Transmission of artificial intelligence network management models in wireless communication systems

Abstract

The present disclosure is generally directed to methods and network devices for assisting intelligent management of wireless communication systems, and in particular to processes for transmitting and receiving artificial intelligence (AI) network management models between various network devices. Depending on the functionality of the AI model, the end point of the AI model transmission may be a wireless terminal device and a wireless base station communicating via an air interface. Alternatively, the end point of the AI model transmission may be a wireless terminal device and a core network device of a wireless communication network via a wireless base station. Within an air interface between a wireless terminal and a wireless base station, the AI model may be transmitted using signaling radio bearer (SRB) or data radio bearer (DRB) resources. These resources may be based on shared SRB or DRB resources, whose configuration is adapted to accommodate AI model transmissions. Alternatively, these resources may be configured as dedicated radio resources for transmission of the AI model. Transmission for an AI model between a core network node and an intermediate wireless base station may be implemented in a control plane (CP) or a user plane (UP) of the wireless network. Various messaging schemes for resource allocation for AI model transmissions may be implemented accordingly.

Description

Technical Field

[0001] The present disclosure generally relates to methods and network devices for assisting in intelligent management of wireless communication systems, and more particularly to processes for transmitting and receiving artificial intelligence (AI) network management models between various network devices. Background Art

[0002] In wireless communication systems, the determination of adaptive network configurations, particularly adaptive network configurations with respect to air communication interfaces and radio resources, may require lengthy measurement processes and / or a large amount of computing power. Such types of configurations may include, but are not limited to, beam management, channel state information (CSI) feedback compression and decompression schemes, and wireless terminal positioning management. The correlation between various network conditions and these adaptive configurations can be learned via artificial intelligence (AI) techniques and models. Such AI models trained and managed on one network device may need to communicate to another network device via an air interface and potentially other network interfaces. Although these AI models are used for control purposes, they are very different from typical control information in that they can contain a large amount of data (e.g., a large number of hyperparameters and actual model parameters) and require a large amount of air communication resources. Summary of the Invention

[0003] The present disclosure generally relates to methods and network devices for assisting in the intelligent management of wireless communication systems, and in particular to a process for transmitting and receiving artificial intelligence (AI) network management models between various network devices. Depending on the function of the AI model, the endpoints of the AI model transmission may be wireless terminal devices and wireless base stations communicating via an air interface. Alternatively, the endpoints of the AI model transmission may be wireless terminal devices and core network devices of a wireless communication network via wireless base stations. Within the air interface between the wireless terminal and the wireless base station, the AI model may be transmitted using signaling radio bearers (SRBs) or data radio bearers (DRBs). These resources may be based on shared SRB or DRB resources, the configuration of which is adapted to accommodate AI model transmission, or alternatively, these resources may be configured as dedicated radio resources for the transmission of the AI model. The transmission of the AI model between the core network node and the intermediate wireless base station may be implemented in the control plane (CP) or user plane (UP) of the wireless network. Various messaging schemes for resource allocation for AI model transmission may be implemented accordingly.

[0004] In one embodiment, a method is disclosed in which a first device communicates with a second device via an air interface in a wireless access network of a wireless communication system. The first device and the second device are respectively one of a wireless terminal and the other of a wireless base station. The method includes: performing a request-response process with the second device on the air interface to initiate transmission of an artificial intelligence (AI) network management model therebetween; determining a radio bearer resource in the air interface for transmission of the AI network management model based on the request-response process and the fact that a core network of the wireless communication system does not respond to the request-response process; and interacting with the second device to transmit the AI network management model via the radio bearer resource.

[0005] In the above embodiment, the request-response process includes transmitting a request for the AI network management model in the form of a Radio Resource Control (RRC) message or a Media Access Control (MAC) Control Element (MAC-CE).

[0006] In any of the above embodiments, when dedicated radio bearer resources in the air interface have not been previously allocated for the purpose of transmission of the AI network management model, the request may be transmitted in the form of an RRC message.

[0007] In any of the above embodiments, the request may be transmitted in the form of a MAC-CE when dedicated radio bearer resources in the air interface have been previously allocated for the purpose of AI network management model transmission.

[0008] In any of the above embodiments, the request may be transmitted in the form of a MAC-CE, and a scheduling request (SR) may be triggered when the MAC CE is suspended and no Physical Uplink Shared Channel (PUSCH) resources are available to accommodate the MAC CE.

[0009] In any of the above embodiments, the request may include an RRC reconfiguration message for allocating new radio bearer resources by the wireless base station for transmission of the AI network management model.

[0010] In any of the above embodiments, the request may be transmitted in the form of an RRC message, the RRC message including at least one of the following: a first identifier for identifying an AI-based function associated with an AI network management model; a second identifier for indicating the AI network management model; a third identifier for identifying a radio bearer resource for transmission of the AI network management model; or a procedure identifier for identifying transmission of the AI network management model.

[0011] In any of the above embodiments, the request may be transmitted in the form of a MAC-CE, which includes at least one of the following: a first identifier for identifying an AI-based function associated with an AI network management model; a second identifier for indicating the AI network management model; a third identifier for identifying a radio bearer resource used for transmission of the AI network management model; a fourth identifier for identifying a PDU session used for transmission of the AI network management model; or a quality of service (QoS) flow identifier (QFI) for indicating a QoS flow used for transmission of the AI network management model.

[0012] In any of the above embodiments, the radio bearer resources may include signaling radio bearer (SRB) resources or data radio bearer (DRB) resources. The DRB resources may be allocated by the wireless base station for exclusive use in AI network management model transmission, as indicated by a flag in a configuration associated with the DRB resources, an SDAP (Service Data Adaptation Protocol) configuration associated with the DRB resources, a PDCP (Packet Data Convergence Protocol) configuration associated with the DRB resources, or an RLC (Radio Link Control) configuration associated with the DRB resources. The DRB resources may be allocated for additional uses other than AI network management model transmission, and the use of the DRB resources for AI network management model transmission is indicated by a flag in the header of an SDAP, PDCP, or RLC PDU; or an AI model transmission specific QFI list in the SDAP configuration associated with the DRB resources.

[0013] In another embodiment, a method for a first device to communicate with a second device via an air interface in a wireless access network of a wireless communication system is disclosed. The first device and the second device are one and the other of a wireless terminal and a wireless base station, respectively. The method may include: interacting with the second device via the air interface in a request process for transmitting at least one AI network management model, the request process being configured to trigger a data tunnel application process between the wireless base station and a core network (CN) node of the wireless communication system, and triggering allocation of radio communication resources in the air interface according to the data tunnel application process; and interacting with the second device to transmit at least one AI network management model via the radio communication resources.

[0014] In the above embodiment, the request process may include transmitting a request for transmission of at least one AI network management model in the form of an RRC message or a MAC-CE. The request may be transmitted from the wireless terminal to the wireless base station and include a non-access stratum (NAS) message from the wireless terminal to the core network node for triggering the data tunnel application process.

[0015] In some of the above embodiments, the request may be transmitted in the form of an RRC message, where the RRC message includes at least one of the following: a first identifier for identifying an AI-based function associated with at least one AI network management model; a second identifier for indicating at least one AI network management model; or an endpoint IP address of a mobile terminal.

[0016] In any of the above embodiments, the request may be transmitted in the form of a MAC CE, and the SR is triggered when the MAC CE is pending and no PUSCH resources are available to accommodate the MAC CE.

[0017] In any of the above embodiments, the request may be transmitted in the form of a MAC CE, which includes at least one of the following: a first identifier for identifying an AI-based function associated with at least one AI network management model; a second identifier for indicating at least one AI network management model; a DRB identifier for identifying a DRB resource for transmitting at least one AI network management model; a PDU session identifier for identifying a PDU session for transmitting at least one AI network management model; or a QFI for indicating a QoS flow used to transmit at least one AI network management model.

[0018] In any of the above embodiments, when data tunnel resources have not been previously allocated in the air interface for the purpose of AI network management model transmission, the request may be transmitted in the form of an RRC message.

[0019] In any of the above embodiments, when at least one data tunnel resource has been previously configured in the air interface for the purpose of transmitting the AI network management model, the request may be transmitted in the form of a MAC-CE. The at least one data tunnel resource may include a DRB resource, a QoS flow of a PDU session, or a PDU session for transmitting the AI network management model.

[0020] In any of the above embodiments, the data tunnel application process may include transmitting a data tunnel request for one or more QoS flows for a PDU session from the wireless base station to the CN node. The data tunnel request may include at least one of the following: a first identifier indicating a mobile terminal-specific Access Management Function (AMF) ID relative to an NG Application Protocol (NGAP) interface; a second identifier indicating a mobile terminal-specific Radio Access Network (RAN) ID relative to the NGAP interface; a PDU session identifier associated with the requested QoS flow; the number of requested QoS flows for each PDU session; the number of requested AI network management models; an endpoint IP address for model transmission; the size of each requested AI network management model; or a QoS request. A response to the data tunnel request from the CN node to the wireless base station may include at least one of the following: a first identifier; a second identifier; a PDU session identifier; a third identifier indicating a QoS flow allocation for each requested AI network management model; a fourth identifier indicating a failed QoS flow allocation for each requested AI network management model; or at least one reason for a failed QoS flow allocation.

[0021] In any of the above embodiments, interacting with the second device includes transmitting or receiving an RRC reconfiguration for mapping one or more QoS flows to at least one AI network management model via at least one of: an AI model transmission specifying a list of QoS flows; a configuration of an SDAP PDU header; or an enable flag for mapping a current QFI list in the SDAP configuration to transmission of at least one AI network management model.

[0022] In any of the above embodiments, the data tunnel application process may include transmitting a data tunnel request for one or more PDU sessions from the wireless base station to the CN node. The data tunnel request may include at least one of the following: a first identifier for indicating that the mobile terminal refers to a specific AMF ID relative to the NGAP interface; a second identifier for indicating a specific RAN ID of the mobile terminal relative to the NGAP interface; an endpoint address for model transmission; the number of one or more PDU sessions; the size of each requested AI network management model; a mobile terminal identifier; and a QoS request. The response to the data tunnel request from the CN node to the wireless base station may include at least one of the following: a first identifier; a second identifier; a PDU session identifier for model transmission; a third identifier for indicating a failed PDU session allocation for each requested AI network management model; or at least one reason for a failed PDU session allocation.

[0023] In any of the above embodiments, interacting with the second device may include transmitting or receiving an RRC reconfiguration for mapping one or more PDU sessions to at least one AI network management model via at least one of: a first flag in a DRB configuration, the first flag indicating that the DRB is for AI network management model transmission; a second flag in an SDAP configuration, the second flag indicating that the corresponding DRB is for AI network management model transmission; a third flag in a PDCP configuration, the third flag indicating that the corresponding DRB is for AI network management model transmission; or a fourth flag in an RLC configuration, the fourth flag indicating that the corresponding DRB is for AI network management model transmission.

[0024] In another embodiment, a method is disclosed for interacting with a core network (CN) node of a wireless communication system to process an application process for transmitting an AI network management model between the first device and a second device, performed by a first device of a radio access network (RAN) of a wireless communication system. The first device and the second device are one and the other of a mobile terminal and a wireless base station, respectively. The method may include: interacting with the second device and the CN node to perform an application process, the application process involving a first request-response process between the mobile terminal and the wireless base station and a second request-response process between the CN node and the wireless base station via an air interface of the RAN; and triggering a data tunnel application process between the UE and the CN node to facilitate the transmission of the AI network management model between the first device and the second device.

[0025] In some other embodiments, an electronic device including a processor and a memory is disclosed, wherein the processor is configured to read computer code from the memory to implement any one of the above methods.

[0026] In yet other embodiments, a computer program product is disclosed that includes a non-transitory computer-readable program medium having computer code stored thereon. The computer code, when executed by a processor, may cause the processor to implement any of the above methods.

[0027] Other aspects and alternatives to the above-described embodiments and their implementations are described in more detail in the drawings, the description, and the following claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 An example wireless communication network including a radio access network, a core network, and a data network is shown.

[0029] Figure 2 An example wireless access network is shown that includes a plurality of mobile stations / terminals or UEs and wireless access network nodes that communicate with each other via an over-the-air radio communication interface.

[0030] Figure 3 An example core network is shown.

[0031] Figure 4 An example communication protocol stack in a wireless access network node or a wireless terminal device including various network layers is shown.

[0032] Figure 5 An example flow chart for transmitting an AI model between a mobile station / terminal and a base station without core network involvement is shown.

[0033] Figure 6 An example flow chart for transmitting an AI model between a mobile station / terminal and a base station with the involvement of a core network is shown.

[0034] Figure 7 Another example flow chart for transmitting an AI model between a mobile station / terminal and a base station with the involvement of a core network is shown.

[0035] Figure 8 An example flow chart for transmitting an AI model between a mobile station / terminal and a core network based on the LTE Positioning Protocol (LPP) is shown.

[0036] Figure 9 Another example flow chart for transmitting an AI model between a mobile station / terminal and a core network based on LPP is shown.

[0037] Figure 10 Another example flow chart for transmitting an AI model between a mobile station / terminal and a core network based on LPP is shown.

[0038] Figure 11 An example flow chart for transferring AI models between a mobile station / terminal and a core network in a control plane is shown.

[0039] Figure 12 Another example flow diagram for transferring an AI model between a mobile station / terminal and a core network in a control plane is shown.

[0040] Figure 13 An example flow chart for transferring AI models between a mobile station / terminal and a core network in a user plane is shown.

