[go: up one dir, main page]

WO2025020076A1 - Ue capability report for ai/ml positioning - Google Patents

Ue capability report for ai/ml positioning Download PDF

Info

Publication number
WO2025020076A1
WO2025020076A1 PCT/CN2023/109092 CN2023109092W WO2025020076A1 WO 2025020076 A1 WO2025020076 A1 WO 2025020076A1 CN 2023109092 W CN2023109092 W CN 2023109092W WO 2025020076 A1 WO2025020076 A1 WO 2025020076A1
Authority
WO
WIPO (PCT)
Prior art keywords
positioning
model
assisted
capabilities
network
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
PCT/CN2023/109092
Other languages
French (fr)
Inventor
Mingwei Jie
Chiao-Yao CHUANG
Pengli YANG
Rao Dai
Xuancheng Zhu
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
MediaTek Singapore Pte Ltd
Original Assignee
MediaTek Singapore Pte Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by MediaTek Singapore Pte Ltd filed Critical MediaTek Singapore Pte Ltd
Priority to PCT/CN2023/109092 priority Critical patent/WO2025020076A1/en
Publication of WO2025020076A1 publication Critical patent/WO2025020076A1/en
Pending legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W64/00Locating users or terminals or network equipment for network management purposes, e.g. mobility management
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00Machine learning
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/04Architecture, e.g. interconnection topology
    • G06N3/045Combinations of networks
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • G06N3/08Learning methods
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W24/00Supervisory, monitoring or testing arrangements
    • H04W24/02Arrangements for optimising operational condition
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W8/00Network data management
    • H04W8/22Processing or transfer of terminal data, e.g. status or physical capabilities
    • H04W8/24Transfer of terminal data

