Search

EP-4740336-A1 - METHOD AND PROCEDURE FOR DATA COLLECTION FOR AI BASED CSI PREDICTION

EP4740336A1EP 4740336 A1EP4740336 A1EP 4740336A1EP-4740336-A1

Abstract

Apparatuses, systems, and methods for Al based CSI feedback with CSI prediction, including systems, methods, and mechanisms for a user equipment device (UE) to encode a data collection request, for training a CSI prediction Al-model, for transmission of the data collection request to a next generation node B (gNB). The UE may decode a data collection request response received from the gNB. The UE may perform UE measurements of CSI Reference Signals (CSI-RS) for training the CSI prediction Al-model based on the data collection request response from the gNB. The UE may encode a data collection stop request for transmission to the gNB. The UE may store the UE measurements of the CSI-RS in a memory coupled to the one or more processors.

Inventors

  • YANG, WEIDONG
  • NIU, HUANING
  • YE, SIGEN
  • SUN, HAITONG
  • ZENG, WEI
  • ZHANG, DAWEI

Assignees

  • Apple Inc.

Dates

Publication Date
20260513
Application Date
20240802

Claims (20)

  1. 1. An apparatus of a user equipment (UE) configured to perform data collection for channel state information (CSI) prediction with an artificial intelligence (Al) model, the apparatus comprising: one or more processors configured to: encode a data collection request, for training a CSI prediction AI- model, for transmission of the data collection request to a next generation node B (gNB); decode a data collection request response received from the gNB; and perform UE measurements of CSI Reference Signals (CSI-RS) for training the CSI prediction Al-model based on the data collection request response from the gNB; and a memory coupled to the one or more processors and configured to store the UE measurements of the CSI-RS.
  2. 2. The apparatus of claim 1 , wherein the one or more processors are further configured to: encode a data collection stop request for transmission to the gNB ; or decode a data collection stop request response received from the gNB.
  3. 3. The apparatus of claim 1, wherein the one or more processors are further configured to encode the UE measurements for transmission to the gNB to enable the gNB to train the CSI prediction Al-model.
  4. 4. The apparatus of claim 1, wherein the one or more processors are further configured to buffer the UE measurements for training the CSI prediction AI- model.
  5. 5. The apparatus of claim 1, further comprising a transceiver configured to: receive the CSI-RS from the gNB for training the CSI prediction Al- model; transmit the data collection request to the UE; and transmit the UE measurements performed by the UE to the gNB.
  6. 6. The apparatus of claim 1, wherein the data collection request includes at least a request to collect input data for the CSI prediction Al-model and output data from the CSI prediction Al-model to enable the gNB to train the CSI prediction Al- model.
  7. 7. The apparatus of any of claims 1 to 6, wherein the input data for the CSI prediction Al-model comprises one or more of: a total number of samples used as input for the CSI prediction Al-model collected in a time domain; measurement distances between one or more of the total number of samples collected in the time domain; and information used for categorizing the total number of samples.
  8. 8. The apparatus of any of claims 1 to 7, wherein output data of the CSI prediction Al-model includes a total number of samples to predict and a prediction distance between one or more of the samples.
  9. 9. The apparatus of claim 1, wherein the one or more processors are further configured to encode the data collection request for transmission from the UE to the gNB via one or more of: an uplink radio resource control (RRC) message; or an uplink Medium Access Control (MAC) control element (CE).
  10. 10. The apparatus of claim 9, wherein the one or more processors are further configured to determine when to trigger the data collection request based on one or more of: the UE has not been trained using the CSI prediction Al-model; the UE has moved from outdoors to indoors; the UE has moved from indoors to outdoors; the UE has moved to an environment with a number of obstructions that have increased or decreased by a predetermined threshold level; or the UE has changed speed by a predetermined amount.
  11. 11. The apparatus of claim 1, wherein the one or more processors are further configured to encode a request for transmission to the gNB for assisted information to categorize the UE measurements based on the assisted information.
  12. 12. The apparatus of claim 11, wherein the assisted information comprises one or more of: a model type of a UE side CSI prediction Al-model; an antenna type of the gNB ; or a use of antenna virtualization by the gNB.
  13. 13. The apparatus of claim 1, wherein the one or more processors are configured to use a UE side CSI prediction Al-model comprising one or more of: a long short-term memory (LSTM) deep recurrent neural network model; a two-dimensional convolutional neural network model with an input tensor and an output tensor including dimensions of time and frequency; or a three-dimensional convolutional neural network model with an input tensor and an output tensor including dimensions of time and frequency and antenna type.
  14. 14. The apparatus of claim 1, wherein the CSI prediction Al-model is a UE side prediction model.
  15. 15. The apparatus of claim 1, wherein the one or more processors are configured to encode the UE measurements of the CSI-RS used for training the CSI prediction Al-model, to enable the UE measurements to be communicated to an external server for offline training of the CSI prediction Al-model.
  16. 16. The apparatus of claim 1, wherein the one or more processors are configured to receive, from the gNB, time domain repetition CSI-RS port transmissions to enable increased CSI-RS measurement accuracy.
  17. 17. The apparatus of claim 1, wherein the one or more processors are configured to measure CSI-RS from multiple OFDM symbols received from the gNB to enable a greater number of CSI-RS measurements to provide to the CSI prediction AI- model to enable higher accuracy training of the CSI prediction Al-model.
  18. 18. The apparatus of claim 1, wherein the data collection request response received from the gNB is a command for the UE to perform the UE measurements of the CSI-RS to collect input samples for training the CSI prediction Al-model.
  19. 19. An apparatus of a user equipment (UE) configured to perform data collection for channel state information (CSI) prediction with an Artificial Intelligence (Al) model, the apparatus comprising: one or more processors configured to: encode a data collection request, for training a CSI prediction AI- model, for transmission of the data collection request to a next generation node B (gNB); decode a data collection request response received from the gNB, wherein the data collection request response received from the gNB includes a command to perform UE measurements of CSI Reference Signals (CSI- RS) for training the CSI prediction Al-model; perform the UE measurements of CSI-RS for training the CSI prediction Al-model based on the data collection request response from the gNB; buffer the UE measurements for training the CSI prediction AI- model; encode a data collection stop request for transmission to the gNB; and decode a data collection stop request response received from the gNB; and a memory coupled to the one or more processors and configured to store the UE measurements of the CSI-RS.
  20. 20. The apparatus of claim 19, wherein the one or more processors are further configured to encode the UE measurements for transmission to the gNB to enable a network side training of the CSI prediction Al-model.