[0041] Figure 14 Another example flow diagram for transferring an AI model between a mobile station / terminal and a core network in a user plane is shown. DETAILED DESCRIPTION

[0042] The techniques and examples of the embodiments and / or examples described in this disclosure can be used to facilitate the transmission and reception of artificial intelligence (AI) network management modules between various wireless network devices or nodes via at least one air interface. In this disclosure, the term "over-the-air interface" is used interchangeably with "air interface" or "radio interface". The term "exemplary" is used to mean "an example of..." and, unless otherwise specified, does not imply an ideal or preferred example, embodiment, or example. Section headings are used in this disclosure to facilitate understanding of the disclosed embodiments and are not intended to limit the techniques disclosed in the sections to only the corresponding sections. The disclosed embodiments can be further embodied in a variety of different forms, and therefore, the scope of the present disclosure or the claimed subject matter is intended to be interpreted as not being limited to any embodiment set forth below. Various embodiments can be embodied as methods, devices, components, systems, or non-transitory computer-readable media. Therefore, the embodiments of the present disclosure can, for example, take the form of hardware, software, firmware, or any combination thereof.

[0043] The present disclosure generally relates to methods and network devices for assisting in the intelligent management of wireless communication systems, and more particularly to processes for transmitting and receiving artificial intelligence (AI) network management models between various network devices. These AI network management models can be used to implement various AI network configuration functions. Such AI network configuration functions can be provided as services for AI-assisted network configuration (AI configuration services, or AICs). Such AICs can be requested and configured via various messaging and signaling mechanisms. AI model lifecycle management, including training, activation, inference, deactivation, switching, performance evaluation, and the like, can be configured, triggered, and otherwise provided via control and data messages and signaling communicated between various network elements in a wireless communication system. In one particular aspect, a trained AI model residing and managed at a particular network node may need to be transferred to another network node. Because these AI models typically contain large amounts of data, the allocation of network communication resources (including radio resources in the air interface and other communication resources) and the configuration for their transmission may be handled differently than other control information communicated in the network.

[0044] Wireless Network Overview

[0045] exist Figure 1An example wireless communication network, shown as 100 in FIG. 1 , may include wireless terminal devices or user equipment (UEs) 110, 111, and 112, a carrier network 102, various service applications 140, and other data networks 150. The wireless terminal devices, or UEs, may alternatively be referred to as wireless terminals. Carrier network 102 may include, for example, access network nodes 120 and 121 and a core network 130. Carrier network 110 may be configured to transmit voice, data, and other information (collectively, data traffic) between UEs 110, 111, and 112, between the UEs and service applications 140, or between the UEs and other data networks 150. Access network nodes 120 and 121 may be configured as various wireless access network nodes (WANNs, alternatively referred to as wireless base stations) to interact with the UEs on one side of a communication session and with the core network 130 on the other side. The term "access network" may be used more broadly to refer to the combination of wireless terminal devices 110, 111, and 112 and access network nodes 120 and 121. The wireless access network may alternatively be referred to as a radio access network (RAN). The core network 130 may include various network nodes configured to control communication sessions and perform network access management and traffic routing. Service applications 140 may be hosted by various application servers deployed outside the core network 130 but connected to the core network 130. Similarly, other data networks 150 may also be connected to the core network 130.

[0046] exist Figure 1 In the example wireless communication network 100 of FIG. 1 , UEs can communicate with each other via a radio access network. For example, UE 110 and UE 112 can be connected and communicate via the same access network node 120. The UEs can communicate with each other via both the access network and the core network. For example, UE 110 can be connected to access network node 120 and UE 111 can be connected to access network node 121, and thus, UE 110 and UE 111 can communicate with each other via access network nodes 120 and 121 and core network 130. The UEs can further communicate with service applications 140 and data network 150 via core network 130. In addition, the UEs can communicate directly with each other via sidelink communications as shown in 113.

[0047] Figure 2Further shown is an example system diagram of a wireless access network 120 including a WANN 202 serving UE 110 and UE 112 via an air interface 204. The wireless transmission resources used for the air interface 204 include a combination of frequency, time and / or space resources. Each of UE 110 and UE 112 can be a mobile or fixed terminal device equipped with a mobile access unit (such as a SIM / USIM module) for accessing the wireless communication network 100. UE 110 and UE 112 can each be implemented as a terminal device, including but not limited to a mobile phone, a smart phone, a tablet computer, a laptop computer, an in-vehicle communication device, a roadside communication device, a sensor device, a smart home appliance (such as a TV, a refrigerator, and an oven), or other devices capable of wireless communication over a network. As Figure 2 As shown, each of the UEs (such as UE 112) may include a transceiver circuit 206 coupled to one or more antennas 208 to enable wireless communication with UE 120 or with another UE (such as UE 110). The transceiver circuit 206 may also be coupled to a processor 210, which may also be coupled to a memory 212 or other storage device. The memory 212 may be temporary or non-temporary and may store computer instructions or code therein that, when read and executed by the processor 210, causes the processor 210 to implement various methods described herein.

[0048] Similarly, WANN 120 may include a wireless base station or other wireless network access point capable of wirelessly communicating with one or more UEs and communicating with core network 130 via air interface 204. For example, WANN 120 may be implemented, without limitation, as a 2G base station, a 3G Node B, an LTE eNB, a 4G LTE base station, a 5G NR base station for a 5G gNB, a 5G central unit base station, or a 5G distributed unit base station. Each of these WANN types may be configured to perform a corresponding set of wireless network functions. WANN 202 may include transceiver circuitry 214 coupled to one or more antennas 216, which may include various forms of antenna towers 218, to enable wireless communication with UE 110 and UE 112. Transceiver circuitry 214 may be coupled to one or more processors 220, which may be further coupled to memory 222 or other storage device. The memory 222 may be transitory or non-transitory and may store therein instructions or code that, when read and executed by the one or more processors 220 , causes the one or more processors 220 to implement the various functions of the WANN 120 described herein.

[0049] Data packets in wireless access networks, such as Figure 2The examples described in , may be transmitted as protocol data units (PDUs). The data included therein may be packaged as PDUs at various network layers wrapped with nested and / or layered protocol headers. Once a connection (e.g., a radio link control (RRC) connection) is established between the transmitting end and the receiving end, the PDUs may be communicated between the transmitting device or transmitting end (the two terms are used interchangeably) and the receiving device or receiving end (the two terms are used interchangeably). Any of the transmitting device or receiving device may be a wireless terminal device (such as Figure 2 devices 110 and 120) or wireless access network nodes such as Figure 2 Each device may be a transmitting device and a receiving device for bidirectional communication.

[0050] Figure 1 The core network 130 may include various network nodes that are geographically distributed and interconnected to provide network coverage for the service area of the carrier network 102. These network nodes may be implemented as dedicated hardware network nodes. Alternatively, these network nodes may be virtualized and implemented as virtual machines or as software entities. These network nodes may each be configured with one or more types of network functions, which together provide the provisioning and routing functions of the core network 130.

[0051] Figure 3 An example division of network node functionality in the core network 130 is shown. Figure 3 Only a single instance of a network node for some functions is shown in FIG, but a person skilled in the art will appreciate that each of these network nodes may be instantiated into multiple instances distributed throughout the core network 130. Figure 3 As shown, the core network 130 may include, but is not limited to, an access management network function (AMF) node 330, a session management function (SMF) node 340, a user plane function (UPF) node 350, a policy control function (PCF) node 320, and an application data management function (AF) node 310.

[0052] The AMF node 230 may communicate with the access network 120, the SMF node 340, and the PCF node 320 via communication interfaces 322, 332, and 324, respectively, and may be responsible for provisioning registration, authentication, and access by the UE to the core network 130, as well as allocating an SMF node to support a specific UE communication session. The SMF node 340 allocated by the AMF node 330 may, in turn, be responsible for allocating a UPF node 350 to support the specific UE communication session and controlling these allocated UPF nodes 350 via communication interface 346. Alternatively or additionally, in some embodiments, the UPF node 350 may be directly allocated by the AMF node 330 via interface 334 and controlled by the SMF node 340 via communication interface 346. The access policy and session routing policy applicable to the UE may be managed by the PCF node 320, which communicates the policy to the AMF node 330 and the SMF node 340 via communication interfaces 324 and 323, respectively. The PCF node 320 may be further responsible for managing user subscriptions 312 to the service application 140 via the AF node 310. Figure 3 The signaling and data exchange between various types of network nodes in the various communication interfaces indicated by the various connection lines can be carried by signaling or data messages that follow a predetermined type of format or protocol.

[0053] In order to support a specific end-to-end communication task requested by the UE, a communication session may be established to support a data traffic pipe for transporting the specific end-to-end data communication traffic. Figure 3 The carrier network portion of the data traffic pipe shown in 370 may involve one or more network nodes in the access network 120 and a set of UPF nodes 352, 354, and 356 in the core network 130, for example, as selected and controlled by a set of SMF nodes 342 and 344, which may be selected and controlled by the AMF node 330 responsible for establishing and managing the communication session. Data traffic is routed between the UE at one end of the data traffic pipe, the carrier network portion of the data traffic pipe (including the set of network nodes in the access network 120 and the selected UPF nodes 352, 354, and 356 in the core network 130), and the other end of the data traffic pipe (including, for example, another UE, a service application or application server 140, or a data network 150) via communication interfaces such as 324, 358, and 359.

[0054] For some communication sessions, data transmitted within core network 130 may terminate at application server 140. In other words, application server 140 may be the destination of data traffic routed within core network 130. Similarly, application server 140 may also be the source of data traffic to be routed by core network 130 to other destinations. In such communication sessions, application server 140 may be accessed by the carrier network portion 370 of the data traffic conduit for the communication session, as indicated by 359. Communication conduit 370 for data traffic may be referred to as the user plane (UP) of carrier network 130, while other core network functions may be referred to as part of the control plane (CP) of carrier network 130. The separation of the UP and CP of carrier network 130 may facilitate efficient resource management, configuration, control, and utilization.

[0055] The application server 140 may further communicate other configuration and control information to the core network 130. The information communicated to the core network 130 may be referred to as application data. Such application data may be provided by Figure 2 The application data may be processed and managed by the AF node 310 in the core network. The application data may be conveyed, for example, in a message from the application server 140 to the AF network node 310 via the communication interface 314. Alternatively, the application server 140 may access the AF node 310 using an open API provided by the core network 130. Figure 2 Only a single application server is shown, but those skilled in the art will appreciate that, in actual implementations, the core network 130 may support multiple service applications of different types.

[0056] Figure 4 Further explained in Figures 1 to 3 FIG. 4 is a simplified diagram of the various network layers involved in transmitting a user plane PDU from a transmitting device 402 to a receiving device 404 in an example wireless access network. Figure 4 It is not intended to include all necessary equipment components or network layers for handling the transmission of PDUs. Figure 4 It shows that the data packaged by the upper network layer 420 at the transmitting device 402 can be transmitted via the packet data convergence protocol layer (PDCP layer, Figure 4 The PDUs are transmitted to corresponding upper layers 430 (such as a radio resource control or RRC layer) at the receiving device 304, through the physical (PHY) layers and radio interfaces (shown as 406) of the transmitting and receiving devices, and the medium access control (MAC) layer 434 and RLC layer 432 of the receiving device. Various network entities in each of these layers may be configured to handle the transmission and retransmission of the PDUs.

[0057] exist Figure 4The upper layer 420 may be referred to as layer-3 or L3, while layers such as the RLC layer and / or the MAC layer and / or the PDCP layer ( Figure 4 The intermediate layers (not shown) may be collectively referred to as Layer-2 or L2, and the term Layer-1 is used to refer to layers such as the physical layer and radio interface associated layers. In some instances, the term "lower layer" may be used to refer to the set of L1 and L2, while the term "higher layer" may be used to refer to Layer-3. The term "lower layer" may be used to refer to layers below the current reference layer among L1, L2, and L3. Control signaling can be initiated and triggered at each of L1 to L3 and within the various network layers therein. These signaling messages can be encapsulated and concatenated into lower layer packets and transmitted via allocated control or data air radio resources and interfaces. The term "layer" generally includes its various corresponding entities. For example, the MAC layer covers the corresponding MAC entities that can be created. Layer-1 includes, for example, a PHY entity. As another example, Layer-2 covers the MAC layer / entity, the RLC layer / entity, the Service Data Adaptation Protocol (SDAP) layer, and / or the PDCP layer / entity.

[0058] AI in Wireless Network Configuration

[0059] At the core of a general AI network management framework are various AI models. AI models typically contain a large number of model parameters, which are determined through a training process in which correlations within a set of training data are learned and embedded in the trained model parameters. Consequently, the trained model parameters can be used to generate inferences from a set of input data sets that may not have existed in the training dataset. AI models are particularly useful in situations where there is little to no traceable deterministic, rule-based, or analytical derivation path between input data and output.

[0060] In wireless communication systems, such as those described above, the determination of adaptive network configurations may rely on empirical characteristics and may further require lengthy measurement processes and / or a significant amount of computing power. This type of configuration may include, but is not limited to, air interface beam management, channel state information (CSI) feedback compression and decompression, and wireless terminal positioning. The correlation between various network conditions and these adaptive configurations can be learned via AI technology. Therefore, the use of artificial intelligence models to assist in network configuration can help reduce the amount of measurements and computing requirements, providing a more flexible network configuration. Therefore, it may be desirable to provide a mechanism for supplying the life cycle of various AI models and applications to assist in adaptively determining these network configurations. Certain key aspects of the AI model life cycle include delivering and transmitting AI models between the various network nodes mentioned above.

[0061] For example, AI technology can be applied to beam management in air communication interfaces. In current embodiments, beam management typically relies on exhaustive search beam scanning. In other words, the network (NW) can perform a full scan of the beam by sending a sufficient number of reference signals. The UE can be configured to monitor and measure each reference signal, and then report the measurement results to the NW for the NW to decide the best beam to which the UE will switch. However, this process is resource and power intensive. Using a trained AI model that embeds learned correlations between various network condition parameters, very few measurements (or fewer reference signals) may be required to accurately infer the best beam. In some embodiments, the AI model can use other network conditions to help identify the inference of the best candidate beam and then only scan and measure the candidate beams to select the beam for current communication. In addition, because the beam configuration is closely related to the location of the UE, AI technology can be further used to infer or predict the UE trajectory or position, thereby indirectly assisting in the selection of the best beam.