Definitions

  • This present disclosure relates generally to wireless communications, and more specifically, to techniques of positioning a user equipment (UE) with Artificial Intelligence (AI) /Machine Learning (ML) .
  • UE user equipment
  • AI Artificial Intelligence
  • ML Machine Learning
  • AI/ML based positioning includes training data collection, model training, model inference and model performance monitoring. There are many different methods and parameters in AI/ML based positioning. UE will support different capabilities based on different requirements. If network can get UE capability information, it will schedule corresponding positioning flow. UE capability reporting is important topic in AI/ML positioning.
  • the first capability is AI/ML positioning modes supported by the UE.
  • AI/ML positioning includes UE-based and UE-assisted positioning, UE-side and non-UE-side model, direct AI/ML and AI/ML assisted positioning.
  • the second capability is the dataset type of AI/ML training input data supported by the UE.
  • the training input could be channel impulse response (CIR) , power delay profile (PDP) and delay profile (DP) of the channel between gNB and UE.
  • CIR channel impulse response
  • PDP power delay profile
  • the third capability is sampling period supported by the UE. Sampling period is related to bandwidth. Different sampling period will cause different performance.
  • the fourth capability is approaches of reduced number of TRPs as model input supported by the UE.
  • the set of TRPs used in AI/ML could be fixed or change dynamically.
  • the fifth capability is for semi-supervised learning. Semi-supervised learning is helpful for improving the positioning accuracy when the same amount of ideal labelled data is used for semi-supervised learning, and the number of ideal labelled data is limited.
  • the sixth capability is for AI/ML model fine-tuning. If UE support model fine-tuning, network will send training data to UE to fine-tune the model.
  • the seventh capability is for AI/ML model monitoring. Model monitoring include monitoring based on provided ground truth label and monitoring without ground truth label.
  • Figure 1 illustrates indoor factory deployment scenario.
  • FIG. 2 illustrates the capabilities transfer procedure
  • FIG. 3 illustrates the sampling period
  • Figure 4 summaries the capabilities of UE reporting.
  • Figure 1 shows an example of indoor factory deployment scenario.
  • There are 18 gNBs (N 18) in indoor factory and the location of all gNBs are fixed. UEs are random distributed in the factory. PRUs (Positioning Reference Unit) as special UEs are deployed in known location for AI/ML training data collection.
  • PRUs Positioning Reference Unit
  • Case 1 UE-based positioning with UE-side model, direct AI/ML or AI/ML assisted positioning;
  • Case 2a UE-assisted/LMF-based positioning with UE-side model, AI/ML assisted positioning;
  • Case 2b UE-assisted/LMF-based positioning with LMF-side model, direct AI/ML positioning;
  • Case 3a NG-RAN node assisted positioning with gNB-side model, AI/ML assisted positioning;
  • Case 3b NG-RAN node assisted positioning with LMF-side model, direct AI/ML positioning.
  • AI/ML model can be deployed in UE-side, Non-UE-side (gNB-side, or LMF-side) .
  • the general positioning includes UE-based and UE-assisted.
  • UE-based means UE may perform measurement and also estimate the own position.
  • UE-assisted means UE may perform measurement and may not estimate the own position.
  • AI/ML positioning includes direct AI/ML and AI/ML assisted positioning.
  • Direct AI/ML positioning means AI/ML model output UE location directly.
  • AI/ML assisted positioning means AIML model output intermediate results. Intermediate results could be time of arrival (TOA) of the channel path between gNB and UE, LOS or NLOS indicator of the channel path between gNB and UE.
  • TOA time of arrival
  • the first capability is AI/ML positioning modes supported by the UE. These modes include:
  • 0101 means UE support “UE-based positioning with UE-side model, AI/ML assisted” and “UE-assisted positioning with Non-UE-side model” . If UE support “UE-assisted positioning with Non-UE-side model” , UE can report measurement results (CIR/PDP/DP) to network for UE-assisted positioning.
  • CIR/PDP/DP measurement results
  • AI/ML training data include AI/ML model input data and ground truth label.
  • gNB can send positioning reference signal to UE and UE will estimate the channel delay profiles based on received reference signal. These estimated channel delay profiles can be used as AI/ML model input. Every measurement of channel delay profile includes delay, power, and phase.
  • CIR is composed of a list of per-sample measurements of delay, power, and phase.
  • PDP is composed of a list of per-sample measurements of delay and power.
  • DP is a degenerated version of PDP, where the path power is not provided. It is only composed of a list of selected delays. The overhead of DP is smallest because it only includes delay information.
  • the AI/ML model input could be one of CIR, PDP and DP.
