CSI Autoencoder Updating Using Pre-Compression Feedback Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication systems face issues with faulty datasets and excessive errors in channel state information (CSI) data compression, particularly in the context of AI models used in autoencoders, which affect the performance of CSI data compression.
Innovation Solution
A method and apparatus are provided to ensure accurate CSI data compression by instructing user equipment (UE) to collect and transmit data before compression, allowing a base station (BS) to update the AI model in the autoencoder, using more bits for compressed data or codebook-based CSI feedback to reduce errors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If the AI model is fine-tuned using faulty datasets, then the model updating process can be completed, but the CSI data compression performance deteriorates with excessive errors
Solution Approach 1:
The system performs preliminary actions by instructing the UE to collect and transmit data before compression (raw CSI data) in advance, so that the BS can use this high-quality data for fine-tuning the AI model without relying on faulty compressed datasets, thereby resolving the contradiction between updating speed and compression performance
Solution Approach 2:
The system implements feedback by using the BS to decode the pre-compression data and update the AI model based on the decoded information, creating a closed-loop system where the model is continuously refined using accurate feedback from the actual channel conditions, thus improving compression performance while maintaining efficient model updates
2Productivity
If compressed data is used for model fine-tuning, then data transmission efficiency is maintained, but the dataset quality becomes faulty leading to excessive errors
Solution Approach 1:
The system extracts the essential training data (pre-compression CSI data) from the UE before compression occurs, separating the high-quality raw data needed for model fine-tuning from the compressed data used for transmission, thereby allowing the model to be trained on reliable data while maintaining transmission efficiency
Solution Approach 2:
The pre-compression data acts as an intermediary between the UE and BS, serving as a high-quality training dataset that mediates the model fine-tuning process, allowing the system to maintain both transmission efficiency (by using compression for actual data transfer) and dataset quality (by using the intermediary pre-compression data for training)
Data Source
AI summary
Disclosed is a method by which a base station (BS) updates an autoencoder (AE) for channel state information (CSI) feedback in a wireless communication system, including transmitting, to a user equipment (UE), a data collect instruction message instructing to collect data before compression corresponding to compressed data, receiving data before the compression from the UE, and updating an AE, based on the received data before compression.