for the AI ​​model, where the request may include the AI ​​model ID.

[0067] Transfer of AI model information between the gNB and the UE, which is similar to the description of Case 2-2 below. In some embodiments, the gNB may transmit AI model information in a system information message. System information messages can be transmitted by SIBs (System Information Blocks) on the broadcast channel. When the UE is in an RRC connected state, system information messages may be transmitted to the UE in RRC messages. o The system information message may only contain the AI ​​model description information of the AI ​​model information. ●Multiple AI models may be included in the system information message, and each AI model has corresponding AI model description information. Each AI model may be uniquely identified by an ID. o The system information message may also include the configuration of resources that can be used by the UE to request AI model information. ● The resource may be a PUCCH resource. ● The resource may be a random access channel (RACH). A resource may be associated with at least one AI model. The UE may send a PUCCH / RACH to request the gNB to provide AI model deployment information for the relevant AI model.

[0068] In some embodiments, the gNB may provide assistance data to the UE (which may aid in AI model inference in the UE), where the assistance data may include any one or more of the following: The boresight direction of the reference signal when the gNB transmits the reference signal The beam width of the reference signal when the gNB transmits the reference signal (for example, 3 dB beam width)

[0069] Case 2-3-2: AI model information is transparent to gNB / NG-RAN nodes

[0070] In this case, the UE determines / performs the AI ​​model inference, and the model control entity resides in the core network entity. The gNB does not have AI model information of the AI ​​model before receiving any AI model information from the UE. The UE may send a request for AI model information to the model control entity. o In some embodiments, the request may include information indicating the purpose of the AI ​​model, e.g., in the case of beam prediction, CSI prediction or LoS / NLoS discrimination In some embodiments, the request may further include any one or more of the following information (which can be used by the model control entity to determine what kind of AI model is required by the UE): Physical Cell Identifier (PCI) ●Transmitting / receiving point identifier (TRP ID) Absolute Radio Frequency Channel Number (ARFCN) Cell bandwidth ● Cell subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ●Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities used as AI model input (which can be used for time-domain beam prediction or time-domain CSI prediction) The expected number of time opportunities used for the AI ​​model output (which can be used for beam prediction in the time domain or CSI prediction in the time domain) A resource mapping pattern in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain) ● The model control entity sends a response in response to a request from the UE. The response may indicate a failure to the request (e.g., the model control entity does not have the requested AI model information). o In some embodiments, the response may include only the AI ​​model description information of the AI ​​model, and each AI model may be uniquely identified by an AI model ID. The UE may send a request to the model control entity for AI model deployment information of the AI ​​model, and the request may include the AI ​​model ID. In some embodiments, the response may include both AI model description information and AI deployment information for the AI ​​model, and each AI model may be uniquely identified by a model ID. In some embodiments, the request and response are carried by NAS (Non-Access Stratum) messages between the UE and the model control entity. In some embodiments, the UE can download an AI model from the model control entity. However, the gNB does not have AI model information for the previous AI model provided by the UE, which is similar to case 2-1 below. In some embodiments, the AI ​​model information includes only a portion of the AI ​​model description information of the corresponding AI model. In some embodiments, the UE does not need to provide AI deployment information to the gNB, or the UE decides not to provide AI deployment information to the gNB. In some embodiments, the UE may send a request to the gNB to obtain assistance data (which may be useful for AI model inference at the UE), and the assistance data may include any one or more of the following: The boresight direction of the reference signal when the gNB transmits the reference signal The beam width of the reference signal when the gNB transmits the reference signal (for example, 3 dB beam width) In some embodiments, the UE may send a request to the gNB to provide a preferred configuration (measurements / assistance data based on the preferred configuration may be used as AI model input), and the preferred configuration indicated in the request may include any one or more of the following: Physical Cell ID (PCI) ●TRP (transmission / reception point) ID ●ARFCN (Absolute Radio Frequency Channel Number) Cell bandwidth ● Cell subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ●Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities used as AI model input (which can be used for time-domain beam prediction or time-domain CSI prediction) The expected number of time opportunities used for the AI ​​model output (which can be used for beam prediction in the time domain or CSI prediction in the time domain) A resource mapping pattern in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain)

