CN-122003905-A - Network-assisted method of indirect ML LCM operation

Abstract

A method of pre-configuring AI/ML operation when enabling a multi-connectivity link in a wireless mobile communication system including a base station (e.g., a gNB) and a mobile station (e.g., a UE) is presented. If the AI/ML model is applied to a radio access network, model performance such as reasoning and/or training is dependent on different model execution environments with different device capabilities. Thus, by reconfiguring any model that operates with multiple devices including relay links, the potential performance impact due to dynamic changes in the applicable model conditions of LCM model operation can be reduced as model performance increases.

Inventors

• Jin Huchen
• A. ANDRE
• R. SHAH
• R. George Stephen

Assignees

• 欧摩威德国有限公司

Dates

Publication Date

20260508

Application Date

20241014

Priority Date

20231027

Claims (18)

1. 1. A method of network assisted indirect ML LCM operation in a wireless network including at least a target UE and at least one relay UE located at a coverage location of a base station, the method comprising the steps of: Initiate direct execution of a dual-sided AI/ML model between the base station and the target UE, during which LCM operations are performed on the target UE, and model management is performed on the base station, Criteria configured for successful execution of the LCM operation to achieve a target performance, The ML condition of the target UE is measured, Comparing the target UE ML condition with the criterion, Generating a decision indication message indicating whether the target UE ML condition satisfies the criterion and achieves the target performance, Estimating a relay UE ML capability of each of the at least one relay UE, Each relay UE ML capability is compared to the criteria, Selecting at least a first relay UE from the at least one relay UE if the decision indication message indicates that the target UE ML condition does not meet the criterion, a comparison of relay UE ML capabilities of the at least first relay UE with the criterion indicating that the at least first relay UE is capable of performing the LCM operation and achieving the target performance, Sending a collaboration indication message to the target UE, thereby indicating that the LCM operation is to be performed on the at least first relay UE, Stopping the LCM operation performed on the target UE, Transmitting an ML configuration to the at least first relay UE via system information or dedicated RRC message, thereby configuring the LCM operation, and The LCM operation is performed by the at least first relay UE, thereby activating an indirect LCM operation.
2. 2. The method of claim 1, wherein the wireless network comprises a plurality of target UEs for which LCM operations are performed indirectly on the at least first relay UE.
3. 3. The method of claim 1 or 2, wherein the base station generates the decision indication message based on a target UE ML condition update message, the target UE ML condition update message being based on an on-device measurement of the target UE ML condition.
4. 4. The method according to claim 1 or 2, wherein the target UE generates the decision indication message based on-device measurements of ML conditions of the target UE.
5. 5. The method according to claim 3 or 4, characterized by sending a periodic message to the target UE and/or the at least one relay UE, the periodic message defining the time to perform the target UE ML condition measurement and/or the relay UE ML capability estimation.
6. 6. The method according to one of the preceding claims, wherein at least one additional UE device performs the LCM operation in combination with a target UE or multiple target UEs.
7. 7. The method according to one of the preceding claims, wherein if the at least first relay UE cannot support the LCM operation alone, selecting a plurality of relay UEs to jointly perform the LCM operation is determined based on a side link ML condition update message from the at least one relay UE comprising an estimated relay UE ML capability.
8. 8. The method according to one of the preceding claims, characterized in that the criterion corresponds to a threshold value, which is configured according to an evaluation of the target UE ML condition to perform any given specific LCM operation.
9. 9. The method of claim 8, wherein the collaboration indication message indicates: If the measured ML condition of the target UE is lower than the corresponding threshold value, performing indirect LCM operation; If the measured target UE ML condition is higher than or equal to the corresponding threshold, direct LCM operation is continued or performed.
10. 10. The method according to one of the preceding claims, characterized in that the target UE measures the target UE ML conditions in a non-periodic way and sends a target UE ML condition update message to the base station, triggering the step of generating the decision indication message.
11. 11. The method according to one of the preceding claims, wherein the generation of the decision indication message is based on one of network traffic congestion or network level power saving.
12. 12. The method of one of the preceding claims, wherein the target UE ML condition is based on an element related to supporting the LCM, the element being at least one of: Device specific HW-/SW ML capability; non-device specific ambient ML conditions and LCM conditions; The target UE calculates processing power, memory size, battery power, tx/Rx configuration/setup information; Availability of frames/libraries required to perform the LCM operation; Geographic location, wireless connection link information, device mobility, neighboring network/device ML information, side link channel quality; list of supported models, open/proprietary model formats, model structure support, LCM operability.
13. 13. The method according to one of the preceding claims, characterized in that the at least first relay UE is configured to provide ML condition updates to the base station periodically or aperiodically based on relay ML configuration information.
14. 14. An apparatus for sidelink positioning in a wireless communication system, the apparatus comprising a wireless transceiver, a processor coupled with a memory, the memory storing therein computer program instructions configured to implement the steps of the method of claims 1 to 13, and the apparatus being designed for use in a base station (gNB).
15. 15. An apparatus for sidelink positioning in a wireless communication system, the apparatus comprising a wireless transceiver, a processor coupled with a memory, the memory storing therein computer program instructions configured to implement the steps of the method of claims 1 to 13, and the apparatus being designed for use in a User Equipment (UE).
16. 16. A base station (gNB) comprising the apparatus of claim 14.
17. 17. A user equipment comprising the apparatus of claim 15.
18. 18. A wireless communication system comprising at least one base station (gNB) according to claim 16 and at least one User Equipment (UE) according to claim 17.

