Search

US-20260128969-A1 - OPERATOR CONTROLLED ARTIFICIAL INTELLIGENCE (AI)/MACHINE LANGUAGE (ML) WTRU DATA COLLECTION EXPOSURE FRAMEWORK

US20260128969A1US 20260128969 A1US20260128969 A1US 20260128969A1US-20260128969-A1

Abstract

A wireless transmit/receive unit (WTRU) may comprise a processor configured to determine that a data collection configuration is not available from a direct data collection client (DDCC) of the WTRU. The processor may send a request associated with the data collection configuration to an application server, wherein the request comprises an application identifier, an indication of a user profile, and a protocol data unit (PDU) session identifier. The processor may receive the data collection configuration, wherein the data collection configuration comprises data collection policies that indicate conditions upon which the WTRU is to collect the data, and wherein the data collection policies indicate whether and where the WTRU is to send the collected data.

Inventors

  • Ulises Olvera-Hernandez
  • Michael Starsinic
  • Oumer Teyeb
  • Taimoor ABBAS
  • Samir Ferdi

Assignees

  • INTERDIGITAL PATENT HOLDINGS, INC.

Dates

Publication Date
20260507
Application Date
20241107

Claims (20)

  1. 1 . A wireless transmit/receive unit (WTRU) comprising: a processor configured to: determine that a data collection configuration is not available from a direct data collection client (DDCC) of the WTRU; send a request associated with the data collection configuration to an application server, wherein the request comprises an application identifier, an indication of a user profile, and a protocol data unit (PDU) session identifier; and receive the data collection configuration, wherein the data collection configuration comprises data collection policies that indicate conditions upon which the WTRU is to collect the data, and wherein the data collection policies indicate whether and where the WTRU is to send the collected data.
  2. 2 . The WTRU of claim 1 , wherein the conditions indicated by the data collection policies comprise a time limit, a size limit, location of the WTRU, a speed of the WTRU, a battery level of the WTRU, or a data activity of the WTRU.
  3. 3 . The WTRU of claim 1 , wherein the data collection policies indicate whether and where the WTRU is to send the collected data based on (i) location of the WTRU, (ii) a data type associated with the collected data, (iii) network conditions, or (iv) time.
  4. 4 . The WTRU of claim 1 , wherein the processor is configured to: send the collected data in accordance with the data collection configuration.
  5. 5 . The WTRU of claim 1 , wherein the processor is configured to: send a request to the DDCC for the data collection configuration, wherein the request comprises the application identifier and the indication of the user profile; receive a response from the DDCC that indicates that the data collection configuration does not exist at the DDCC; and determine that the data collection configuration is not available from the DDCC based on the response from the DDCC.
  6. 6 . The WTRU of claim 1 , wherein the processor is configured to determine the PDU session identifier via the use of AT-Commands.
  7. 7 . The WTRU of claim 1 , wherein the processor is configured to receive the PDU session identifier from the DDCC.
  8. 8 . The WTRU of claim 1 , wherein the processor is configured to determine the PDU session identifier via an instantiation of a DCAL layer.
  9. 9 . The WTRU of claim 1 , wherein the processor is configured to receive the data collection configuration at an application layer via the DDCC.
  10. 10 . The WTRU of claim 1 , wherein the processor is configured to: validate the data collection configuration based on a comparison between the data collection policies of the data collection configuration and a configuration received from a radio access network (RAN).
  11. 11 . A method performed by a wireless transmit/receive unit (WTRU), the method comprising: determining that a data collection configuration is not available from a direct data collection client (DDCC) of the WTRU; sending a request associated with the data collection configuration to an application server, wherein the request comprises an application identifier, an indication of a user profile, and a protocol data unit (PDU) session identifier; and receiving the data collection configuration, wherein the data collection configuration comprises data collection policies that indicate conditions upon which the WTRU is to collect the data, and wherein the data collection policies indicate whether and where the WTRU is to send the collected data.
  12. 12 . The WTRU of claim 11 , wherein the conditions indicated by the data collection policies comprise a time limit, a size limit, location of the WTRU, a speed of the WTRU, a battery level of the WTRU, or a data activity of the WTRU.
  13. 13 . The WTRU of claim 11 , wherein the data collection policies indicate whether and where the WTRU is to send the collected data based on (i) location of the WTRU, (ii) a data type associated with the collected data, (iii) network conditions, or (iv) time.
  14. 14 . The WTRU of claim 11 , wherein the method further comprises: sending the collected data in accordance with the data collection configuration.
  15. 15 . The WTRU of claim 11 , wherein the method further comprises: sending a request to the DDCC for the data collection configuration, wherein the request comprises the application identifier and the indication of the user profile; receiving a response from the DDCC that indicates that the data collection configuration does not exist at the DDCC; and determining that the data collection configuration is not available from the DDCC based on the response from the DDCC.
  16. 16 . The WTRU of claim 11 , wherein the method further comprises determining the PDU session identifier via the use of AT-Commands.
  17. 17 . The WTRU of claim 11 , wherein the method further comprises receiving the PDU session identifier from the DDCC.
  18. 18 . The WTRU of claim 11 , wherein the method further comprises determining the PDU session identifier via an instantiation of a DCAL layer.
  19. 19 . The WTRU of claim 11 , wherein the method further comprises receiving the data collection configuration at an application layer via the DDCC.
  20. 20 . The WTRU of claim 11 , wherein the method further comprises: validating the data collection configuration based on a comparison between the data collection policies of the data collection configuration and a configuration received from a radio access network (RAN).