[0071] Case 2-3-3: Partial AI model information is transparent to the gNB (or NG-RAN node) In this case, the UE determines / performs AI model inference, and the model control entity resides in the core network entity. Partial AI model information being transparent to the gNB means that the gNB has AI model description information of the AI ​​model, but the gNB does not have AI model information of the AI ​​deployment information of the AI ​​model. The gNB may send a request to the model control entity for AI model description information. o In some embodiments, the request may include information indicating the purpose of the AI ​​model, e.g., in the case of beam prediction, CSI prediction or LoS / NLoS discrimination In some embodiments, the request may further include any one or more of the following information (which can be used by the model control entity to determine what kind of AI model is required by the gNB): Physical Cell ID (PCI) ●TRP (transmission / reception point) ID ●ARFCN (Absolute Radio Frequency Channel Number) Cell bandwidth ● Cell subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ●Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities used as AI model input (which can be used for time-domain beam prediction or time-domain CSI prediction) The expected number of time opportunities used for the AI ​​model output (which can be used for beam prediction in the time domain or CSI prediction in the time domain) A resource mapping pattern in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain) ●The model control entity sends a response in response to a request from the gNB. The response may indicate a failure to the request (e.g., the model control entity does not have the requested AI model). o In some embodiments, the response may include only the AI ​​model description information of the AI ​​model, and each AI model may be uniquely identified by a model ID. In some embodiments, the UE may obtain the AI ​​model information via one of the following methods: The gNB can instruct the UE to download AI model deployment information from the model control entity. The instructions may include either a model ID of the AI ​​model or AI model description information. The UE may send a response to the gNB indicating that the requested AI model has been obtained by the UE. The gNB can request the model control entity to provide the UE with AI model deployment information. The instructions may include either a model ID of the AI ​​model or AI model description information. ●The model control entity may send a response to the gNB indicating which AI model has been provided to the UE. The UE can request the model control entity to provide AI model information. In the above method, the transfer of AI model information between the model control entity and the UE is similar to the description of Case 2-3-2 below. The UE may send a request for AI model information to the model control entity. In some embodiments, the request may include information indicating the purpose of the AI ​​model, e.g., in the case of beam prediction, CSI prediction or LoS / NLoS discrimination. In some embodiments, the request may further include any one or more of the following information (which can be used by the model control entity to determine what type of AI model is required by the UE): Physical Cell ID (PCI) ○TRP (transmission / reception point) ID ARFCN (Absolute Radio Frequency Channel Number) Cell bandwidth ○ Cell subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ○Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities to be used for the AI ​​model input (can be used for beam prediction in the time domain or CSI prediction in the time domain). The expected number of time opportunities to be used for the AI ​​model output (can be used for beam prediction in the time domain or CSI prediction in the time domain). o Resource mapping patterns in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain) ● The model control entity sends a response in response to a request from the UE. The response may indicate a failure to the request (e.g., the model control entity does not have the requested AI model information). The response may include only the AI ​​model description information of the AI ​​model, and each AI model may be uniquely identified by an AI model ID. In some embodiments, the response may include only the AI ​​model description information of the AI ​​model, and each AI model may be uniquely identified by a model ID. The UE may send a request to the model control entity for AI model deployment information for the AI ​​model, and the request may include the AI ​​model ID. ● In some embodiments, the response may include both AI model description information and AI deployment information for the AI ​​model, and each AI model may be uniquely identified by an AI model ID. In some embodiments, the request and response are carried by NAS (Non-Access Stratum) messages between the UE and the model control entity.