  • the second capability reporting is the supported dataset type of AI/ML model input. If one UE only support PDP as model input, it will report to network the capability. Network will send PDP-format training data to this UE. If one UE support all three formats as model input, network can send one of format training data to UE. UE also can send request to indicate which format it wants for training. Three bits can be used to indicate the format supported or not.
  • the sampling period is given by multiple of Tc.
  • the sampling period is 64Tc (32.55ns) .
  • the sampling period is 16Tc (8.14ns) .
  • the third capability reporting is the supported sampling period. After UE report the supported sampling period to network, network will send training data with corresponding sampling period to UE.
  • Supported sampling period could be a set of sampling period values, or the indication of the smallest sampling period that could be supported.
  • the number of TRPs (N’ TRP ) that provide measurements to model input varies.
  • N’ TRP ⁇ N TRP the remaining (N TRP -N’ TRP ) TRPs do not provide measurements to model input, i.e., measurement value is set such that the (N TRP -N’ TRP ) TRPs do not affect model output.
  • TRPs (N’ TRP ) that provide measurements can change dynamically.
  • TRP dimension of model input is equal to the number of TRPs (N’ TRP ) that provide measurements as model input.
  • N’ TRP ⁇ N TRP the remaining (N TRP -N’ TRP ) TRPs are ignored by the given model.
  • the fourth capability reporting is the approaches for reduced number of TRP as model input. If UE support fixed set of TRPs (approach 1-A, 2-A) , network will send training data of fixed set of TRPs to UE. If UE support dynamic set of TRPs (approach 1-B 2-B) , network can send training data of dynamic chosen TRPs to UE. Two bits can be used to indicate fixed and dynamic.
  • Ground truth label may be location of UE/PRU, time of arrival (TOA) of the channel path between gNB and UE/PRU, LOS or NLOS indicator of the channel path between gNB and UE/PRU.
  • TOA time of arrival
  • LOS or NLOS indicator of the channel path between gNB and UE/PRU.
  • Training data could include part of labelled data and part of un-labelled data.
  • the fifth capability reporting is for semi-supervised learning. If UE don’t support semi-supervised learning, it can only use labelled data to train the AI/ML model. Network won’t send un-labelled data to this UE. But if UE support semi-supervised learning, Network will send labelled and un-labelled data to this UE. And It can use labelled and un-labelled data to improve the model performance. One bit can be used to indicate semi-supervised learning supported or not.
  • Model fine-tuning is to fine-tune the pretrained AI/ML model with the newly collected training data, which is more flexible without strict requirement on the data for model pretraining. Model fine-tuning can improve model generalization when the well-trained AI/ML model is transferred to other scenarios.
  • the sixth capability reporting is for model fine-tuning. If UE transferred to a different scenario with current model, network could send new training data to UE and UE will fine-tuning the model. One bit can be used to indicate model fine-tuning supported or not.
  • Model monitoring is monitoring the model performance in different scenario. It can be based on provided ground truth label to calculate statistics of the difference between model output and provided ground truth label. It also can be monitoring without ground truth label. Monitoring metrics include statistics of measurement (s) compared to the statistics associated with the training data, statistics associated with the model output.
  • the seventh capability reporting is for model monitoring. If UE transferred to a different scenario with current model, it could monitor model performance. Two bits can be used to indicate model monitoring capability with/without ground truth label.
  • Combinations such as “at least one of A, B, or C, ” “one or more of A, B, or C, ” “at least one of A, B, and C, ” “one or more of A, B, and C, ” and “A, B, C, or any combination thereof” include any combination of A, B, and/or C, and may include multiples of A, multiples of B, or multiples of C.
  • combinations such as “at least one of A, B, or C, ” “one or more of A, B, or C, ” “at least one of A, B, and C, ” “one or more of A, B, and C, ” and “A, B, C, or any combination thereof” may be A only, B only, C only, A and B, A and C, B and C, or A and B and C, where any such combinations may contain one or more member or members of A, B, or C.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Artificial Intelligence (AREA)
  • Signal Processing (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Mathematical Physics (AREA)
  • General Physics & Mathematics (AREA)
  • Computing Systems (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Biomedical Technology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Medical Informatics (AREA)
  • Databases & Information Systems (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