[0062] For another example, AI technology can be applied to channel state information (CSI) feedback. Traditionally, CSI feedback can be implemented using a codebook known to the UE and NW. The UE can measure the CSI and obtain the measurement result, and then map the measurement result to the closest vector of the codebook, and transmit the index of the vector to the NW to save air interface resource consumption. However, because the codebook is not infinite or dynamically variable over time, there will always be a mismatch, resulting in uncontrolled CSI feedback errors as the wireless environment changes. AI can therefore be applied to compression-decompression for CSI feedback. Specifically, the CSI report can be compressed by the UE-side AI model and decompressed by the corresponding NW-side AI model. Such AI models may be initially trained and continue to evolve over time and with the accumulation of network conditions.

[0063] For another example, AI technology can be applied to UE positioning. Traditional methods for UE positioning depend on PRS or SRS (such as DL positioning reference signal and uplink sounding reference signal). Regardless of which alternative method is adopted, LOS (Line-Of-Sight) beam is the key beam to be identified in order to generate the most accurate position estimate through triangulation on the NW side. However, in most cases, it is difficult to identify LOS beam from other NLOS (Non-Line-Of-Sight) beams to provide accurate UE positioning. On the other hand, the trained AI model can identify various patterns and correlations in PRS and SRS to extract LOS information and provide more accurate UE positioning.

[0064] These AI models can be trained and managed at the various network nodes described above and may need to be delivered or transmitted to another network node. For example, in some cases, it may be necessary to train or manage the AI models in either the uplink (UL) or downlink (DL) communication direction. Figure 1 and Figure 3 In some other example cases, the AI model may need to be transmitted between the UE and the wireless base station as the endpoint, such as Figure 1 and Figure 2 As shown, or between the UE as an endpoint and the core network node and via the base station (such as Figure 1 and Figure 3 The following disclosure provides example embodiments of interaction procedures between various network nodes in a wireless communication network for resource request, allocation, and management to transmit an AI network management model in either the CP, UP, or both the CP and UP of the network.

[0065] AI network management model transmission between UE as endpoint and base station

[0066] In some example cases, the AI model trained and / or managed at the UE may need to be transmitted to the base station, and vice versa. Such AI model transmission can be carried out in the air interface. The communication resources in the air interface can typically include multiple radio bearers (RB) resources. In some example embodiments, such RB resources between the UE and the base station, either UL RBs, DL RBs, or bidirectional RBs, can be further managed and allocated as separate categories of signaling RBs (SRBs) or data RBs (DRBs). For example, SRB resources can be allocated to carrier signaling and control information, while DRB resources can be allocated to carrier data information. Therefore, SRB or DRB resources can be allocated for the transmission of AI network management models.

[0067] The allocation of SRBs by the base station generally does not require the establishment of a PDU session and therefore may not require the involvement of the core network. Therefore, the existing radio resource request and allocation schemes of existing wireless networks can be used to transmit the AI network management model between the UE and the base station via SRB resources without involving the core network. However, because DRB allocations in existing network management schemes generally involve data transmission in the user plane (UP) of the core network, they are associated with one or more PDU sessions. The base station may not be able to allocate any DRB resources for the transmission of the AI network management model between the UE and the base station in accordance with the DRB resource allocation scheme in the existing network without involving the core network in performing various processes to establish and specify one or more PDU sessions.

[0068] The following various example embodiments describe schemes, (1) in which SRB resources are requested and allocated for transmission of an AI network management model between a UE and a base station without involving any core network node, (2) in which DRB resources are requested and allocated with a new procedure without involving any core network node for transmission of an AI network management model between a UE and a base station, and (3) in which DRB resources are requested and allocated in the presence of a core network involved in establishing one or more PDU sessions for transmission of an AI network management model, but in which there is no model transmission endpoint in the core network (the AI model transmission endpoints are located at the UE and the base station, both of which are outside the core network).

[0069] Further in the following embodiments, either the UE or the base station can be the requesting device for AI model transmission, and the AI model transmission can be in the UL direction or the DL direction, resulting in four different scenarios:

[0070] S1: The base station requests UL AI model transmission from the UE;

[0071] S2: UE requests DL AI model transmission from the eNodeB.

[0072] S3: UE requests UL AI model transmission to the base station; or

[0073] S4: The base station requests DL AI model transmission to the UE.

[0074] Implementation of SRB or DRB resources in the air interface without CN participation

[0075] Figure 5An example general process 500 for initiating, triggering, and terminating AI model transmission between a UE and a base station without core network involvement is shown. This example process is generally applicable to all scenarios described above (scenarios S1 to S4 described above regarding which of the UE and base station acts as the requester or responder of the AI model transmission and which acts as the origin or destination of the model).

[0076] Figure 5 The process 500 involves a requesting network device 502 and a responding network device 506 for AI model transmission that first implement a request-response process. The requesting network device 502 can be one of a UE and a base station, and the responding network device 506 can be the other of the UE and the base station. The requesting network device 502, whether it is a UE or a base station, can be the origin or destination of the AI model being transmitted, and thus can serve as an AI model origin network device to transmit the AI model, or as a model destination network device to receive the AI model. Similarly, the responding network device 506, whether it is a UE or a base station, can also be the origin or destination of the AI model being transmitted, and thus can serve as an AI model origin network device to transmit the AI model, or as a model destination network device to receive the AI model.

[0077] Figure 5 The general process 500 may include:

[0078] The requesting network device 502 transmits an AI model transmission request 504 to the responding network device 506 .

[0079] The model transfer request 504 is received by the responding network device 506 and the responding network device 506 transmits an AI model transfer response 508 to the requesting network device 502 .

[0080] In 510 , network resources for AI model transmission are established (particularly for the air interface).

[0081] Transmitting the AI network management model from the AI model origin device (one of the requesting network device and the responding network device) to the AI model destination network device (the other of the requesting network device and the responding network device), as shown at 512, with double arrows indicating either UP or DL transmission direction.

[0082] SRB-based implementation

[0083] In some example embodiments, the radio resources in the air interface allocated for transmission of the AI network management model in 510 may be one or more SRB resources. Such embodiments may include the following example steps:

[0084] Step 1: The requesting network device 502 may generate an AI model transmission request 504 based on the AI-based functions associated with the AI model to be transmitted (e.g., beam management function, CSI reporting function, and positioning function) and send it to the responding network device 506.

[0085] In some example implementations, the AI model transmission request 504 may be transmitted in the form of a UL RRC or DL RRC message. In some alternative implementations, the AI model transmission request 504 may be transmitted in the form of a UL or DL MAC control element (MAC CE).

[0086] In some other example embodiments, both RRC messages and MAC CEs may be used as available and permissible forms for transmitting the AI model transmission request 504. Either of these forms may be selected by the requester device 502 based on whether a previous SRB resource has been established for AI model transmission. For example, if no previous SRB resource has been established for AI model transmission, the requesting network device 502 may determine that a UL RRC or DL RRC message may be used to transmit the AI model transmission request 504. Otherwise, if one or more previous SRB resources have been established for AI model transmission, the requesting network device 502 may determine that a DL or UL MAC CE should be used to transmit the AI model transmission request 504.

[0087] In some example embodiments, when the requesting network device 502 determines that a UL MAC CE is to be used to transmit the AI model transmission request 504, and if the AI model transmission request has been triggered and pending and there are no available PUSCH resources to accommodate the MAC CE, a scheduling request (SR) may be triggered to schedule PUSCH resources for the MAC CE.

[0088] In some example embodiments, a DL RRC message may be used to transmit the AI model transmission request 504 (in the case where the requesting network device 502 is a base station), which may include a DL RRC reconfiguration message sent by the base station in a message to add one or more SRB resource allocations for AI model transmission. Such SRB resources can be added separately from other existing types of UL, DL, or bidirectional SRBs. For example, SRB type 5 can be added on top of normal SRB types 1 to 4. The added UL, DL, or bidirectional SRB can be dedicated to AI model transmission between the UE and the base station. Allocating a dedicated SRB for AI model transmission can be advantageous due to the distinctive characteristics and requirements for AI model transmission compared to other signaling information typically transmitted in SRB resources (e.g., data volume, transmission reliability, delay tolerance). Dedicating SRBs to the transmission of the AI network management model also helps minimize the degradation of other signaling information in existing SRBs.

[0089] In some example embodiments, the AI model transmission request 504 may include at least one of the following information (in the form of an RRC message or a MAC CE in a UL AI model transmission request or a DL AI model transmission request):

[0090] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0091] An AI model identifier or signature indicating the requested AI model. This identifier may be an AI model ID, or an AI model description that may be used to identify the AI model to be transmitted.

[0092] In some example embodiments, if the AI model transmission request 504 is a DL RRC message, it may further include an SRB indicator or identifier to specify the SRB resource to be used for transmission of the AI network management model when an existing SRB resource rather than a new SRB is to be used for AI model transmission. In some example embodiments, the AI model transmission request 504 may implicitly indicate that the SRB being used to send the AI model transmission request 504 is the SRB used for transmission of the AI network management model.

[0093] In some example embodiments, when an RRC message is used for transmission of the AI model transmission request 504 and when one or more SRB resources to be used for transmission of the requested AI model are shared with existing signaling information, the AI model transmission request 504 may further include an RRC transaction identifier for indicating the RRC procedure associated with the AI model transmission request, so that it can be distinguished from other RRC messages used to transmit signaling information in a shared SRB, so as to provide differentiated processing of AI model transmission and other signaling information when necessary.

[0094] Step 2: The responding network device 506 may respond to the requesting network device 502 with an AI model transmission response 508 when the AI model transmission request 504 is received and accepted.

[0095] In some example embodiments, the AI model transmission response 508 may be an UL RRC message (the UE is the responding network device and the base station is the requesting network device) or a DL RRC message (the base station is the responding network device and the UE is the requesting network device). The UL RRC or DL RRC message may indicate an acknowledgement (ACK) or a rejection (NACK) to the AI model transmission request 504. If the UL RRC or DL RRC message includes a rejection, it may further include information about one or more reasons for the rejection.

[0096] In some example embodiments, if the AI model transmission response 508 is a DL RRC message (NW is the responding network device and the base station is the requesting network device), the AI model transmission response 508 may further include an SRB indicator or identifier to specify the SRB resources to be used for transmission of the AI network management model when existing SRB resources rather than new SRBs are to be used for AI model transmission.

[0097] In some alternative example embodiments, the AI model transmission response 508 may be a UL MAC CE (the UE is the responding network device and the base station is the requesting network device) or a DL MAC CE (the base station is the responding network device and the UE is the requesting network device). The UL MAC CE or DL MAC CE may indicate an acknowledgment (ACK) or a rejection (NACK) to the AI model transmission request 504. If the UL MAC CE or DL MAC CE message includes a rejection, it may further include information about one or more reasons for the rejection.

[0098] In some example embodiments, when the AI model transmission request 504 is an RRC reconfiguration message for configuration of dedicated SRB resources for transmission of the AI network management model, the AI model transmission response 508 may correspondingly be an RRC-reconfiguration-complete message for indicating confirmation or establishment of the SRB resources.

[0099] In some example embodiments, when the model transmission request 504 is a UL RRC message or a UL MAC CE (from the UE to the base station), the AI model transmission response 508 may be an RRC reconfiguration message from the base station to the UE, requesting the UE to establish one or more dedicated SRB resources for transmission of the AI network management model.

[0100] In some example embodiments, when the model transmission request 504 is a DL RRC message (from the base station to the UE), which may include an RRC reconfiguration message for establishing one or more dedicated SRB resources for transmission of the AI network management model, the AI model transmission response 508 may correspondingly be an RRC-reconfiguration-complete message for indicating confirmation and establishment of the SRB resources.

[0101] Step 3: If the AI model transmission request is confirmed and the SRB has not been previously configured, the base station may initiate an RRC reconfiguration procedure to establish one or more new SRB resources. In the case where the AI model transmission request is transmitted from the base station to the UE in an RRC reconfiguration message, the request message acts to initiate an RRC reconfiguration procedure for establishing new SRB resources. For AI model transmission using already allocated SRBs, RRC reconfiguration is not required.

[0102] Step 4: The originating network device (UE or base station, as the requesting network device or the responding network device) may start transmitting the requested AI network management model to the destination network device (base station or UE, as the requesting network device or the responding network device) via the allocated or newly established SRB resources.

[0103] DRB-based implementation

[0104] In some example embodiments, Figure 5 The radio resources in the air interface allocated for transmission of the AI network management model in 510 may be one or more DRB resources. Such implementation may include the following example steps:

[0105] Step 1: The requesting network device 502 may generate a model transfer request 504 based on the AI-based functionality associated with the AI model to be transferred and transmit the request to the responding network device 506 .

[0106] In some example embodiments, the model transmission request 504 may be transmitted in the form of a UL RRC or DL RRC message. In some other embodiments, the AI model transmission request may be transmitted in the form of a UL MAC CE or a DL MAC CE.

[0107] In some other example embodiments, as in the SRB case described above, both RRC messages and MAC CEs may be used as available and permissible forms for transmitting the AI model transmission request 504. Either of these forms may be selected by the requester device 502 based on whether a previous DRB resource has been established for AI model transmission. For example, if no DRB resource has been previously established or no DRB resource is available for AI model transmission, the requesting network device 502 may determine that a UL RRC or DL RRC message may be used to transmit the AI model transmission request 504. Otherwise, if one or more previous DRB resources have been established for AI model transmission, the requesting network device 502 may determine that a DL or UL MAC CE should be used to transmit the AI model transmission request 504.