Description

BACKGROUND Mechanisms and frameworks have been specified for using AI/ML based approaches for both the air interface level (e.g., CSI-feedback enhancements (e.g., CSI compression), beam management/prediction and WTRU positioning) and/or network level (e.g., network energy saving, load balancing and mobility). AI/ML operations are based on models (e.g., neural networks) that have been trained using a substantial amount of data under different scenarios/conditions. The conditions may be WTRU-side conditions (e.g., speed) or network-side conditions (e.g., antenna pattern, load, etc.,). These conditions are referred to as WTRU-side additional conditions and network-side additional conditions, respectively. For a given AI/ML functionality, there could be several models (e.g., each model trained and/or suitable for different network and/or WTRU side additional conditions). Once a model is well trained, it may be deployed in a test environment (e.g., test network) for performance testing. Model monitoring may need to be performed even after deployment in a real network, as the current network and/or WTRU conditions may become different from the scenarios and/or conditions in which the model was trained and/or tested. If the model monitoring is shown to provide undesirable WTRU and/or network performance, a decision may be made to switch to another model and/or stop using AI/ML based operations for the concerned function, etc., The performance monitoring can also be used to determine whether a model needs to be retained with new sets of data. Model training and/or monitoring may be performed at the WTRU, at the network, and/or in collaboration between the two. Model training and/or monitoring may be performed offline or online. A given AI/ML model may be trained under certain WTRU and/or network conditions. For example, a WTRU condition may be the speed of the WTRU. On the other hand, network conditions may be something that may be related to some network configurations and/or settings that the WTRU may not be aware of but may impact the performance of the model. For example, an AI/ML beam management model may perform differently if it is trained when the network was using a certain antenna pattern, beam pattern, power levels, and so on. Also, there may be aspects related to network load, that may have impact on the model performance. Since the WTRU doesn't necessarily need to know all these, and network may also not want to expose some of these implementations, the network may hide these details by signaling to the WTRU (e.g., via broadcast signaling, dedicated signaling, etc.,) a network configuration index and/or associated ID. For example, when data is being collected for training a model, tagging may be performed indicating under which network conditions the model is being trained. When a WTRU is being configured to perform an AI/ML operation, it may be configured to check the consistency between the conditions under which the AI/ML model is trained on and current conditions (e.g., current WTRU conditions, current associated ID signaled by the network indicating current network conditions/settings, etc.,). The topic of data collection has been studied. Data collection is a function that provides input data to the model training, management, and inference functions. Data collection is performed for the purpose of AI/ML model training. Model training is a function that performs AI/ML model training, validation, and testing which may generate model performance metrics that may be used as part of the model testing procedure. The model training function may also be responsible for data preparation (e.g., data pre-processing and cleaning, formatting, and transformation) based on training data delivered by a data collection function, if required. If a model storage function is used, the model training function may be used to deliver trained, validated and tested AI/ML models to the model storage function. According to studies, for all types of offline model training (i.e., WTRU-side, network (NW)-side, and/or two-sided mode training), there is no latency requirement for data collection. When designing system enhancements to support data collection procedures, consideration should be given to what type of data may be reported and what type of data needs to be reported. Examples of the type data the needs to be reported may be data related to model input, data related to ground truth, data that can provide assistance when categorizing the data, and information about the quality of the data (e.g. relevance of the data to the training, uniformity of the data contents, etc.,). When designing system enhancements to support data collection procedures, consideration should be given to how the WTRU may be configured to perform the reporting of the collected data. When designing system enhancements to support data collection procedures, consideration should be given to whether the WTRU is able to report information that is pro