This disclosure describes AI/ML positioning capabilities of UE to be reported. These capabilities include AI/ML positioning mode, AI/ML model input dataset type, sampling period of AI/ML model input, approaches for reduced number of TRP, semi-supervised learning, model fine-tuning and model monitoring. After get UE capability information, network will schedule corresponding AI/ML positioning flow.

Description

UE CAPABILITY REPORT FOR AI/ML POSITIONING FIELD
This present disclosure relates generally to wireless communications, and more specifically, to techniques of positioning a user equipment (UE) with Artificial Intelligence (AI) /Machine Learning (ML) .
BACKGROUND
3GPP (The 3rd Generation Partnership Project) approved a study item on AI/ML for positioning accuracy enhancement. AI/ML based positioning includes training data collection, model training, model inference and model performance monitoring. There are many different methods and parameters in AI/ML based positioning. UE will support different capabilities based on different requirements. If network can get UE capability information, it will schedule corresponding positioning flow. UE capability reporting is important topic in AI/ML positioning.
SUMMARY
The following presents a simplified summary of one or more aspects in order to provide a basic understanding of such aspects. This summary is not an extensive overview of all contemplated aspects, and is intended to neither identify key or critical elements of all aspects nor delineate the scope of any or all aspects. Its sole purpose is to present some concepts of one or more aspects in a simplified form as a prelude to the more detailed description that is presented later.
We propose several capabilities of UE reporting in this disclosure. The first capability is AI/ML positioning modes supported by the UE. AI/ML positioning includes UE-based and UE-assisted positioning, UE-side and non-UE-side model, direct AI/ML and AI/ML assisted positioning. The second capability is the dataset type of AI/ML training input data supported by the UE. The training input could be channel impulse response (CIR) , power delay profile (PDP) and delay profile (DP) of the channel between gNB and UE. The third capability is sampling period supported by the UE. Sampling period is related to bandwidth. Different sampling period will cause different performance. The fourth capability is approaches of reduced number of TRPs as model input supported by the UE. The set of TRPs used in AI/ML could be fixed or change dynamically. The fifth capability is for semi-supervised learning. Semi-supervised learning is helpful for improving the positioning accuracy when the same amount of ideal labelled data is used for semi-supervised learning, and the number of ideal labelled data is limited. The sixth  capability is for AI/ML model fine-tuning. If UE support model fine-tuning, network will send training data to UE to fine-tune the model. The seventh capability is for AI/ML model monitoring. Model monitoring include monitoring based on provided ground truth label and monitoring without ground truth label.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 illustrates indoor factory deployment scenario.
Figure 2 illustrates the capabilities transfer procedure.
Figure 3 illustrates the sampling period.
Figure 4 summaries the capabilities of UE reporting.
DETAILED DESCRIPTION
The detailed description set forth below in connection with the appended drawings is intended as a description of various configurations and is not intended to represent the only configurations in which the concepts described herein may be practiced. The detailed description includes specific details for the purpose of providing a thorough understanding of various concepts. However, it will be apparent to those skilled in the art that these concepts may be practiced without these specific details. In some instances, well known structures and components are shown in block diagram form in order to avoid obscuring such concepts.
Several aspects of telecommunication systems will now be presented with reference to various apparatus and methods. These apparatus and methods will be described in the following detailed description and illustrated in the accompanying drawings by various blocks, components, circuits, processes, algorithms, etc. (collectively referred to as “elements” ) . These elements may be implemented using electronic hardware, computer software, or any combination thereof. Whether such elements are implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system.
Figure 1 shows an example of indoor factory deployment scenario. There are 18 gNBs (N=18) in indoor factory and the location of all gNBs are fixed. UEs are random distributed in the factory. PRUs (Positioning Reference Unit) as special UEs are deployed in known location for AI/ML training data collection.
There are five basic poisoning cases in 3GPP for further study. There cases are
Case 1: UE-based positioning with UE-side model, direct AI/ML or AI/ML assisted positioning;
Case 2a: UE-assisted/LMF-based positioning with UE-side model, AI/ML assisted positioning;
Case 2b: UE-assisted/LMF-based positioning with LMF-side model, direct AI/ML positioning;
Case 3a: NG-RAN node assisted positioning with gNB-side model, AI/ML assisted positioning;
Case 3b: NG-RAN node assisted positioning with LMF-side model, direct AI/ML positioning.
Among these cases, AI/ML model can be deployed in UE-side, Non-UE-side (gNB-side, or LMF-side) . The general positioning includes UE-based and UE-assisted. UE-based means UE may perform measurement and also estimate the own position. UE-assisted means UE may perform measurement and may not estimate the own position. AI/ML positioning includes direct AI/ML and AI/ML assisted positioning. Direct AI/ML positioning means AI/ML model output UE location directly. AI/ML assisted positioning means AIML model output intermediate results. Intermediate results could be time of arrival (TOA) of the channel path between gNB and UE, LOS or NLOS indicator of the channel path between gNB and UE.
The first capability is AI/ML positioning modes supported by the UE. These modes include:
UE-based positioning with UE-side model, direct AI/ML;
UE-based positioning with UE-side model, AI/ML assisted;
UE-assisted positioning with UE-side model, AI/ML assisted;
UE-assisted positioning with Non-UE-side model.
Four bits can be used to indicate the mode supported or not. For example, “0101” means UE support “UE-based positioning with UE-side model, AI/ML assisted” and “UE-assisted positioning with Non-UE-side model” . If UE support “UE-assisted positioning with Non-UE-side model” , UE can report measurement results (CIR/PDP/DP) to network for UE-assisted positioning.
AI/ML training data include AI/ML model input data and ground truth label. gNB can send positioning reference signal to UE and UE will estimate the channel delay profiles based on received reference signal. These estimated channel delay profiles can be used as AI/ML model input. Every measurement of channel delay profile includes delay, power, and phase. CIR is composed of a list of per-sample measurements of delay, power, and phase. PDP is composed of a list of per-sample measurements of delay and power. DP is a degenerated version of PDP, where the path power is not provided. It is only composed of a list of selected delays. The overhead of DP is smallest because it only includes delay information. The AI/ML model input could be one of CIR, PDP and DP.
