AI-Based CSI Compression for Low-Bit Uplink Feedback
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The omission of part or all information in Channel State Information (CSI) leads to a loss of channel information and affects communication performance when a large number of bits are carried in CSI.
Innovation Solution
The use of Artificial Intelligence (AI)/Machine Learning (ML)-based models to compress CSI by adjusting a first model, reducing the number of bits while ensuring the integrity of channel information is maintained, allowing complete recovery at the network device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If CSI omission is performed to reduce the number of bits, then the uplink channel transmission is enabled, but channel information is lost and communication performance deteriorates
Solution Approach 1:
The patent changes the parameter of model precision (from high precision to low precision) to reduce the number of CSI bits. By adjusting the precision parameter of the AI/ML model, the system achieves compression while maintaining acceptable channel information integrity for scheduling purposes.
Solution Approach 2:
The patent creates a compressed copy of the CSI through the AI/ML model that captures the essential channel characteristics. This compressed representation serves as a functional copy that enables scheduling decisions without requiring the full original CSI data.
2Quantity of substance
If AI/ML-based model is used to compress CSI, then the number of bits is reduced, but the complexity of the system increases
Solution Approach 1:
The terminal device autonomously adjusts the AI/ML model parameters based on local channel conditions and scheduling requirements. This self-service approach eliminates the need for complex centralized coordination, reducing overall system complexity while achieving effective CSI compression.
Solution Approach 2:
The system changes model parameters dynamically based on transmission needs. By adjusting precision parameters rather than using fixed complex models, the system achieves compression with manageable complexity that adapts to different communication scenarios.
Data Source
AI summary
An information compression method and apparatus, and a terminal device and a network device are provided. The method includes that a terminal device determines, according to a priority of first channel state information (CSI) among at least one piece of CSI carried by an uplink channel, that the first CSI needs to be compressed. The first CSI is AI/MI-based CSI, and the uplink channel is used to carry the first CSI. The terminal device compresses the first CSI by adjusting the first model. The first model is used to generate the first CSI.