[0072] In some embodiments, the gNB may provide assistance data to the UE (which may aid in AI model inference in the UE), where the assistance data may include any one or more of the following: The boresight direction of the reference signal when the gNB transmits the reference signal The beam width of the reference signal when the gNB transmits the reference signal (for example, 3 dB beam width)

[0073] V.Case 3: 2-side model

[0074] Case 3-1: The UE can provide some AI model information to the gNB / NG-RAN node

[0075] In some embodiments, the UE can download the encoder portion of the AI ​​model from a cloud server, but the gNB does not have the AI ​​model information of the AI ​​model before any AI model information is provided by the UE.

[0076] In some embodiments, some AI model information includes only a portion of the AI ​​model description information corresponding to an encoder portion of the AI ​​model, and each encoder portion may be uniquely identified by an ID. In some embodiments, the UE does not need to provide the AI ​​deployment information to the gNB.

[0077] In some embodiments, the UE may send a request to the gNB requesting assistance data (which may be useful for AI model inference at the UE), and the assistance data may include any one or more of the following: The boresight direction of the reference signal when the gNB transmits the reference signal The beam width of the reference signal when the gNB transmits the reference signal (for example, 3 dB beam width)

[0078] In some embodiments, the UE may send a request to the gNB to provide a preferred configuration (measurements based on the preferred configuration may be used as AI model input), and the preferred configuration indicated in the request may include any one or more of the following: Physical Cell ID (PCI) ●TRP (transmission / reception point) ID ●ARFCN (Absolute Radio Frequency Channel Number) Cell bandwidth and subcarrier spacing Number of beams (e.g., number of CSI-RS resources in a CSI-RS resource set) The direction of each beam (e.g., the direction can be the boresight direction of the beam) Number of ports (e.g., number of ports for CSI-RS resources) ●Time periodicity of the reference signal Subband granularity (e.g., the number of resource blocks contained in a subband) The expected number of time opportunities used as AI model input (which can be used for time-domain beam prediction or time-domain CSI prediction) The expected number of time opportunities used for the AI ​​model output (which can be used for beam prediction in the time domain or CSI prediction in the time domain) A resource mapping pattern in the frequency domain (e.g., to indicate which resource elements carry reference signals, which may be used for channel estimation or CSI prediction in the frequency domain)

[0079] In some embodiments, the gNB may indicate which AI model (or encoder portion) should be used for AI model inference at the UE, where the indication may include at least an AI model ID to the encoder portion.

[0080] Case 3-2: gNB / NG-RAN node can provide some AI model information to UE

[0081] In some embodiments, the gNB may transmit the AI ​​model information in a system information message. ● System information messages may be transmitted by SIBs (System Information Blocks) in the broadcast channel. ● System information messages may be transmitted to the UE in RRC messages when the UE is in an RRC connected state. ●The system information message may only include the AI ​​model description information of the AI ​​model information. Multiple AI models may be included in the system information message, with each AI model having its corresponding AI model description information. Each AI model may be uniquely identified by an AI model ID. In some embodiments, the AI ​​model ID may be one of the following: ●Encoder part ID ●Decoder part ID ● {Encoder part ID, Decoder part ID} pair

[0082] In some embodiments, the system information message may also include a configuration of resources that may be used by the UE to request AI model information. ● The resource may be a PUCCH resource. ● The resource may be a random access channel (RACH). A resource may be associated with at least one AI model. The UE may transmit a PUCCH / RACH to request the gNB to provide AI model information for the encoder portion of the AI ​​model.

[0083] In some embodiments, the UE may inform the gNB (e.g., via an encoder portion ID) which encoder portion was used for the AI ​​model inference.

[0084] Case 3-3: Model control entity resides in core network entity

[0085] In the absence of conflicts, the transfer of AI model information related to the decoder portion of the AI ​​model occurring between the model control entity and the gNB can reuse the procedure of Case 1 for the network-side model.