The second capability reporting is the supported dataset type of AI/ML model input. If one UE only support PDP as model input, it will report to network the capability. Network will send PDP-format training data to this UE. If one UE support all three formats as model input, network can send one of format training data to UE. UE also can send request to indicate which format it wants for training. Three bits can be used to indicate the format supported or not.
The sampling period is given by multiple of Tc. Tc=1/ (Δfmax·Nf) where Δfmax=480·103 Hz and Nf=4096. For example, for 15KHz subcarrier spacing and 20MHz bandwidth (Nf=  2048) , the sampling period is 64Tc (32.55ns) . The maximum estimation error is (32.55e-9) * (3e8) /2 = 4.88meters. For 30KHz subcarrier spacing and 100MHz bandwidth (Nf=4096) , the sampling period is 16Tc (8.14ns) . The maximum estimation error is (8.14e-9) * (3e8) /2 = 1.22meters. It is smaller than 32.55ns sampling period.
The third capability reporting is the supported sampling period. After UE report the supported sampling period to network, network will send training data with corresponding sampling period to UE. Supported sampling period could be a set of sampling period values, or the indication of the smallest sampling period that could be supported.
There are 4 approaches agreed in 3GPP for reduced number of TRP.
1. Approach 1: Model input size stays constant as NTRP=18. The number of TRPs (N’TRP) that provide measurements to model input varies. When N’TRP < NTRP, the remaining (NTRP -N’TRP) TRPs do not provide measurements to model input, i.e., measurement value is set such that the (NTRP -N’TRP) TRPs do not affect model output.
i. Approach 1-A. The set of TRPs (N’TRP) that provide measurements is fixed.
ii. Approach 1-B. The set of TRPs (N’TRP) that provide measurements can change dynamically.
2. Approach 2: The TRP dimension of model input is equal to the number of TRPs (N’TRP) that provide measurements as model input. When N’TRP < NTRP, the remaining (NTRP -N’TRP) TRPs are ignored by the given model.
i. Approach 2-A. The set of active TRPs (N’TRP) that provide measurements is fixed.
ii. Approach 2-B. The set of active TRPs (N’TRP) that provide measurements can change dynamically.
The fourth capability reporting is the approaches for reduced number of TRP as model input. If UE support fixed set of TRPs (approach 1-A, 2-A) , network will send training data of fixed set of TRPs to UE. If UE support dynamic set of TRPs (approach 1-B 2-B) , network can send training data of dynamic chosen TRPs to UE. Two bits can be used to indicate fixed and dynamic.
Ground truth label may be location of UE/PRU, time of arrival (TOA) of the channel path between gNB and UE/PRU, LOS or NLOS indicator of the channel path between gNB and UE/PRU. In some cases, it is difficulty to collect enough labeled data to train the AI/ML model. For example, we need to deploy more PRUs to get ground truth label, but it will cost too much. Training data could include part of labelled data and part of un-labelled data.
The fifth capability reporting is for semi-supervised learning. If UE don’t support semi-supervised learning, it can only use labelled data to train the AI/ML model. Network won’t send un-labelled data to this UE. But if UE support semi-supervised learning, Network will send labelled and un-labelled data to this UE. And It can use labelled and un-labelled data to improve  the model performance. One bit can be used to indicate semi-supervised learning supported or not.
Model fine-tuning is to fine-tune the pretrained AI/ML model with the newly collected training data, which is more flexible without strict requirement on the data for model pretraining. Model fine-tuning can improve model generalization when the well-trained AI/ML model is transferred to other scenarios.
The sixth capability reporting is for model fine-tuning. If UE transferred to a different scenario with current model, network could send new training data to UE and UE will fine-tuning the model. One bit can be used to indicate model fine-tuning supported or not.
Model monitoring is monitoring the model performance in different scenario. It can be based on provided ground truth label to calculate statistics of the difference between model output and provided ground truth label. It also can be monitoring without ground truth label. Monitoring metrics include statistics of measurement (s) compared to the statistics associated with the training data, statistics associated with the model output.
The seventh capability reporting is for model monitoring. If UE transferred to a different scenario with current model, it could monitor model performance. Two bits can be used to indicate model monitoring capability with/without ground truth label.
The previous description is provided to enable any person skilled in the art to practice the various aspects described herein. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects. Thus, the claims are not intended to be limited to the aspects shown herein, but is to be accorded the full scope consistent with the language claims, wherein reference to an element in the singular is not intended to mean “one and only one” unless specifically so stated, but rather “one or more. ” The word “exemplary” is used herein to mean “serving as an example, instance, or illustration. ” Any aspect described herein as “exemplary” is not necessarily to be construed as preferred or advantageous over other aspects. Unless specifically stated otherwise, the term “some” refers to one or more. Combinations such as “at least one of A, B, or C, ” “one or more of A, B, or C, ” “at least one of A, B, and C, ” “one or more of A, B, and C, ” and “A, B, C, or any combination thereof” include any combination of A, B, and/or C, and may include multiples of A, multiples of B, or multiples of C. Specifically, combinations such as “at least one of A, B, or C, ” “one or more of A, B, or C, ” “at least one of A, B, and C, ” “one or more of A, B, and C, ” and “A, B, C, or any combination thereof” may be A only, B only, C only, A and B, A and C, B and C, or A and B and C, where any such combinations may contain one or more member or members of A, B, or C. All structural and functional equivalents to the elements of the various aspects described throughout this disclosure that are known or later come to be known to those of ordinary skill in the art are expressly incorporated herein by reference and are intended to be encompassed by the claims.  Moreover, nothing disclosed herein is intended to be dedicated to the public regardless of whether such disclosure is explicitly recited in the claims. The words “module, ” “mechanism, ” “element, ” “UE, ” and the like may not be a substitute for the word “means. ” As such, no claim element is to be construed as a means plus function unless the element is expressly recited using the phrase “means for. ”
While aspects of the present disclosure have been described in conjunction with the specific embodiments thereof that are proposed as examples, alternatives, modifications, and variations to the examples may be made. Accordingly, embodiments as set forth herein are intended to be illustrative and not limiting. There are changes that may be made without departing from the scope of the claims set forth below.