Description

Network-assisted method of indirect ML LCM operation Technical Field The present disclosure relates to AI/ML operation pre-configuration, in which techniques for reconfiguring and signaling specific information to improve efficient signaling of a side-link based model are presented. Background In 3GPP (3 rd Generation partnership project), one of the selected research projects as an approved version 18 package is AI/ML (Artificial Intelligence/machine learning), as described in the related document (RP-213599) submitted in 3GPP TSG RAN (technical Specification group radio Access network) conference #94 e. The formal name of the AI/ML Study item is "Study on AI/ML for NR AIR INTERFACE (Study on AI/ML of NR air interface)", and the current RANs WG1 and WG2 are actively formulating specifications. The goal of this research project is to establish a generic AI/ML framework and the field of earning revenue using AI/ML based technology and use cases and LCM (lifecycle management). According to 3GPP, the main goal of this research project is to exploit the target use case to study the AI/ML framework for the air interface by taking into account performance, complexity and potential specification effects. In particular, the AI/ML model, terms and descriptions used to establish the common and specific characteristics of the framework will be one of the key working areas. With respect to the AI/ML framework, surveys are being conducted considering various aspects, and one of the key items is lifecycle management with respect to AI/ML models, wherein multiple phases are forcibly included for model training, model deployment, model reasoning, model monitoring, model updating, etc. Earlier, in 3GPP TR 37.817, version 17, entitled "Study on enhancement for Data Collection for NR and EN-DC (study of NR and EN-DC data collection enhancement)", UE movement was also seen as one of the AI/ML use cases, and one of the scenarios of model training/reasoning was that both functions are located within the RAN node. Subsequently, in release 18, a new work item "ARTIFICIAL INTELLIGENCE (AI)/MACHINE LEARNING (ML) for NG-RAN (artificial intelligence (AI)/Machine Learning (ML) for NG-RAN))" is initiated to specify data collection enhancements and signaling support within existing NG-RAN interfaces and architectures, including mobility optimization as one of the targeted use cases. For the standardization work of the above activities, the UE ML conditions for supporting the RAN-based AI/ML model may be considered very important for both the gNB and the UE to meet any desired model operation (e.g., model training/reasoning/selection/switching/updating/monitoring, etc.). Currently, no specification is defined for signaling methods or gNB-UE behavior with respect to split LCM distribution over side link relay links as RAN-based AI/ML model operations continue. Therefore, it is necessary to study any canonical effect by considering the model operation through the side link relay link. There is also a need to address any mechanism of additional signaling methods and/or gNB-UE behavior to support relay-based model operation between gNB and UE such that any potential impact of UE ML conditions on model operation in the RAN should be minimized under service continuity. On the other hand, in 3GPP, the terms in the working list contain a high-level description set on AI/ML model training, reasoning, verification, testing, UE-side model, network-side model, single-side model, double-side model, etc. The UE-side model and the network-side model indicate that the AI/ML model is located on the UE-side and the network-side, respectively, for operation. In a similar context, a single-sided model and a double-sided model indicate that the AI/ML model is located on one side and on both sides, respectively. All signaling aspects supporting the above have not been specified so far, as the definition of terms is still under discussion for further modification. Any potential standard impact under the new or enhanced mechanisms supporting the AI/ML model with the above-described work list items is one of the key areas to investigate in the AI/ML study. US 2021 203 565 A1 discloses methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training and using a machine learning model to classify network traffic as IoT traffic or non-IoT traffic and to manage traffic based on the classification. In some implementations, the machine learning parameters of the local machine learning model trained by the edge devices are received from each of at least a subset of the edge devices in the set of edge devices. The machine learning parameters received from the edge device are parameters of a local machine learning model trained by the edge device based on local network traffic handled by the edge device and used to classify network traffic as internet of things (IoT) traffic or non-IoT traffic. The machine learning parameters