[0108] In some example embodiments, when the requesting network device 502 determines that a UL MAC CE is to be used to transmit the AI model transmission request 504, and if the AI model transmission request has been triggered and pending and there are no available PUSCH resources to accommodate the MAC CE, a scheduling request (SR) may be triggered to schedule PUSCH resources for the MAC CE.

[0109] In some example embodiments, the AI model transmission request 504 may include at least one of the following information (in the form of an RRC message or a MAC CE in a UL AI model transmission request or a DL AI model transmission request):

[0110] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0111] An AI model identifier or signature indicating the requested AI model. This identifier may be an AI model ID, or an AI model description that may be used to identify the AI model to be transmitted.

[0112] A DRB indicator or identifier for indicating a DRB resource that can be used for transmission of the AI network management model.

[0113] A PDU session indicator or identifier for indicating a PDU session that may be used for transmission of the AI network management model.

[0114] A QoS flow indicator or identifier for indicating a QoS flow associated with a PDU session indicated by the current PDU session indicator or identifier as being available for transmission of the AI network management model.

[0115] In some example embodiments, where the model transmission request is a DL RRC message, the DL RRC message may be or may include an RRC reconfiguration message that configures the DRBs to be used for transmission of the AI network management model. In one embodiment of the DRBs used for transmission of the AI network management model, the RRC reconfiguration message may be used to add / modify DRBs that are only used for transmission of the AI network management model. In another embodiment, the RRC reconfiguration message may be used to add / modify DRBs that can be used for transmission of the AI network management model and that can be associated with a PDU session via SDAP-Config.

[0116] In some example embodiments, when an RRC message is used for transmission of the AI model transmission request 504 and when one or more SRB resources to be used for transmission of the requested AI model are shared with existing signaling information, the AI model 504 may further include an RRC transaction identifier for indicating the RRC procedure associated with the AI model transmission request, so that it is distinguished from other RRC messages used to transmit signaling information in a shared SRB, so as to provide differentiated handling of the AI model transmission request and other signaling information when necessary.

[0117] Step 2: The responding network device 506 may respond to the requesting network device 502 with an AI model transmission response 508 when the AI model transmission request 504 is received and accepted.

[0118] In some example embodiments, the AI model transmission response 508 may be an UL RRC message (the UE is the responding network device and the base station is the requesting network device) or a DL RRC message (the base station is the responding network device and the UE is the requesting network device). The UL RRC or DL RRC message may indicate an acknowledgement (ACK) or a rejection (NACK) to the AI model transmission request 504. If the UL RRC or DL RRC message includes a rejection, it may further include information about one or more reasons for the rejection.

[0119] In some alternative example embodiments, the AI model transmission response 508 may be a UL MAC CE (the UE is the responding network device and the base station is the requesting network device) or a DL MAC CE (the base station is the responding network device and the UE is the requesting network device). The UL MAC CE or DL MAC CE may indicate an acknowledgment (ACK) or a rejection (NACK) to the AI model transmission request 504. If the UL MAC CE or DL MAC CE message includes a rejection, it may further include information about one or more reasons for the rejection.

[0120] In some example embodiments, when the AI model transmission response 508 is a UL RRC message (ie, from the UE to the NW), the AI model transmission response may be or may include an RRC reconfiguration complete message indicating an ACK or establishment of the specified DRB resources.

[0121] In some example embodiments, when the AI model transmission response 508 is a DL RRC message (i.e., from the network to the UE), the AI model transmission response may be or may include an RRC reconfiguration message for establishing and / or configuring DRB resources for AI model transmission.

[0122] In some example embodiments, the AI model transmission response 508 may include at least one of: one or more PDU session identifiers associated with one or more PDU sessions to be used for transmission of the AI network management model, one or more QFIs associated with one or more QoS flows to be used for transmission of the AI network management model, and / or one or more DRB IDs indicating air interface DRB resources to be utilized for transmission of the AI network management model.

[0123] Step 3: If the AI model transmission request is confirmed and the DRB for transmission of the AI network management model has not been previously configured, the base station may initiate an RRC reconfiguration procedure to establish one or more new DRB resources.

[0124] In some example embodiments of configuring one or more specific DRB resources, one or more DRB resources may be dedicated to AI model transmission. Such RRC configuration for DRB resources may include one or more flags in the DRB configuration (i.e., RadioBearerConfig) to indicate that one or more DRB resources are established for AI model transmission. In an alternative embodiment, such a flag may be present in the SDAP configuration associated with the DRB resource (i.e., SDAP-Config). In another alternative embodiment, such a flag may be present in the PDCP configuration associated with the DRB resource (i.e., PDCP-Config). In yet another alternative embodiment, such a flag may be present in the RLC configuration associated with the DRB resource (i.e., RLC-Config). In other words, one or more flags for indicating one or more DRB resources may be placed at various levels in the radio bearer configuration stack.

[0125] In some example embodiments, the SDAP-configuration may not be present in the RRC configuration of the DRB resources used for AI model transmission. In other words, the DRB resources used for AI network-managed transmission may not be associated with any PDU session and / or QoS flow.

[0126] In some example embodiments, the RLC mode of the RLC entity associated with the DRB resources used for transmission of AI network management may be set to Acknowledgement Mode (AM) to ensure reception of the AI network management model.

[0127] In some example embodiments, when an SDAP configuration is present in an RRC configuration associated with a DRB for AI model transmission (e.g., within the radiobearerConfig of the RRC configuration), the PDU session ID may not be present in the SDAP configuration because the AI model transmission is between the UE and the base station and therefore does not need to rely on any PDU session.

[0128] In some example embodiments, rather than using a dedicated DRB for AI model transmission, legacy DRB resources may be configured for AI model transmission. Configuring such DRB resources may include at least one of the following:

[0129] An indicator bit in the header of an SDAP or PDCP or RLC PDU to indicate that the corresponding PDU is for AI model transmission; or

[0130] One or more QFIs in a special QFI list in SDAP-Config, indicating that data with this type of QoS flow is for AI model transmission, where this special QFI list is introduced to distinguish between regular data transmission and AI model transmission (for example, using mappedQoS-FlowsForModelTransferToAdd to construct this special QFI list). In such an embodiment, the UL or DL SDAP header in which the QoS flow ID is present can be configured to be present.

[0131] Step 4: The originating network device (UE or base station, as the requesting network device or the responding network device) may start transmitting the requested AI network management model to the destination network device (base station or UE, as the requesting network device or the responding network device) via the allocated DRB resources for transmission of the AI network management model.

[0132] Implementation Method of DRB Resources in Air Interface with CN Participation

[0133] In some embodiments where the AI network management model is to be transmitted between the UE and the base station, the configuration of the air interface resources may involve the core network.

[0134] An example of a general implementation is shown in Figure 6 In the flowchart 600. Figure 6 In the example embodiment, the requesting network device (UE 602 or base station 604) can transmit an AI model transmission request 608 to the responding network device (base station 604 or UE 602). The responding network device can send an AI model transmission response 610 after receiving the AI model transmission request 608. When the UE is the AI model transmission requester, the AI request response 610 may be optional. In that case, the response to the UE can be delayed after the base station successfully interacts with the core network to configure communication resources for the transmission of the AI model (for example, the response can be effectively implemented through the RRC reconfiguration message 616 described below).

[0135] The interaction between the UE and the base station in 608 and 610 may be referred to as an AI model transmission request-response procedure. Either the UE or the base station may be a requesting network device, and the other of the UE or the base station may be a responding network device. In the case where the UE is the requester and no immediate response to the AI model transmission request is sent back to the UE, the corresponding process without a response may be referred to as an AI model requisition procedure.

[0136] like Figure 6As shown, following the AI model transmission request-response process or the application process, the base station 604 may initiate an AI model transmission tunnel request 612 with a CN node (e.g., an AMF node 606), and then an AI model transmission tunnel response 614 is transmitted from the CN node 606 to the base station 604. After receiving the AI model transmission tunnel response 614, and if the CN node 606 authorizes the requested tunnel, the base station 604 may perform network configuration, for example, by generating an RRC reconfiguration message 616 and sending it to the UE 602. The UE then determines and establishes communication resources for AI model transmission in the air interface according to the RRC reconfiguration message, as shown in FIG. Figure 6 As shown in 618, and the confirmation of RRC reconfiguration 620 is conveyed to the base station 604. Thereafter, in 622, the requested AI network management model can be transmitted between the AI model originating device and the AI model destination device. The originating AI device can be a UE or a base station. Correspondingly, the destination AI device can be a base station or a UE.

[0137] Figure 6 The example embodiment of substantially utilizes a data tunnel request and establishment process between a wireless base station 604 and a core network, and uses radio resources associated with the data tunnel so established to transmit the AI network management model between the UE 602 and the base station 604 as endpoints of the AI model transmission. Figure 6 Involving the CN for utilizing the CN-assisted data tunnel, DRB resources in the air interface between the UE 602 and the base station 604 may be used as radio resources for transmitting the requested AI network management model.

[0138] Figure 6 The general implementation can be carried out in the following steps.

[0139] Step 1: The requesting network device (UE 602 or base station 604) transmits an AI model transmission request 608 to the responding network device (base station 604 or UE 602). The request may be associated with one or more AI-based network management functions (e.g., beam management function, CSI reporting function, positioning function, etc.).

[0140] In some example implementations, the AI model transmission request 608 may be transmitted in the form of a UL RRC or DL RRC message. In some alternative implementations, the AI model transmission request 504 may be transmitted in the form of a UL or DL MAC control element (MAC CE).

[0141] In some other example embodiments, both RRC messages and MAC CEs may be used as available and allowable forms for transmitting the AI model transmission request 608. Any of these forms may be selected by the requester device (base station or UE) depending on whether a previous data tunnel or resource has been established for AI model transmission. For example, if no previous data tunnel or resource has been established or configured for AI model transmission, the requesting network device may determine that a UL RRC or DL RRC message may be used to transmit the AI model transmission request 606. Otherwise, if one or more previous data tunnels or resources have been established or configured for AI model transmission, the requesting network device may determine that a DL or UL MAC CE should be used to transmit the AI model transmission request 608. The previously available resources may be one or more DRBs, one or more QoS flows of a PDU session, or one or more PDU sessions indicated by the configuration to be used for AI model transmission.

[0142] In some example embodiments, when the AI model transmission request 608 is a UL RRC message (the UE is the requesting network device), such RRC message may include a non-access stratum (NAS) container that contains a request message as an AI model transmission tunnel request (for requesting a data tunnel for AI model transmission). This NAS request message may be destined for the core network node 606 in the control plane (CP), for example, and may be transparent to the base station 604. In some example embodiments, when a data tunnel is already available for AI model transmission, then the NAS message container may not be present.

[0143] In some example embodiments, the AI model transmission request 608 (either UL from the UE to the base station or DL from the base station to the UE, which may be an RRC message or a MAC CE) may include at least one of the following:

[0144] A request message that is an AI model transport tunnel request (for requesting a data tunnel for AI model transport). A flag may be present in such a request to indicate that the tunnel is being requested for AI model transport. In one example embodiment, such a request message may be present in a non-access stratum (NAS) container that is transparent to the gNB.

[0145] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0146] An AI model identification or identifier that indicates the requested AI model. This identifier can be an AI model ID, or an AI model description that can be used to request other endpoints to determine or search for a matching AI model to be transmitted.

[0147] AI model information, used to indicate information about the requested AI model, such as the AI model size.

[0148] In some example embodiments, the AI model transmission request 608 may be a UL RRC message (e.g., the UE is the requesting network device) and may further include an endpoint IP address for the UE for AI model transmission (the UE being the AI model origin device or the AI model destination device). Such endpoint IP address of the UE may be used by the core network to identify or look up a UPF associated with a data tunnel to be allocated for transmission of the AI network management model.

[0149] In some example embodiments, the AI model transmission request 608 may be a UL MAC CE or a DL MAC CE. In the event that a data tunnel may already be available for transmission of the AI model, such a UL MAC CE or a DL MAC CE may further include at least one of the following aspects for the MAC CE request message to identify the available data tunnel:

[0150] One or more PDU session identifiers to indicate the PDU session or sessions to be used for the requested AI model transmission.

[0151] One or more QoS flow identifiers, indicating one or more QoS flows to be used to transmit the requested AI model.

[0152] One or more DRB identifiers indicating one or more DRB resources to be used for transmitting the requested AI model.

[0153] An AI-based function indication or identifier, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0154] An AI model identifier or signature, which indicates the requested AI model. This identifier may be an AI model ID.

[0155] AI model information, used to indicate information about the requested AI model, such as the AI model size.

[0156] The endpoint of the UE's IP address.

[0157] Step 2: The responding network device in the UE or base station may generate an AI model transmission response 610. Such a response may be an RRC message or a MAC CE, respectively. As described above, such an immediate response from the base station to the UE may be optional.

[0158] In some example embodiments, when the AI model transmission response 610 is a UL RRC message (the UE is the responding network device), such RRC message may include a NAS container that contains a request message that is an AI model transmission tunnel request (for requesting a data tunnel for AI model transmission). As described above, the NAS request message may be destined for the core network node 606 in the control plane (CP) and may be transparent to the base station 604.

[0159] In case the response 610 is a UL RRC message, it may further include at least one of the following:

[0160] A request message that is an AI model transmission tunnel request (used to request a data tunnel for AI model transmission). A flag may be present in such a request to indicate that the tunnel is requested for AI model transmission. In one embodiment, such a request message may be present in a non-access stratum (NAS) container that is transparent to the gNB.

[0161] An identifier used to indicate an AI-based function associated with the AI network management model.