Claims (14)

  1. Amethod of wireless communication of a user equipment (UE) , comprising:
    receiving AI/ML positioning capabilities request message from a network; and
    sending, to the network, AI/ML positioning capabilities.
  2. The method of claim 1, wherein network doesn’t send capabilities request message to UE and UE sends capabilities to network directly.
  3. The method of claim 1, wherein AI/ML positioning capabilities include AI/ML positioning mode.
  4. The method of claim 3, wherein AI/ML positioning mode include:
    UE-based positioning with UE-side model, direct AI/ML;
    UE-based positioning with UE-side model, AI/ML assisted;
    UE-assisted positioning with UE-side model, AI/ML assisted; and
    UE-assisted positioning with Non-UE-side model.
  5. The method of claim 1, wherein AI/ML positioning capabilities include AI/ML model input dataset type.
  6. The method of claim 5, wherein AI/ML positioning dataset type include CIR, PDP, DP, the format of dataset may contain sampling period, the delay index corresponding to non-zero power value, the phase corresponding to the index with non-zero power for CIR, the power corresponding to the index with non-zero power for CIR and PDP.
  7. The method of claim 1, wherein AI/ML positioning capabilities include sampling period of AI/ML model input.
  8. The method of claim 7, wherein the sampling period is given by multiple of Tc.
  9. The method of claim 1, wherein AI/ML positioning capabilities include approaches for reduced number of TRP.
  10. The method of claim 9, wherein approaches for reduced number of TRP include fixed and dynamic.
  11. The method of claim 1, wherein AI/ML positioning capabilities include semi-supervised learning.
  12. The method of claim 1, wherein AI/ML positioning capabilities include model fine-tuning.
  13. The method of claim 1, wherein AI/ML positioning capabilities include model monitoring.
  14. The method of claim 13, wherein AI/ML model monitoring include monitoring with and without ground truth labels.
PCT/CN2023/109092 2023-07-25 2023-07-25 Ue capability report for ai/ml positioning Pending WO2025020076A1 (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2023/109092 WO2025020076A1 (en) 2023-07-25 2023-07-25 Ue capability report for ai/ml positioning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2023/109092 WO2025020076A1 (en) 2023-07-25 2023-07-25 Ue capability report for ai/ml positioning