[0086] In the absence of conflicts, the transfer of AI model information related to the encoder portion of the AI ​​model occurring between the model control entity and the gNB or between the model control entity and the UE can reuse the procedures of Cases 2-3 for the UE-side model.

[0087] FIG. 5 shows an example block diagram of a hardware platform 500 that may be part of a network device (e.g., a base station) or a communication device (e.g., user equipment (UE)). The hardware platform 500 includes at least one processor 510 and a memory 505 having instructions stored thereon. The instructions, when executed by the processor 510, configure the hardware platform 500 to perform the operations described in FIGS. 1-4B and various embodiments described in this patent document. The transmitter 515 transmits or sends information or data to another device. For example, a network device transmitter can send a message to a user equipment. The receiver 520 receives information or data transmitted or sent by another device. For example, a user equipment can receive a message from a network device.

[0088] The above-described embodiments apply to wireless communications. Figure 6 illustrates an example of a wireless communications system (e.g., a 5G or NR cellular network) including a base station 620 and one or more user equipments (UEs) 611, 612, and 613. In some embodiments, the UE accesses a BS (e.g., a network) using a communication link to the network (indicated by dashed arrows 631, 632, and 633, sometimes referred to as the uplink direction), which then enables subsequent communication from the BS to the UE (indicated by arrows 641, 642, and 643, sometimes referred to as the downlink direction, which is shown in the network-to-UE direction). In some embodiments, the BS transmits information to the UE (indicated by arrows 641, 642, and 643, sometimes referred to as the downlink direction), which then enables subsequent communication from the UE to the BS (indicated by dashed arrows 631, 632, and 633, sometimes referred to as the uplink direction, which is shown in the UE-to-BS direction). The UE may be, for example, a smartphone, a tablet, a mobile computer, a machine-to-machine (M2M) device, an Internet of Things (IoT) device, etc.

[0089] The term "exemplary" is used herein to mean "an example" and does not refer to an ideal or preferred embodiment, unless expressly stated otherwise.

[0090] Some of the embodiments described herein are described in the general context of a method or process that may be implemented in one embodiment by a computer program product embodied in a computer-readable medium containing computer-executable instructions, such as program code, executed by computers in a network environment. Computer-readable media may include removable and non-removable storage devices, including, but not limited to, read-only memory (ROM), random access memory (RAM), compact discs (CDs), digital versatile discs (DVDs), and the like. Thus, computer-readable media may include non-transitory storage media. Generally, program modules may include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Computer- or processor-executable instructions, associated data structures, and program modules represent examples of program code for executing steps of the methods disclosed herein. The particular sequence of such executable instructions or associated data structures represents examples of corresponding acts for implementing the functions described in such steps or processes.

[0091] Some of the disclosed embodiments may be implemented as devices or modules using hardware circuits, software, or a combination thereof. For example, a hardware circuit implementation may include discrete analog and / or digital components integrated, for example, as part of a printed circuit board. Alternatively or additionally, the disclosed components or modules may be implemented as application-specific integrated circuits (ASICs) and / or field-programmable gate array (FPGA) devices. Some implementations may also or alternatively include a digital signal processor (DSP), which is a dedicated microprocessor having an architecture optimized for the operational needs of digital signal processing associated with the disclosed functionality of the present application. Similarly, various components or subcomponents within each module may be implemented in software, hardware, or firmware. Connections between modules and / or components within a module may be provided using any one of the connection methods and mediums known in the art, including, but not limited to, communication via the Internet, wired, or wireless networks using appropriate protocols.