[0162] Confirmation of AI model transfer request.

[0163] The endpoint of the UE's IP address.

[0164] Acknowledgement of failure of AI model transfer request.

[0165] One or more reasons for failed validation.

[0166] In some example embodiments, when the AI model transmission response 610 is a UL MAC CE or a DL MAC CE, such MAC CE response may include at least one of the following fields:

[0167] • An identifier indicating the AI-based function to which the validated AI model to be transferred belongs.

[0168] AI model identifier, used to indicate the AI model being confirmed.

[0169] One or more PDU session identifiers to indicate one or more PDU sessions to be used to perform the AI model transfer.

[0170] One or more DRB identifiers indicating one or more DRB resources to be used to perform the AI model transfer.

[0171] One or more QFIs indicating one or more QoS flows over which the AI model transmission is to be executed.

[0172] Step 3: After the above AI model transmission request-response process, the base station 604 may then generate an AI model transmission tunnel request 612 and / or forward the AI model transmission tunnel request 612 to a core network node 606, such as an AMF, to obtain a model transmission tunnel according to the AI model transmission request 608.

[0173] In some example embodiments, the AI model transport tunnel request may be a request for one or more QoS flows of a PDU session from the AMF and may include at least one of the following:

[0174] AMF UE NGAP (NG Application Protocol) Identifier, used to indicate the UE specific AMF ID via the NGAP interface.

[0175] RAN UE NGAP Identifier, used to indicate the UE specific RAN ID via the NGAP interface.

[0176] • One or more PDU session identifiers associated with the requested QoS flow.

[0177] • The number of requested QoS flows for each of one or more PDU Sessions.

[0178] The number of AI network management models requested.

[0179] The size of each requested AI network management model.

[0180] One or more QoS requirements associated with the requested AI model or models.

[0181] UE endpoint IP address.

[0182] In some example embodiments, the AI model transmission tunnel request may be a request from the AMF for one or more PDU sessions for AI model transmission and may include at least one of the following:

[0183] ●UE endpoint IP address.

[0184] The number of PDU sessions requested.

[0185] One or more PDU session identifiers: Identifiers used to indicate the requested PDU session for the transmission of the AI network management model.

[0186] The size of each requested AI network management model.

[0187] The number of AI network management models requested.

[0188] AMF UE NGAP Identifier, used to indicate the UE specific AMF ID via the NGAP interface.

[0189] RAN UE NGAP Identifier, used to indicate the UE specific RAN ID via the NGAP interface.

[0190] One or more QoS requirements associated with the requested AI model or models.

[0191] Step 4: After receiving the AI model transmission tunnel request 612, the core network node (e.g., AMF) 606 may then transmit an AI model transmission tunnel response 614 to the base station.

[0192] In some example embodiments, when the AI model transmission tunnel request 612 is a request for one or more QoS flows as described above, the corresponding AI model transmission tunnel response 614 may include at least one of the following:

[0193] AMF UE NGAP Identifier, used to indicate the UE specific AMF ID via the NGAP interface (assuming the core network node 606 is an AMF).

[0194] RAN UE NGAP Identifier, used to indicate the UE specific RAN ID via the NGAP interface.

[0195] One or more QoS flow identifiers indicating the successful allocation of one or more QoS flows for the requested AI model.

[0196] One or more failed QoS flow identifiers to indicate failed allocation of QoS flows for the requested AI model.

[0197] One or more reasons for failed QoS flow allocation for the requested AI model

[0198] One or more PDU session identifiers to which the current QoS flow belongs.

[0199] In some example embodiments, when the AI model transmission tunnel request 612 is a request for one or more PDU sessions as described above, the corresponding AI model transmission tunnel response 614 may include at least one of the following:

[0200] AMF UE NGAP Identifier, used to indicate the UE specific AMF ID via the NGAP interface (assuming the core network node 606 is an AMF).

[0201] RAN UE NGAP Identifier, used to indicate the UE specific RAN ID via the NGAP interface.

[0202] One or more PDU session identifiers to indicate the successful allocation of one or more PDU sessions for the requested AI model.

[0203] One or more failed PDU session identifiers, indicating one or more failed allocations of PDU sessions for the requested AI model.

[0204] One or more reasons for failed PDU session allocation for the requested AI model.

[0205] Step 5: The base station may then initiate an RRC reconfiguration process 610 to allocate air interface resources for the requested tunnel according to the received AI model transmission tunnel response 614.

[0206] In the case where one or more QoS flows are to be allocated with air interface resources, in order to map the QoS flows to the AI model transmission using RRC reconfiguration, at least one of the following configurations may be adapted:

[0207] ●In SDAP configuration, a QoS flow list can be introduced to represent only AI model transfers, through, for example, QFIForModelTransfer-addModlist.

[0208] ●For SDAP configuration, when the QFI present in the SDAP configuration is configured for AI model transmission, sdap-HeaderDL / UL can be set to a value indicating that the SDAP PDU header is "present" so that the transmitted / received SDAPPDU has a header in which the QoS flow ID is included.

[0209] In the SDAP configuration, if no additional QFI list is introduced, a flag can be introduced in the QoS flow configuration to indicate that the QoS flows in the current QFI list in the SDAP configuration are used for AI model transmission.

[0210] In the event that one or more PDU sessions are to be allocated air interface resources for AI model transmission, in order to use RRC reconfiguration to map the PDU sessions to the AI model transmission, at least one of the following configurations may be adapted: For each DRB configured for one or more PDU sessions, a flag may be present in the corresponding radio bearer configuration to indicate that the DRB is for AI model transmission. Alternatively, a flag may be present in the corresponding SDAP configuration to indicate that the corresponding DRB is for AI model transmission. Alternatively, a flag may be present in the corresponding PDCP configuration to indicate that the corresponding DRB is for AI model transmission. Alternatively, a flag may be present in the corresponding RLC configuration to indicate that the corresponding DRB is for AI model transmission.

[0211] Step 6: Once the data tunnel and the corresponding air interface resources are established, the AI model origin network device (UE 602 or base station 604) can use the established tunnel for AI model transmission to start AI model transmission to the AI model destination network device (base station 604 or UE 602) via the air interface.

[0212] In some further example embodiments for transmitting the AI network management model between the UE and the base station, a CN network node acting as a data network node can be used as a proxy for the base station and act as an endpoint for the AI model transmission. In this way, the configuration for the transmission of the AI network management model can be based on the normal communication between the UE and the data network via the PDU session.

[0213] Such example implementations are shown in Figure 7 700. Figure 7 As shown, OAM 706 can be designated as an agent for base station 704 and an endpoint for the requested AI model transmission. The other endpoint of the AI model transmission is UE 702. OAM 706 can act as a data network node. AI model transmission request 710 and corresponding AI model transmission response 712 can thus be communicated between OAM 706 and UE 702, either of which is a requesting network device and the other is a responding network device. AI model transmission request 710 and AI model transmission response 712 can pass through base station 704 (as shown by bidirectional arrow 710 passing through base station 704). The AI model transmission request-response process is thus between the UE and OAM.

[0214] Following this type of AI model transmission request-response process, two options for AI model transmission may be implemented. In the first option, the transmission may be performed in a cellular network. For this option, the establishment of one or more PDU sessions in the cellular network for the transmission of the requested AI model and the actual UL or DL transmission of the model may follow normal cellular data communication between the UE and the data network, as shown in the process in dashed box 720. In the second option, the UE may abandon the cellular network and trigger the transmission of the AI network management model with the OAM as the data network via other communication paths or channels, as shown by 750. For example, an IP network such as a Wi-Fi network may exist between the UE and the OAM, and the AI model may be transmitted therebetween via an IP communication channel. If desired, the transmission of the AI model may be further performed between the base station and the OAM. For example, for transmitting the AI model from the base station to the UE, there may be two implementation methods: (1) if the AI model is stored at the base station, the base station will first transmit the model to the OAM in order to use the cellular network option to transmit the AI model from the OAM to the UE via the core network or use the IP network option to transmit the AI model from the OAM to the UE; (2) if the AI model is stored at the OAM but is to be deployed at the base station, the base station will transmit some relevant information about the requested AI model transmission to the OAM, and the OAM will use the cellular network option to transmit the AI model from the OAM to the UE via the core network or use the IP network option to transmit the AI model from the OAM to the UE. For another example, there may also be two implementations for transmitting the AI model from the UE to the base station: (1) if the AI model is stored at the base station, then after the UE has transmitted the model to the OAM via the first option or the second option described above, the OAM transmits the model to the base station; and (2) if the AI model is stored at the OAM but is to be deployed at the base station, then after the UE has transmitted the model to the OAM via the first option or the second option described above, the OAM may need to deploy the model to the base station.

[0215] Therefore, in the first option above, the AI model transmission from the UE to the OAM is performed in the user plane (UP) and the AI model is transmitted via the core network. In the first option, an AI model transmission tunnel request is made to the core network to establish a PDU session. This tunnel request can be initiated by the AI model request-response process between the UE and the OAM (base station) described above. Therefore, the tunnel request can be considered to be initiated by either the UE or the OAM (base station).

[0216] Figure 7 The process outlined in can be implemented as the following steps.

[0217] Step 1: The AI model requesting device (OAM 706 or UE 702) may generate an AI model transmission request 710 and transmit the request to the responding network device (UE 702 or OAM 706) via the base station 704. The base station 704 connects to the OAM 706 using a private interface. For UE-initiated requests, the AI model transmission request may first be sent from the UE to the base station, and then the base station forwards the request to the OAM. For OAM-initiated requests, the AI model transmission request may first be sent from the OAM to the base station, and then the base station forwards the request to the UE.

[0218] In some example embodiments, the AI model transfer request 710 may be implemented as a DL or UL RRC message (eg, AIModelTransferRequest). Alternatively, the AI model transfer request 710 may be implemented as a DL or UL MAC CE (eg, AI Model Transfer Request MAC CE).

[0219] In some example embodiments, such an AI model transmission request via an RRC message or a MAC CE in the UL direction or the DL direction may include at least one of the following:

[0220] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0221] An AI model identification or identifier indicating the requested AI model. This type of identifier can be an AI model ID, or can be used to identify an AI model description to be transmitted.

[0222] • Endpoint IP address at the OAM side or UE side (depending on which is the requesting network device).

[0223] Network type indication for model transmission: used to indicate the network type used for model transmission, such as NR network, WLAN or other IP network.

[0224] One or more transmission latency requirements (e.g., if the AI model transmission is over a cellular network, or the first option, there may be network charges or fees for that transmission, and the user may want to limit those charges by setting a time limit).

[0225] One or more location requirements (used to restrict model transfer to a specific location or range of locations).

[0226] One or more transport quality of service requirements.

[0227] In some example embodiments, if the AI model transmission request 710 is made by the UE via an UL RRC message to transmit the AI model via the cellular network, such AI model transmission request may include a NAS message (e.g., PDU SESSION ESTABLISHMENT REQUEST). One embodiment of the NAS message is for requesting the CN to establish a PDU session for model transmission.

[0228] In some example embodiments, if the AI model transmission request 710 is made by the UE via a UL RRC message to transmit the AI model via a cellular network, and if the AI model transmission request may not include a NAS message (e.g., PDU SESSION ESTABLISHMENT REQUEST) used to request to establish a PDU session for model transmission, this means that the UE may select an IP network instead of a cellular network to perform model transmission.

[0229] Step 2: The responding network device (UE 702 or OAM 706) transmits an AI model transmission response message 712 to the requesting network device (OAM 706 or UE 702) via the base station 704 as a forwarding network device.

[0230] In some example implementations, the AI model transmission response 712 may be implemented as a UL RRC or DL RRC message. In some other alternative implementations, the AI model transmission response 712 may be implemented as a UL MAC CE or a DL MAC CE (eg, a model transmission response MAC CE).

[0231] In some example embodiments, the AI model transmission response 712 in the UL direction or the DL direction via an RRC message or a MAC CE may include at least one of the following:

[0232] Confirmation of the requested AI model.

[0233] Failure confirmation of the requested AI model.

[0234] Network type indication for model transmission: used to indicate the network type used for model transmission, such as NR network, WLAN or other IP network.

[0235] • Endpoint IP address on the UE side (when the UE is the responding network device).

[0236] In some example embodiments, if the AI model transmission response 712 is made by the UE via an UL RRC message to transmit the AI model via a cellular network, such AI model transmission response 712 may include a NAS message (e.g., PDU SESSION ESTABLISHMENT REQUEST) to indicate that the model transmission will be performed via a cellular network (e.g., an NR wireless network), if the model transmission response 712 may not include a NAS message (e.g., PDU SESSION ESTABLISHMENT REQUEST) to indicate that the model transmission will be formed via an IP network (e.g., WLAN) rather than a cellular network (e.g., an NR wireless network). One embodiment of the NAS message is for requesting the CN to establish a PDU session for model transmission.

[0237] Step 3: If the cellular path 720 has been selected by the UE, the base station 706 may start the PDU session establishment procedure for model transmission, and then perform AI model transmission between the UE and the gNB / OAM via the NR network.

[0238] In some example embodiments, the UE may initiate a PDU session establishment procedure for AI model transmission. In some other example embodiments, the OAM may initiate a PDU session establishment procedure for AI model transmission.

[0239] The PDU establishment process may, for example, include: the UE sending a request 722 for a tunnel to the core network 708; the core network 708 responding to the OAM and the base station, as shown in 724; the base station 704 sending an RRC reconfiguration message 726 to the UE, an RRC reconfiguration process; the base station receiving an RRC reconfiguration completion message 728 from the UE; and the base station sending a model transmission tunnel resource setup response message 729 to the further network 708.

[0240] like Figure 7 As shown, the UL AI model transmission via the cellular network may include UL transmission of the AI model from the UE to the core network via an established PDU session (as shown by 730) and subsequent transmission of the AI model from the core network to the OAM (as shown by 732).

