Online Learning Autoencoder for CSI Feedback Adaptation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 6G communication systems face challenges in efficiently transmitting feedback information for channel state information-reference signals (CSI-RS), particularly in high-speed scenarios and dynamic channel environments, where pre-trained AI models are ineffective due to environmental changes.
Innovation Solution
A method where a UE and a base station perform real-time encoding and decoding of CSI-RS feedback information using an autoencoder, with the UE generating a training dataset and transmitting training result information, allowing for online learning and adaptation to the current channel environment without relying on pre-trained AI models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pre-trained AI models are used for CSI feedback compression, then the initial transmission efficiency is improved, but the system performance deteriorates in high-speed scenarios and dynamic channel environments
Solution Approach 1:
The patent implements dynamic adaptation by enabling the UE to retrain the autoencoder model online based on current channel conditions. The system transitions from a static pre-trained model to a dynamic model that continuously adapts to changing environments through online learning, resolving the contradiction between initial efficiency and environmental adaptability.
Solution Approach 2:
The UE performs self-training of the autoencoder model using locally available CSI-RS data and decoder information from the base station. This self-service mechanism allows the system to adapt to changing channel conditions without requiring external retraining, maintaining both transmission efficiency and adaptability.
2Adaptability or versatility
If online learning-based autoencoder training is performed by the UE, then the adaptability to changing environments is improved, but the device complexity and computation burden increase
Solution Approach 1:
The base station performs preliminary actions by determining the decoder information and training completion time point in advance, then providing these to the UE. This preliminary preparation reduces the computational burden on the UE during online learning, as the UE only needs to train the encoder portion with pre-configured decoder parameters.
Solution Approach 2:
The base station acts as an intermediary by providing decoder information and training guidance to the UE. This intermediary role allows the system to distribute computational complexity, with the base station handling decoder configuration and the UE focusing on encoder training, thereby reducing overall device complexity.
3Measurement precision
If the UE generates training dataset from received CSI-RS, then the model accuracy is improved, but the training time and latency increase
Solution Approach 1:
The system implements partial action by having the UE generate training datasets from a subset of received CSI-RS signals rather than requiring complete datasets. The base station determines a training completion time point that balances accuracy requirements with time constraints, allowing training to proceed with sufficient but not excessive data, thereby reducing training time while maintaining acceptable accuracy.
Data Source
AI summary
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.