[0092] While this document contains many details, these should not be construed as limitations on the scope of the claimed invention or the invention that may be claimed, but rather as descriptions of features specific to particular embodiments. Certain features described herein in the context of separate embodiments may also be implemented in combination in a single embodiment. Conversely, various features described in the context of a single embodiment may also be implemented in multiple embodiments separately or in any suitable subcombination. Furthermore, while features may be described above as acting in a particular combination and initially claimed as such, one or more features from a claimed combination may, in some cases, be excised from the combination, and the claimed combination may be directed to a subcombination or a variation of the subcombination. Similarly, while operations are shown in the figures in a particular order, this should not be understood as requiring such operations to be performed in the particular order shown, or in any sequential order, or that all of the operations shown be performed, to achieve desirable results.

[0093] Based on what has been described and illustrated in this disclosure, only some implementations and examples have been described, and other implementations, enhancements, and variations may be made.

Claims

1. A wireless communication method, the wireless communication method comprising: a first wireless device transmitting a first message requesting model information to a second device located in a network, the first message including one or more of: a number of expected time opportunities to be used for model input for beam prediction in the time domain or channel state information (CSI) prediction in the time domain; a number of expected time opportunities to be used for model output for beam prediction in the time domain or CSI prediction in the time domain; or a resource mapping pattern in the frequency domain indicating resource elements that will transmit reference signals, resource elements used for channel estimation, or resource elements for CSI prediction; receiving, by the first wireless device, a second message in response to transmitting the first message; Including, The second message includes model description information that describes one or more characteristics of a model to be used by the first wireless device.

2. The wireless communication method of claim 1 , wherein the model description information includes an identifier for the model.

3. The method of claim 1 , wherein the first message includes a field identifying a purpose for requesting the model information.

4. The wireless communication method comprises: transmitting a third message from the first wireless device to the second device in the network requesting model deployment information, the third message including the identifier of the model; and receiving a fourth message in response to transmitting the third message, the fourth message including the model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model; The wireless communication method of claim 2 , further comprising:

5. The method of claim 1 , wherein the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model.

6. The wireless communication method of claim 1, further comprising the first wireless device receiving a fault message in response to transmitting the first message, the fault message indicating that the second device in the network does not have the model information.

7. A wireless communication method, the wireless communication method comprising: a second device located in a network receiving from a first wireless device a first message requesting model information, the first message including one or more of: a number of expected time opportunities to be used for model input for beam prediction in the time domain or channel state information (CSI) prediction in the time domain; a number of expected time opportunities to be used for model output for beam prediction in the time domain or CSI prediction in the time domain; or a resource mapping pattern in the frequency domain indicating resource elements that will carry a reference signal, resource elements used for channel estimation, or resource elements for CSI prediction; the second device transmitting a second message in response to receiving the first message; Including, The second message includes model description information that describes one or more characteristics of a model to be used by the first wireless device.

8. The wireless communication method of claim 7 , wherein the model description information includes an identifier of the model.

9. The method of claim 7 , wherein the first message includes a field identifying a purpose for requesting the model information.

10. The wireless communication method comprises: receiving, by the second device in the network, a third message from the first wireless device requesting model deployment information, the third message including the identifier of the model; transmitting a fourth message containing the model deployment information, the fourth message including one or more values ​​corresponding to one or more parameters to be used by the model, in response to receiving the third message; The wireless communication method of claim 8, further comprising:

11. The method of claim 7 , wherein the second message includes model deployment information including one or more values ​​corresponding to one or more parameters to be used by the model.

12. The wireless communication method described in claim 7, further comprising the second device transmitting a fault message in response to receiving the first message, the fault message indicating that the second device in the network does not have the model information.

13. A wireless communication device comprising a processor configured to perform the wireless communication method according to any one of claims 1 to 6.

14. 10. A non-transitory computer readable program storage medium having stored thereon code that, when executed by a processor, causes the processor to perform a wireless communication method according to any one of claims 1 to 6.

15. A wireless communication device comprising a processor configured to execute the wireless communication method of any one of claims 7 to 12.

16. A non-transitory computer-readable program storage medium having stored thereon code which, when executed by a processor, causes the processor to perform a wireless communication method described in any one of claims 7 to 12.

Images