[0241] like Figure 7 As shown, the DL AI model transmission via the cellular network may include transmission of the AI model from the OAM to the core network (as shown by 732) and subsequent transmission of the AI model from the core network to the UE via the established PDU session (as shown by 736).

[0242] Step 4: If another IP network path has been selected by the UE, the UE is at most responsible for implementing the transmission of the AI model via the IP network, such as Figure 7 As shown in 752. In some example embodiments, for DL AI model transmission (from gNB to UE), the UE may send ACK information to the cellular network to notify the cellular network of the completion of the model transmission via another IP network.

[0243] Transmission of AI network management models between the UE as an endpoint and the core network

[0244] In some example cases, the AI model trained and / or managed at the UE may need to be transmitted to the core network node, and vice versa. Such AI model transmission can be carried out at least partially in the air interface and partially in the core network. The following various example embodiments provide resources for configuring such AI model transmission and solutions for implementing the model transmission. Generally, the transmission of AI models involving the core network can be implemented in the control plane (CP) of the core network, the user plane (UP) of the core network, or a mixture of these two planes.

[0245] AI model transmission between UE and core network based on LTE positioning protocol in the control plane

[0246] In some example embodiments, in the control plane (CP), AI model transmission in the core network can be based on the LTE Positioning Protocol (LPP), both over the air interface and in the core network. The AI model transmission requesting device can be a core network node involved in the transmission, or a UE. Correspondingly, the AI model responding network device can be a UE or a core network node.

[0247] Example implementations are shown in Figure 8 The flowchart 800 includes an AI model transmission request-response process 810 or 820 (depending on whether the UE or the core network is the requester) and an AI model transmission process 830.

[0248] An implementation method of AI model transmission based on LPP may include the following example steps.

[0249] Step 1: AI Model Transmission The requesting device (UE 802 or core network node 808) may use the LPP protocol to send an AI model transmission request to the responding device (core network node 808 or UE 802), as shown in 812 or 822. The core network node 808 may be, for example, a Location Management Function (LMF) node. Therefore, this example embodiment is particularly suitable for transmitting AI models for positioning and location management.

[0250] In some example embodiments, the AI model transmission request 812 or 822 may be implemented as a UL or DL LPP message and may include at least one of the following information:

[0251] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0252] An AI model identification or identifier indicating the requested AI model. This identifier can be an AI model ID, or it can be an AI model description that can be used to uniquely identify the AI model to be transmitted, such as the model size and other model information.

[0253] Service information, which indicates the requested AI-based function service.

[0254] UE identifier, used to indicate the UE to the LMF to assign the appropriate AI model to the UE (when the UE is the requester).

[0255] ● Physical cell identifier or cell group identifier (CGI), used to indicate the cell to which the UE is connected.

[0256] Position information, indicating the current position obtained from legacy information.

[0257] Step 2: The AI model transmission response device (LMF 808 or UE 802) may use the LPP protocol to send a model transmission response message to the AI model transmission request device (UE 802 or LMF 808), as shown in 814 or 824.

[0258] In some example embodiments, the AI model transmission response 814 or 824 may be implemented as a UL or DL LPP message and may include at least one of the following information:

[0259] Confirmation of the requested AI model.

[0260] Rejection of the requested AI model.

[0261] One or more reasons for the rejection.

[0262] Step 3: After the AI model transmission response message is received by the requesting device (UE 802 or LMF 808 ), UE 802 and LMF start AI model transmission (UL or DL transmission) using the LPP protocol, as shown by 830 .

[0263] Further Figure 8 As shown in , the communication between the base station and the LMF may be indirect, as such communication passes through another intermediate core network node, such as Figure 8 The AMF network node 806 shown in .

[0264] In the above example implementation, the LPP protocol can be used. The transmission of the AI model therefore uses SRB resources in the air interface and control plane resources in the core network. For example, the core network interfaces involved in the transmission of the AI model can be the NG-C interface between the base station and the AMF and the NL1 interface between the AMF and the LMF. In other words, the LPP protocol path used to transmit the AI model can be expressed as LPP=SRB+NG-C+NL1. Therefore, in this example implementation, the transmission of the AI model occurs in the CP and air interface of the core network.

[0265] AI model transmission based on LTE positioning protocol between UE and core network in the control plane of the core network and the user plane of the air interface

[0266] In some example embodiments, the AI model transmission in the core network may also be based on LPP, but may occur in a mix of control plane (CP) and user plane (UP). Specifically, the AI model may be transmitted between the UE and the base station via DRB resources in the air interface in the UP, and then transmitted between the base station and the core network and within the core network in the CP. The AI model transmission requesting device may be a UE, or a core network node involved in the transmission. Correspondingly, the AI model responding network device may be a core network node or a UE.

[0267] Example implementations are shown in Figure 9 The flowchart 900 includes (1) an AI model transmission request-response process, which involves the UE sending an AI model transmission request 910 to a core network node (such as LMF 908) via a base station 904 and an AMF 906, and receiving an AI model transmission response 924 from the LMF 908; (2) a data tunnel and resource establishment process, which involves 912-922 after the AI model transmission request 910 and before the response 924; and (3) AI model transmission 926 in the air interface through UP resources and AI model transmission 928 between the base station and the core network and in the core network through CP.

[0268] Figure 9 Such an implementation based on LPP and using mixed UP and CP for transmission may include the following example steps.

[0269] Step 1: UE 902, as an AI model requester, may generate an AI model transmission request 910 and send it to LMF 908 via base station 904 and AMF 906 (other type of core network node) using the LPP protocol.

[0270] In some example embodiments, the AI model transmission request 910 may be transmitted as a UL LPP message and may include at least one of the following information:

[0271] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0272] An AI model identification or identifier or model information indicating the requested AI model. Such an identifier can be an AI model ID, or an AI model description that can be used to uniquely identify the AI model to be transmitted, such as the model size and other model information.

[0273] Service information, which indicates the requested AI-based function service.

[0274] UE identifier, used to indicate the UE to which the LMF instructs the appropriate AI model to be assigned (when the UE is the requester).

[0275] • Physical cell identifier or Cell Group Identifier (CGI), used to indicate the cell to which the UE is connected.

[0276] Position information, indicating the current position obtained from legacy information.

[0277] Step 2: Once the AI model transmission request is confirmed, the LMF 908 may generate an AI model transmission tunnel request 912 to the AMF 906 via, for example, the NL1 interface to obtain air interface resources / tunnels for the requested AI model transmission. Thus, the AI model transmission tunnel request may be an NL1 message and may include at least one of the following information:

[0278] AI model information, including, for example, the size of each requested AI model, and the number of AI models requested.

[0279] One or more QoS requirements for each requested AI model.

[0280] Step 3: If the received AI model transmission tunnel request is accepted, the AMF 906 may generate an AI model transmission resource request 914 and send it to the base station via, for example, the NG-C interface.

[0281] In some example embodiments, the AI model transmission resource request 914 may be structured to request air interface resources for one or more QoS flows for AI model transmission. Such a resource request may include at least one of the following information:

[0282] ●One or more PDU session identifiers to identify one or more PDU sessions associated with the current QoS flow for AI model transmission.

[0283] One or more QFIs, used to identify one or more QoS flows for model transmission.

[0284] Endpoint IP address for AI model transmission.

[0285] AI model information of the requested AI model, such as the identifier and size of the AI model.

[0286] AI-based function information, which indicates the AI-based function to which the requested AI model belongs.

[0287] Delay requirement, used to indicate the delay requirement for model transmission.

[0288] Location requirement: used to indicate the location requirement for model transmission.

[0289] In one embodiment, the AI model transmission resource request 914 may be configured to request air interface resources for one or more PDU sessions for AI model transmission. Such a resource request may include at least one of the following information:

[0290] One or more PDU session identifiers, used to indicate one or more PDU sessions for AI model transmission.

[0291] AI model information of the requested AI model, such as the identifier and size of the AI model.

[0292] Endpoint IP address for AI model transmission.

[0293] AI-based function information, which indicates the AI-based function to which the requested AI model belongs.

[0294] QoS requirements associated with the requested AI model, e.g. guaranteed bitrate, maximum bitrate, etc.

[0295] A list of QoS flows for each requested AI model associated with the current PDU Session.

[0296] In another example embodiment of steps 2 and 3, once the model transmission request 910 is confirmed, the AI model transmission resource request can be constructed to directly request air interface resources from the LMF 908 to the base station (combined 912 and 914). Such resource request may include at least one of the following information:

[0297] AI model information of the requested AI model, such as the identifier and size of the AI model.

[0298] AI-based function information, which indicates the AI-based function to which the requested AI model belongs.

[0299] ●QoS requirements for each requested AI model, including, for example, guaranteed bitrate, maximum bitrate, etc.

[0300] Latency requirements for each requested AI model.

[0301] Step 4: If the AI model transmission resource request is accepted, the base station 904 may then configure DRB resources for AI model transmission with the UE 902 via the RRC reconfiguration message 916 .

[0302] Step 5: The UE 902 may then send an RRC reconfiguration complete message 918 to the base station 904 when the RRC configuration is successfully performed.

[0303] In one embodiment of steps 4 and 5, if 912 and 914 exist separately, the RRC reconfiguration process of steps 4 and 5 may follow the same procedure as described above regarding QoS flow or PDU session depending on whether QoS flow or PDU session is requested. Figure 6 A similar process is described in the processes of 616, 618 and 620.

[0304] In another alternative example embodiment of steps 4 and 5, a direct AI model transmission tunnel request 913 may be transmitted from the LMF to the base station instead of separate steps 912 and 914. In that case, the RRC reconfiguration process of steps 4 and 5 may follow a similar process as described above with respect to process 510 of the DRB-based solution.

[0305] Step 6: When the RRC reconfiguration process is successfully completed and terminated, the base station 904 may send a model transmission resource response 920 to the AMF 906.

[0306] In the case where the AI model transmission resource request 914 involves one or more QoS flows as described above, it may include at least one of the following:

[0307] AMF UE NGAP Identifier, used to indicate the UE specific AMF ID via the NGAP interface.

[0308] RAN UE NGAP Identifier, used to indicate the UE specific RAN ID via the NGAP interface.

[0309] One or more QoS flow identifiers indicating the successful allocation of one or more QoS flows for the requested AI model.

[0310] ●One or more failed QoS flow identifiers to indicate failed allocation of QoS flows for the requested AI model.

[0311] ●One or more reasons for failed QoS flow allocation for the requested AI model.

[0312] In the case where the AI model transmission resource request 914 involves one or more PDU sessions as described above, it may include at least one of the following:

[0313] AMF UE NGAP Identifier, used to indicate the UE specific AMF ID via the NGAP interface (assuming the core network node 606 is an AMF).

[0314] RAN UE NGAP Identifier, used to indicate the UE specific RAN ID via the NGAP interface.

[0315] One or more PDU session identifiers to indicate the successful allocation of one or more PDU sessions for the requested AI model.

[0316] One or more failed PDU session identifiers, indicating one or more failed allocations of PDU sessions for the requested AI model.

[0317] One or more reasons for failed PDU session allocation for the requested AI model.

[0318] Step 7: The AMF 906 may generate an AI model transport tunnel response 922 to the LMF 908 via, for example, the NL1 interface based on the received AI model transport resource response 920.

[0319] In another alternative example embodiment of steps 6 and 7, the base station 904 may generate a direct model transmission resource response 923 and transmit it directly to the AMF 908 via, for example, the NRL and NG-C interfaces, in combination with Figure 9 For steps 920 and 922, such model transmission resource response 923 may include at least one of the following information:

[0320] One or more DRB indications or identifiers: used to indicate the successful allocation of one or more DRB resources used to transmit the requested AI model.

[0321] One or more failed DRB identifiers, indicating one or more failed allocations of DRB resources for the requested AI model.

[0322] One or more reasons why DRB resource allocation for the requested AI model failed.

[0323] Step 8: The LMF 908 may then generate an AI model transmission response 924 and send it to the UE 902 using the LPP protocol.

[0324] In some example embodiments, the AI model transmission response 924 may include at least one of the following information:

[0325] One or more PDU session identifiers, used to indicate one or more PDU sessions for AI model transmission.

[0326] ●One or more QoS flow identifiers to indicate one or more QoS flows for AI model transmission.

[0327] One or more DRB indications or identifiers indicating successful allocation of one or more DRB resources for transmission of the requested AI model.

[0328] AI model information of the requested AI model, such as the identifier and size of the AI model.

[0329] AI-based capabilities information associated with the requested AI model. In particular, for AI model transmission between UE and LMF, the AI-based capabilities may be for positioning management and enhancement.

[0330] Step 9: The LMF 908 and the UE 902 may then begin transmitting the requested AI model via the LPP protocol mixed with the user plane (UP) and control plane (CP). Specifically, transmission 926 between the UE 902 and the base station 904 may be via DRB resources in the UP, while transmission 928 between the base station 904 and the AMF 906 and within the core network may be via the CP. Transmission between the base station 904 and the AMF 906 may be via the NG-C interface, while transmission between the AMF 906 and the LMF 908 may be via the NL1 interface. This example hybrid LPP method may be represented as LPP_Mixed = DRB + NG-C + NL1.

[0331] For LMF as the requesting network device, Figure 10Another example implementation is shown in flowchart 1000 of , including (1) an AI model transmission request process, which involves the LMF sending an AI model transmission request 1010 to the UE 1002 via the AMF 1006 and the base station 1004, and receiving an AI model transmission response 1026 from the UE 1002 to the LMF 1008 via the base station 1004 and the AMF 1006; (2) a data tunnel and resource establishment process, which involves 1012-1024 after the AI model transmission request 1010 and before the response 1026; and (3) AI model transmission 1028 in the air interface through UP resources and AI model transmission 1030 between the base station and the core network and in the core network through CP.