Publications (1)

Publication Number Publication Date
WO2025020076A1 true WO2025020076A1 (en) 2025-01-30

Family

ID=94373748

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/109092 Pending WO2025020076A1 (en) 2023-07-25 2023-07-25 Ue capability report for ai/ml positioning

Country Status (1)

Country Link
WO (1) WO2025020076A1 (en)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115349279A (en) * 2022-06-23 2022-11-15 北京小米移动软件有限公司 AI model determining method, device, communication equipment and storage medium
CN115835373A (en) * 2021-12-20 2023-03-21 中兴通讯股份有限公司 A positioning mode acquisition method, electronic equipment and storage medium
US20230164817A1 (en) * 2021-11-24 2023-05-25 Lenovo (Singapore) Pte. Ltd. Artificial Intelligence Capability Reporting for Wireless Communication
CN116170871A (en) * 2021-11-22 2023-05-26 维沃移动通信有限公司 Positioning method, device, terminal and network side equipment
WO2023098661A1 (en) * 2021-11-30 2023-06-08 维沃移动通信有限公司 Positioning method and communication device

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116170871A (en) * 2021-11-22 2023-05-26 维沃移动通信有限公司 Positioning method, device, terminal and network side equipment
US20230164817A1 (en) * 2021-11-24 2023-05-25 Lenovo (Singapore) Pte. Ltd. Artificial Intelligence Capability Reporting for Wireless Communication
WO2023098661A1 (en) * 2021-11-30 2023-06-08 维沃移动通信有限公司 Positioning method and communication device
CN115835373A (en) * 2021-12-20 2023-03-21 中兴通讯股份有限公司 A positioning mode acquisition method, electronic equipment and storage medium
CN115349279A (en) * 2022-06-23 2022-11-15 北京小米移动软件有限公司 AI model determining method, device, communication equipment and storage medium

Similar Documents

Publication Publication Date Title
US8018972B2 (en) Timing over packet performance
US9634782B2 (en) Clock synchronization system, clock synchronization method, and storage medium whereupon clock synchronization program is stored
EP2023456A1 (en) Estimating a time offset between stationary clocks
US20160043823A1 (en) Method, apparatus, and system for generating timestamp
US20170195982A1 (en) Method for time synchronization of domain based on time information of vehicle
US11411666B2 (en) Clock fault detection and correction between synchronized network devices
CN108599885B (en) High-precision time synchronization method
CN113141648B (en) Timing information reporting method, terminal and network side equipment
WO2025020076A1 (en) Ue capability report for ai/ml positioning
CN115702590A (en) Position estimation
US12439370B2 (en) Automated position estimation based on coordinating access points
WO2024148576A1 (en) Scenario monitor mechanism for ai/ml positioning
US20240137246A1 (en) Network node for a vehicle
WO2024145925A1 (en) Model performance monitor mechanism for ai/ml positioning
WO2024152311A1 (en) Serving cell-based data collection and model monitoring mechanism for ai/ml positioning
Guchhait et al. Joint minimum variance unbiased and maximum likelihood estimation of clock offset and skew in one-way packet transmission
WO2018188080A1 (en) Method and device for measuring time difference
WO2024197735A1 (en) Model performance monitor mechanism for ai/ml assisted positioning
WO2024164116A1 (en) Auxiliary positioning reference units mechanism for ai/ml positioning
US12113614B2 (en) AIML positioning receiver for flexible carrier aggregation
WO2024065731A1 (en) Mechanism for integer cycle report based on carrier phase in ofdm systems
US20250151015A1 (en) Methods, apparatuses, and computer program products for valid reporting of timing measurements for new radio carrier phase positioning
WO2024168807A1 (en) A general training data collection procedure for ai positioning
CN112383379B (en) Method and device for calculating node time error of wireless information system
US20250056477A1 (en) Model monitoring for positioning

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23946169

Country of ref document: EP

Kind code of ref document: A1