18351220. METHOD AND APPARATUS FOR CSI FEEDBACK PERFORMED BY ONLINE LEARNING-BASED UE-DRIVEN AUTOENCODER simplified abstract (SAMSUNG ELECTRONICS CO., LTD.)

From WikiPatents
Jump to navigation Jump to search

METHOD AND APPARATUS FOR CSI FEEDBACK PERFORMED BY ONLINE LEARNING-BASED UE-DRIVEN AUTOENCODER

Organization Name

SAMSUNG ELECTRONICS CO., LTD.

Inventor(s)

Wonjun Kim of Suwon-si (KR)

Suhwook Kim of Suwon-si (KR)

Seunghyun Lee of Suwon-si (KR)

Hyeondeok Jang of Suwon-si (KR)

METHOD AND APPARATUS FOR CSI FEEDBACK PERFORMED BY ONLINE LEARNING-BASED UE-DRIVEN AUTOENCODER - A simplified explanation of the abstract

This abstract first appeared for US patent application 18351220 titled 'METHOD AND APPARATUS FOR CSI FEEDBACK PERFORMED BY ONLINE LEARNING-BASED UE-DRIVEN AUTOENCODER

Simplified Explanation

The disclosed patent application is about a communication system (5G or 6G) that supports higher data transfer rates compared to a 4G system like LTE. The method involves a terminal transmitting training capability information to a base station, which relates to the terminal's ability to train an artificial intelligence (AI) model called an autoencoder. This autoencoder is designed to compress and reconstruct feedback information for a channel state information-reference signal (CSI-RS). The terminal receives information from the base station regarding the completion time point of the training and decoder information determined based on the training capability. The terminal then generates a training dataset for the autoencoder based on at least one received CSI-RS, trains the autoencoder using the decoder information, training completion time point, and the generated training dataset, and finally transmits the training result information of the autoencoder back to the base station.

  • The patent application describes a method for a terminal in a wireless communication system to train an AI model called an autoencoder.
  • The autoencoder is specifically designed to compress and reconstruct feedback information for a channel state information-reference signal (CSI-RS).
  • The terminal transmits training capability information to the base station, which indicates the terminal's ability to train the autoencoder.
  • The base station provides information on the training completion time point and decoder information based on the training capability.
  • The terminal generates a training dataset for the autoencoder based on received CSI-RS.
  • The autoencoder is trained using the decoder information, training completion time point, and the generated training dataset.
  • The terminal transmits the training result information of the autoencoder back to the base station.

Potential Applications:

  • This technology can be applied in 5G or 6G communication systems to improve data transfer rates and overall system performance.
  • It can be used in wireless communication networks to enhance the efficiency of channel state information-reference signal (CSI-RS) feedback.

Problems Solved:

  • The technology addresses the need for higher data transfer rates in communication systems beyond 4G.
  • It solves the problem of efficiently compressing and reconstructing feedback information for a channel state information-reference signal (CSI-RS) using an AI model.

Benefits:

  • The use of an autoencoder and AI model training allows for more efficient compression and reconstruction of feedback information, leading to improved system performance.
  • The technology enables higher data transfer rates, enhancing the overall user experience in wireless communication systems.
  • It provides a method for training the autoencoder based on training capability information, allowing for customization and optimization based on the terminal's capabilities.


Original Abstract Submitted

The disclosure relates to a 5G or 6G communication system for supporting a higher data transfer rate than a 4G communication system such as LTE. According to an embodiment, a method performed by a terminal in a wireless communication system may include transmitting, to a base station, training capability information of the terminal relating to artificial intelligence (AI) model training of an autoencoder configured to compress and reconstruct feedback information for a channel state information-reference signal (CSI-RS), receiving, from the base station, information on a training completion time point and decoder information determined based on the training capability information, generating a training dataset for the autoencoder, based on at least one received CSI-RS, training the autoencoder, based on the decoder information, the information on the training completion time point, and the generated training dataset, and transmitting training result information of the autoencoder to the base station.

Retrieved from "http://wikipatents.org/index.php?title=18351220._METHOD_AND_APPARATUS_FOR_CSI_FEEDBACK_PERFORMED_BY_ONLINE_LEARNING-BASED_UE-DRIVEN_AUTOENCODER_simplified_abstract_(SAMSUNG_ELECTRONICS_CO.,_LTD.)&oldid=30164"