[0332] Figure 10 Such an implementation of AI model transmission based on LPP and using a mixture of UP and CP may include the following example steps.

[0333] Step 1: The LMF 1008 may generate an AI model transmission request message 1010 and may send it to the UE 1002 via the LPP protocol. The requested AI model may be based on a positioning function associated with the LMF 1008, for example.

[0334] In some example embodiments, the AI model transfer request 1010 may be transmitted as a DL LPP message and may include at least one of the following information:

[0335] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0336] An AI model identification or identifier or model information indicating the requested AI model. Such an identifier can be an AI model ID, or an AI model description that can be used to uniquely identify the AI model to be transmitted, such as the model size and other model information.

[0337] Service information, which indicates the requested AI-based function service.

[0338] Step 2a: The UE 1002 may generate an AI model transmission tunnel request message 1012 based on the AI model transmission request message 1010 and when the AI model transmission is confirmed, and send it to the AMF 1006 to request an available tunnel for AI model transmission.

[0339] In some example embodiments, the AI model transmission tunnel request message 1012 may be an N1 message that uses NAS signaling to make it transparent to the base station 1004. The AI model transmission tunnel request message 1012 may request the establishment of one or more PDU sessions for AI model transmission. The AI model transmission tunnel request message 1012 may include at least one of the following information:

[0340] One or more PDU session identifiers, used to indicate one or more PDU sessions for AI model transmission.

[0341] The number of QoS flows requested for each PDU Session.

[0342] The number of AI models requested.

[0343] The size of each requested AI model.

[0344] One or more QoS requirements.

[0345] A flag indicating that this request message is for AI model transmission.

[0346] Step 2b: As an alternative to step 2a, the UE 1002 may generate another model transmission request message 1014 based on the AI model transmission request message 1010 when the AI model transmission is confirmed, and send it to the base station 1004 to request the resources available for the AI model transmission. The base station 1004 may further generate a model transmission tunnel request message 1016 based on the received model transmission request message 1014 and send it to the AMF 1006 to request tunnel resources for the AI model transmission. The model transmission request message 1014 may be implemented as an UL RRC message, for example, and may include at least the following information:

[0347] The number of AI models requested.

[0348] The size of each requested AI model.

[0349] In some example embodiments, the AI model transmission tunnel request message 1016 may be transmitted as NG-C signaling and may request modification of one or more PDU sessions to add one or more QFI flows for AI model transmission. Such an AI model transmission tunnel request message may include at least one of the following information:

[0350] One or more PDU session identifiers, used to indicate one or more PDU sessions.

[0351] • The number of QoS flows requested.

[0352] The number of AI models requested.

[0353] The size of each requested AI model.

[0354] ●One or more QoS requirements.

[0355] A flag indicating that this type of request message is for AI model transmission.

[0356] Step 3: AMF 1006 can generate an AI model transmission resource request message 1018 based on the received AI model transmission tunnel request message 1012 or 1016, and send it to the base station 1004.

[0357] Step 4: The base station may then begin the RRC reconfiguration process 1020 and 1022 to configure resources according to the received AI model transmission resource request message 1018. Depending on whether a QoS flow or a PDU session is requested, the RRC reconfiguration process of step 4 may follow the same procedure as described above regarding Figure 6 The process is similar to that described in 612, 618 and 620.

[0358] Step 5: The base station 1004 may then generate an AI model transmission resource response message 1024 after receiving the RRC reconfiguration complete message 1022 and send it to the AMF.

[0359] Step 6: UE 1002 may then generate and send an AI model transmission response message 1026 to LMF 1008 when the RRC reconfiguration procedure for establishing resources for model transmission is successfully completed and terminated. In some example embodiments, model transmission response message 1026 may be transmitted as an LPP message that may include at least one of the following information:

[0360] Confirmation of the requested AI model.

[0361] Rejection of the requested AI model.

[0362] One or more reasons for the rejection.

[0363] Step 7: LMF 1008 and UE 1002 may then perform AI model transmission using the LPP-Mixed protocol of LPP_Mixed=DRB+NG-C+NL1, where AI model transmission 1028 occurs in the user plane via DRB, and AI transmission 1030 occurs in the control plane between the base station and the core network and within the core network.

[0364] Generic AI model transmission between UE and core network in the control plane

[0365] The above LPP-based implementation can be extended to cases where LMF and LPP are not involved and thus become also suitable for UE / core network transmission of AI models associated with network functions other than positioning. Similarly, the transmission of AI models can occur entirely in the control plane of the network, including on the air interface, between the base station and the core network, and within the core network.

[0366] Such general CP-based model transmission schemes for the core network and UE as AI model requesting devices are shown in Figure 11 Flowchart 1100 and Figure 12 1200. Figure 8 Compared with the implementation method, Figure 11 and Figure 12 The implementation method does not involve LMF or LPP. Figure 11 In the , the core network (such as AMF) acts as the requesting device, and Figure 12 In both scenarios, the UE acts as the requesting device. They both include the AI model request-response process and the AI model transmission process in the control plane.

[0367] like Figure 11 As shown, this type of implementation with the core network node (taking AMF as an example) as the requesting device may include the following example steps.

[0368] Step 1: The AMF 1106 generates an AI model transfer request 1108 and sends it to the base station via, for example, the NG-C interface. It may contain at least one of the following information:

[0369] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0370] An AI model identification or identifier indicating the requested AI model. This identifier can be an AI model ID, or an AI model description that can be used to uniquely identify the AI model to be transmitted, such as the model size and other model information.

[0371] • Indicates the identifier of the target UE, such as AFM UE NGAP ID or RAN UE NGAP ID.

[0372] Step 2: In response to receiving the AI model transmission request 1108, the base station 1104 may generate an AI model transmission request 1110 and send it (eg, as a DL RRC message) to the target UE 1102 via SRB resources in the air interface according to the target UE identifier.

[0373] In one embodiment, the DL RRC message containing the AI model transmission request may include at least one of the following information:

[0374] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0375] An AI model identification or identifier indicating the requested AI model. This identifier can be an AI model ID, or an AI model description that can be used to uniquely identify the AI model to be transmitted, such as the model size and other model information.

[0376] Transaction ID.

[0377] In some example embodiments, upon receiving the AI model transmission request 1108 from the AMF, if an SRB has not been previously established before forwarding the AI model transmission request 1110 to the UE 1102 via the established SRB resources, the base station may first establish one or more dedicated SRBs for AI model transmission (as described above).

[0378] Step 3: If the requested model transmission is confirmed, the UE 1102 sends an AI model transmission response 1112 to the base station.

[0379] In some example embodiments, the AI model transmission response 1112 may be transmitted as a UL RRC message including at least one of the following information:

[0380] The data for the requested AI model.

[0381] · Validation of each AI model requested.

[0382] A rejection of each requested AI model.

[0383] The reason for the refusal, if any.

[0384] Transaction ID.

[0385] Step 4: The base station 1104 may then forward the AI model transmission response 1114 to the AMF 1116.

[0386] Step 5: The UE 1102 and the AMF 1106 may then start AI model transmission 1120 using the allocated SRB resources and the NG-C control interface between the base station and the AMF. This transmission may thus be expressed as: CP solution = SRB + NG-C.

[0387] Figure 12 Further shown is a flowchart 1200 for implementing AI model transmission between UE and core network (e.g., AMF) via the control plane, similar to Figure 11 , but the UE is the requesting device. Figure 12 An implementation of may include the following example steps.

[0388] Step 1: The UE 1202 may generate an AI model transmission request 1208 according to the AI-based function and send it to the base station 1204 via SRB resources.

[0389] In some example embodiments, the AI model transmission request 1208 may be sent only via SRB resources dedicated for AI model transmission. If such SRB resources are not configured, the UE's model transmission request may not be allowed. In some other example embodiments, the AI model transmission request may be sent via legacy SRB resources shared with the transmission of other signaling / control information.

[0390] In some example embodiments, the AI model transmission request message may be transmitted as a UL RRC message, which may include at least one of the following information:

[0391] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0392] An AI model identification or identifier indicating the requested AI model. This identifier can be an AI model ID, or an AI model description that can be used to uniquely identify the AI model to be transmitted, such as the model size and other model information.

[0393] Transaction identifier, used to indicate the AI transmission transaction process.

[0394] Step 2: The base station 1204 may then generate or forward an NG-C message AI Model Transfer Request 1210 for transmission to the AMF 1206. In addition to the information included in the UL RRC message, the UL RRC message may further include a target UE ID, such as an AMF UE NGAP ID or a RAN UE NGAP ID.

[0395] Step 3: If the requested model transmission is confirmed, the AMF 1206 may generate an AI model transmission response 1212 and send it to the base station 1204. In one embodiment, the AI model transmission response 1212 may be transmitted as an NG-C message, which may include at least one of the following information:

[0396] · Validation of each AI model requested.

[0397] Rejection of the requested AI model.

[0398] · The reason or reasons for the rejection.

[0399] AMF UE NGAP ID.

[0400] RAN UE NGAP ID.

[0401] The requested resources for AI model transfer.

[0402] Step 4: The base station may then generate an AI model transmission response message 1214 or forward it to the UE via an SRB in the air interface. In some embodiments, if an SRB has not been previously established before forwarding the AI model transmission response 1214 to the UE 1202 via the established SRB resources, the base station may first establish one or more dedicated SRBs for AI model transmission (as described above) by including an RRC reconfiguration message. In some other example embodiments, the AI model transmission response message may include at least one of the following information:

[0403] Confirmation of the requested model.

[0404] Rejection of the requested model.

[0405] The reason for the refusal.

[0406] Transaction ID.

[0407] RRC reconfiguration to establish one or more SRBs for AI model transmission.

[0408] Step 5: The UE 1202 and the AMF 1206 may then start AI model transmission via the allocated SRB resources in the air interface and in the control plane between the base station 1204 and the AMF 1206 via, for example, the NG-C interface. Thus, the transmission may be expressed as: CP solution = SRB + NG-C, similar to Figure 11 In some example embodiments, a NAS message within an RRC message may be used for AI model transmission, meaning that the AI model information may be transparent to the base station.

[0409] Using virtual DNN or OAM in the user plane for AI model transmission between UE and core network

[0410] For general UP-based AI model transmission terminated between the UE and the CN, a virtual DNN (vDNN) or OAM in the CN can be used for this model transmission. In one embodiment, the virtual DNN or OAM can connect to the AMF using a private interface. In another embodiment, the virtual DNN or OAM can connect to the UPF using a private interface. In this case, the normal UP data tunnel establishment process can be relied upon to allocate and configure resources for AI model transmission.

[0411] The example is shown in Figure 13 In the flowchart 1300, a request for AI model transmission is made from the network side (e.g., from vDNN or OAM 1310). This type of solution utilizes PDU sessions to transmit AI models. This transmission can thus be performed entirely in the user plane. Figure 13 An example implementation may include a model transmission request-response process 1312-1318, a PDU session and PDU session resource establishment process 1320 and 1322, and an AI model transmission 1324, which may include the following steps.

[0412] Step 1: The AMF may first send an AI model transfer request 1314 to the base station 1304. In some embodiments, such a request may be obtained by the AMF 1306 from the OAM and / or vDNN 1310 via a private interface. In some example embodiments, the AI model transfer request 1314 may be transmitted as an NG-C message and may include at least one of the following information:

[0413] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0414] Information about each requested AI model, including but not limited to a model ID or model description.

[0415] ●QoS (Quality of Service) requirements for each requested AI model, e.g. guaranteed bitrate, maximum bitrate, etc.

[0416] Target UE identifier, including but not limited to AMF UE NGAP ID or RAN UE NGAP ID.

[0417] Endpoint IP address, used to indicate the endpoint of the IP address used for UL model transmission.

[0418] Step 2: The base station 1304 may then generate an AI model transmission request message 1316 to the UE 1302 .

[0419] In some example embodiments, the AI model transmission request message 1316 may be transmitted as a DL RRC message and may include at least one of the following information:

[0420] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0421] Information about the requested AI model, including but not limited to the model ID or model description.

[0422] Endpoint IP address, used to indicate the endpoint of the IP address used for UL model transmission.

[0423] Transaction ID.

[0424] Step 3: If the model transfer request is confirmed, the UE 1302 generates an AI model transfer response message 1318 to the AMF 1306.

[0425] In some example embodiments, the AI model transmission response message 1318 may be transmitted as a UL RRC message and may include a NAS message (i.e., PDU SESSION ESTABLISHMENT REQUEST), which is sent to the AMF 1306 to trigger the PDU session establishment procedure for AI model transmission.

[0426] In some example embodiments, the model transfer response message may include at least one of the following information:

[0427] Confirmation of the requested AI model.

[0428] Rejection of the requested AI model, if any.

[0429] One or more reasons for the rejection.

[0430] Endpoint IP address: used to indicate the endpoint of the IP address used for model transmission.

[0431] Step 4: The AMF 1306 may then start the PDU session establishment procedure 1320 within the core network based on the response message 1318 from the UE 1302.

[0432] Step 5: After the AMF 1306 starts the PDU session establishment process, the UE 1302, the base station 1304 and the AMF 1306 may start the PDU session resource establishment process 1322 to configure DRB resources for the PDU session.

[0433] Step 6: After the PDU session resource request process 1322 (within which at least one DRB resource was established for the PDU session) is successfully terminated, the UE 1302 and the vDNN / OAM 1310 may then start model transmission 1324 via the UPF 1308 using the established PDU session.