Description

METHOD AND PROCEDURE FOR DATA COLLECTION FOR Al BASED CSI PREDICTION FIELD [0001] The invention relates to wireless communications, and more particularly to apparatuses, systems, and methods for data collection for artificial intelligence (Al) based channel state information (CSI) prediction, including systems, methods, and mechanisms for a UE to initiate data collection to train an Al prediction model during 5G NR communications. DESCRIPTION OF THE RELATED ART [0002] Wireless communication systems are rapidly growing in usage. In recent years, wireless devices such as smart phones and tablet computers have become increasingly sophisticated. In addition to supporting telephone calls, many mobile devices now provide access to the internet, email, text messaging, and navigation using the global positioning system (GPS), and are capable of operating sophisticated applications that utilize these functionalities. Additionally, there exist numerous different wireless communication technologies and standards. [0003] Long Term Evolution (LTE), also referred to as the Evolved Universal Terrestrial Radio Access Network (E-UTRAN, has been the technology of choice for the majority of wireless network operators worldwide, providing mobile broadband data and high-speed Internet access to their subscriber base. LTE was first proposed in 2004 and was first standardized in 2008. Since then, as usage of wireless communication systems has expanded exponentially, demand has risen for wireless network operators to support a higher capacity for a higher density of mobile broadband users. Thus, in 2015 study of a new radio access technology began and, in 2017, a first release of the Third Generation Partnership Project (3GPP) Fifth Generation New Radio (5G NR) was standardized. 5th generation mobile networks or 5th generation wireless systems, referred to as 3GPP NR (otherwise known as 5G-NR or NR-5G for 5G New Radio, also simply referred to as NR). NR proposes a higher capacity for a higher density of mobile broadband users, also supporting device-to-device, ultra-reliable, and massive machine communications, as well as lower latency and lower battery consumption, than LTE standards. [0004] 5G-NR provides, as compared to LTE, a higher capacity for a higher density of mobile broadband users, while also supporting device-to-device, ultra-reliable, and massive machine type communications with lower latency and/or lower battery consumption. Further, NR may allow for more flexible UE scheduling as compared to current LTE. Consequently, efforts are being made in ongoing developments of 5G-NR to take advantage of higher throughputs possible at higher frequencies. [0005] One aspect of wireless communication systems, e.g., systems for NR cellular wireless communications, is the transmission and measurement of reference signals, including channel- state information reference signals (CSI-RS). SUMMARY [0006] Embodiments relate to wireless communications, and more particularly to apparatuses, systems, and methods for data collection for Al based CSI prediction, including systems, methods, and mechanisms for a UE to initiate and collect data to train an Al prediction model during 5G NR communications. [0007] For example, in some embodiments, a UE may transmit, to a base station, a data collection request, for training a CSI prediction Artificial Intelligence (Al) model, for transmission of the data collection request to a base station (e.g., a next generation node B (gNB)). The UE may receive a data collection request response received from the gNB. The UE may perform UE measurements of CSI Reference Signals (CSI-RS) for training the CSI prediction Al-model based on the data collection request response from the gNB. [0008] The UE may transmit, to the gNB, a data collection stop request. The UE may store the UE measurements of the CSI-RS in a memory coupled to one or more processors. The configuration for the data collection request may include at least a request to collect input data for the CSI prediction Al-model and output data from the CSI prediction AI- model to enable the gNB to train the CSI prediction Al-model. The configuration for the input data for the CSI prediction Al-model include a total number of samples used as input for the CSI prediction Al-model collected in a time domain, a measurement distances between one or more of the total number of samples collected in the time domain, and information used for categorizing the total number of samples. The configuration for the output data of the CSI prediction Al-model may include a total number of samples to predict and a prediction distance between one or more of the samples. [0009] The techniques described herein may be implemented in and/or used with a number of different types of devices, including but not limited to base stations, access points, cellular phones, tablet computers, wearable computing devices, portable media players, vehicles, and any of various other comput