[0434] Another example is shown in Figure 14 In the flowchart 1400, the AI model transmission request is made from the UE side. This type of solution also uses PDU sessions to transmit the AI model. This transmission can thus be performed entirely in the user plane. Figure 14 An example implementation may include model transmission request-response processes 1404 and 1410, PDU session and PDU session resource establishment processes 1406 and 1408, and AI model transmission 1324, which may include the following steps.

[0435] Step 1: UE 1402 may first send an AI model transmission request to AMF 1406 via base station 1404. In some example embodiments, AI model transmission request 1404 may include at least one of the following information:

[0436] Message used to request the AMF to establish a PDU session for AI model transmission (e.g., PDU SESSION ESTABLISHMENT REQUEST).

[0437] An AI-based function indication, used to indicate the AI-based function to which the AI model to be transmitted belongs (the AI-based function may be one of the following, for example, an AI-based beam management function, an AI-based CSI enhancement function, or an AI-based positioning enhancement function).

[0438] Information about the requested AI model, including but not limited to the model ID or model description.

[0439] ●QoS requirements for each requested AI model, e.g., guaranteed bitrate, maximum bitrate, etc.

[0440] Step 2: If the request from UE 1402 can be confirmed by AMF 1406, the AMF can start the PDU session establishment procedure 1406 in the core network according to the received PDU SESSION ESTABLISHMENT REQUEST.

[0441] Step 3: The AMF 1406, base station 1404 and UE 1402 may then start the PDU session resource establishment procedure 1408 to establish DRBs in the air interface for the requested AI model transmission.

[0442] Step 4: If the DRB resources are successfully established, the AMF 1406 may generate an AI model transmission response message 1410 to the UE 1402.

[0443] In some example embodiments, the AI model transmission response message 1410 may be an N1 message and may include at least one of the following information:

[0444] Confirmation of the requested AI model.

[0445] Failure confirmation of the requested AI model,

[0446] · The reason or reasons for the rejection.

[0447] Step 5: The UE 1402 and the vDNN or OAM 1410 may then start the AI model transmission 1412 via the UPF, i.e., in the air interface and in the UP between the base station 1404 and the core network and in the core network over the established DRB resources.

[0448] The above description and accompanying drawings provide specific example embodiments and implementations. However, the subject matter described may be embodied in a variety of different forms, and therefore, the subject matter covered or claimed is intended to be construed as not being limited to any example embodiment set forth herein. A reasonably broad scope of the subject matter is intended to be claimed or covered. For example, the subject matter may be embodied as a method, device, component, system, or non-transitory computer-readable medium for storing computer code, among other things. Thus, the embodiments may, for example, take the form of hardware, software, firmware, storage media, or any combination thereof. For example, the method embodiments described above may be implemented by a component, device, or system comprising a memory and a processor by executing computer code stored in the memory.

[0449] Throughout the specification and claims, terms may have nuanced meanings that are suggested or implied by the context, beyond those explicitly stated. Likewise, the phrase "in one embodiment / implementation" as used herein does not necessarily refer to the same embodiment, and the phrase "in another embodiment / implementation" as used herein does not necessarily refer to a different embodiment. It is intended that, for example, claimed subject matter includes all or part of the combination of the example embodiments.

[0450] In general, terms can be understood, at least in part, from their usage in the context. For example, terms such as "and," "or," or "and / or," as used herein, can include multiple meanings, which can depend, at least in part, on the context in which such terms are used. Typically, "or," if used in a list of associations, such as A, B, or C, is intended to mean A, B, and C, where used in an inclusive sense, and A, B, or C, where used in an exclusive sense. Furthermore, the term "one or more," as used herein, can be used to describe any feature, structure, or characteristic in a singular sense or can be used to describe a combination of features, structures, or characteristics in a plural sense, depending, at least in part, on the context. Similarly, terms such as "a," "an," or "the" can be understood to convey singular usage or to convey plural usage, depending, at least in part, on the context. Furthermore, the term "based on" can be understood to not necessarily be intended to convey an exclusive set of factors and can instead allow for the presence of additional factors that are not necessarily explicitly described, again, depending, at least in part, on the context.

[0451] References throughout this specification to features, advantages, or similar language do not imply that all features and advantages that can be achieved with the present solution should be or are included in any single embodiment of the solution. Rather, language referring to features and advantages is understood to mean that a specific feature, advantage, or characteristic described in connection with an embodiment is included in at least one embodiment of the present solution. Thus, discussions of features and advantages, and similar language, throughout this specification may, but do not necessarily, refer to the same embodiment.

[0452] Furthermore, the described features, advantages, and characteristics of the present solution may be combined in any suitable manner in one or more embodiments. Those skilled in the relevant art will recognize, based on the description herein, that the present solution may be practiced without one or more of the specified features or advantages of a particular embodiment. In other instances, additional features and advantages may be recognized in certain embodiments that may not be present in all embodiments of the present solution.

Claims

1. A method for communicating with a second device via an air interface in a wireless access network of a wireless communication system by a first device, the first device and the second device being one of a wireless terminal and the other of a wireless base station, respectively, and comprising: performing a request-response procedure with the second device over the air interface to initiate transmission of an artificial intelligence (AI) network management model therebetween; determining radio bearer resources in the air interface for transmission of the AI network management model based on the request-response procedure and the core network of the wireless communication system not responding to the request-response procedure; and interacting with the second device to transmit the AI network management model through the radio bearer resources.

2. The method according to claim 1, wherein The request-response process includes transmitting a request for the AI network management model in the form of a radio resource control (RRC) message or a medium access control (MAC) control element (MAC-CE).

3. The method according to claim 2, wherein: The request is transmitted in the form of an RRC message when dedicated radio bearer resources in the air interface have not been previously allocated for the purpose of AI network management model transmission.

4. The method according to claim 2, wherein: The request is transmitted in the form of a MAC-CE when dedicated radio bearer resources in the air interface have been previously allocated for the purpose of AI network management model transmission.

5. The method according to claim 2, wherein: The request is transmitted in the form of a MAC-CE, and a scheduling request (SR) is triggered when the MAC CE is pending and no physical uplink shared channel (PUSCH) resources are available to accommodate the MAC CE.

6. The method according to claim 2, wherein: The request includes an RRC reconfiguration message for allocating new radio bearer resources by the wireless base station for transmission of the AI network management model.

7. The method according to claim 2, wherein: The request is transmitted in the form of an RRC message, wherein the RRC message includes at least one of the following: a first identifier for identifying an AI-based function associated with the AI network management model; A second identifier, used to indicate the AI network management model; A third identifier, used to identify the radio bearer resource used for transmission of the AI network management model; or A process identifier, used to identify the transmission of the AI network management model.

8. The method according to claim 2, wherein: The request is transmitted in the form of a MAC-CE, where the MAC-CE includes at least one of the following: a first identifier for identifying an AI-based function associated with the AI network management model; A second identifier, used to indicate the AI network management model; a third identifier, used to identify the radio bearer resource used for transmission of the AI network management model; a fourth identifier, used to identify a PDU session for transmission of the AI network management model; or A Quality of Service (QoS) Flow Identifier (QFI) is used to indicate a QoS flow used for transmission of the AI network management model.

9. The method according to any one of claims 2 to 8, wherein The radio bearer resources include signaling radio bearer (SRB) resources.

10. The method according to any one of claims 2 to 8, wherein The radio bearer resources include data radio bearer (DRB) resources.

11. The method according to claim 10, wherein: The DRB resources are allocated by the wireless base station for exclusive use in AI network management model transmission, as indicated by a configuration associated with the DRB resources, an SDAP (Service Data Adaptation Protocol) configuration associated with the DRB resources, a PDCP (Packet Data Convergence Protocol) configuration associated with the DRB resources, or a flag in an RLC (Radio Link Control) configuration associated with the DRB resources.

12. The method according to claim 10, wherein: The DRB resources may be allocated for additional usage in addition to AI network management model transmission, and wherein usage of the DRB resources for AI network management model transmission is indicated by: A flag in the header of an SDAP, PDCP or RLC PDU; or The AI model transfer specific QFI list in the SDAP configuration associated with the DRB resource.

13. A method for communicating, by a first device, with a second device via an air interface in a radio access network of a wireless communication system, the first device and the second device being one and the other of a wireless terminal and a wireless base station, respectively, and comprising: interacting with the second device over the air interface in a request process for transmitting at least one AI network management model, the request process being configured to trigger a data tunnel application process between the wireless base station and a core network (CN) node of the wireless communication system, and triggering allocation of radio communication resources in the air interface according to the data tunnel application process; as well as interacting with the second device to transmit the at least one AI network management model over the radio communication resources.

14. The method according to claim 13, wherein The request process includes transmitting a request for transmission of the at least one AI network management model in the form of an RRC message or a MAC-CE.

15. The method according to claim 14, wherein The request is transmitted from the wireless terminal to the wireless base station and includes a non-access stratum (NAS) message from the wireless terminal to the core network node for triggering the data tunnel application process.

16. The method according to claim 14, wherein The request is transmitted in the form of an RRC message, wherein the RRC message includes at least one of the following: a first identifier for identifying an AI-based function associated with the at least one AI network management model; a second identifier, used to indicate the at least one AI network management model; or Endpoint IP address of the mobile terminal.

17. The method according to claim 14, wherein: The request is transmitted in the form of a MAC CE, and the SR is triggered when the MAC CE is pending and no PUSCH resources are available to accommodate the MAC CE.

18. The method according to claim 14, wherein The request is transmitted in the form of a MAC CE, where the MAC CE includes at least one of the following: a first identifier for identifying an AI-based function associated with the at least one AI network management model; a second identifier, used to indicate the at least one AI network management model; a DRB identifier, configured to identify a DRB resource used to transmit the at least one AI network management model; A PDU session identifier, used to identify a PDU session used to transmit the at least one AI network management model; or A QFI is used to indicate a QoS flow used to transmit the at least one AI network management model.

19. The method according to claim 14, wherein The request is transmitted in the form of an RRC message when data tunnel resources have not been previously allocated in the air interface for the purpose of AI network management model transmission.

20. The method according to claim 14, wherein The request is transmitted in the form of a MAC-CE when at least one data tunnel resource has previously been configured in the air interface for the purpose of AI network management model transmission.

21. The method according to claim 20, wherein The at least one data tunnel resource includes a DRB resource, a QoS flow of a PDU session, or a PDU session for AI network management model transmission.

22. The method according to claim 13, wherein The data tunnel application process includes transmitting a data tunnel request for one or more QoS flows of a PDU session from the wireless base station to the CN node, wherein the data tunnel request includes at least one of the following: A first identifier, used to indicate a mobile terminal specific access management function (AMF) ID relative to an NG application protocol (NGAP) interface; a second identifier for indicating a mobile terminal specific Radio Access Network (RAN) ID relative to the NGAP interface; The PDU session identifier associated with the requested QoS flow; The number of requested QoS flows for each PDU session; The number of AI network management models requested; Endpoint IP address for model transfer; The size of each requested AI Network Management model; or QoS request.

23. The method according to claim 22, wherein The response to the data tunnel request from the CN node to the wireless base station includes at least one of the following: a first identifier; a second identifier; PDU session identifier; a third identifier for indicating QoS flow allocation for each requested AI network management model; a fourth identifier for indicating failed QoS flow allocation for each requested AI network management model; or At least one reason for failed QoS flow allocation.

24. The method according to claim 23, wherein Interacting with the second device includes transmitting or receiving an RRC reconfiguration for mapping one or more QoS flows to a transmission of at least one AI network management model via at least one of the following: The AI model transmits a list of specific QoS flows; The presence of a SDAP PDU header configuration; or Mapping a current QFI list in the SDAP configuration to a transmitted enable flag of the at least one AI network management model.

25. The method according to claim 13, wherein The data tunnel application process includes transmitting a data tunnel request for one or more PDU sessions from the wireless base station to the CN node, wherein the data tunnel request includes at least one of the following: A first identifier indicating the mobile terminal specific AMF ID relative to the NGAP interface; a second identifier for indicating a mobile terminal specific RAN ID relative to the NGAP interface; Endpoint address for model transfer; the number of the one or more PDU sessions; The size of each requested AI network management model; mobile terminal identifier; and QoS request.

26. The method according to claim 25, wherein The response from the CN node to the wireless base station to the data tunnel request includes at least one of the following: the first identifier; the second identifier; The PDU session identifier used for model transmission; a third identifier for indicating failed PDU session allocation for each requested AI network management model; or At least one reason for a failed PDU Session allocation.

27. The method of claim 26, wherein interacting with the second device comprises transmitting or receiving an RRC reconfiguration for mapping the one or more PDU sessions to transmission of the at least one AI network management model via at least one of: The first flag in the DRB configuration indicates that the DRB is used for AI network management model transmission; The second flag in the SDAP configuration indicates that the corresponding DRB is used for AI network management model transmission; The third flag in the PDCP configuration indicates that the corresponding DRB is used for AI network management model transmission; or The fourth flag in the RLC configuration indicates that the corresponding DRB is used for AI network management model transmission.

28. A method performed by a first device of a radio access network (RAN) of a wireless communication system, interacting with a core network (CN) node of the wireless communication system to process an application process for transmitting an AI network management model between the first device and a second device, wherein the first device and the second device are respectively one of a mobile terminal and the other of a wireless base station, and the method comprising: interacting with the second device and the CN node to perform the application process, the application process involving a first request-response process between the mobile terminal and the radio base station via an air interface of the RAN and a second request-response process between the CN node and the radio base station; as well as Trigger a data tunnel application process between the UE and the CN node to facilitate transmission of the AI network management model between the first device and the second device.

29. A first device comprising a memory for storing instructions and a processor for executing the instructions to implement any one of claims 1 to 28.

30. A computer-readable non-transitory medium for storing computer instructions, the computer instructions being configured to cause the first device of claims 1 to 28 to implement any one of claims 1 to 28 when executed by a processor of the first device.

Images