CSI Feedback Overhead Reduction via Eigenvalue Decomposition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In MIMO wireless communication systems, the high CSI feedback overhead restricts performance improvement, particularly for multiple sub-bands, as existing CSI quantization feedback methods either have low precision or excessive overhead.
Innovation Solution
The method involves decomposing the CSI matrix into orthogonal matrices Ud and Vd, feeding back amplitude and phase information of their eigenvectors, which allows for precise CSI feedback with reduced overhead by utilizing singular value decomposition (SVD) and determining the number of eigenvalues based on threshold values.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If traditional DFT-based codebooks are used for CSI quantization feedback, then the feedback overhead is relatively small, but the CSI quantization precision is low and performance is limited
Solution Approach 1:
The patent transforms the CSI feedback from traditional DFT-based parameter representation to eigenvalue decomposition parameters (eigenvalues and eigenvectors). By changing the mathematical representation parameters, the system achieves higher CSI quantization precision while controlling feedback overhead through selective feedback of only the largest eigenvalues and corresponding eigenvectors.
Solution Approach 2:
The patent segments the full CSI matrix into dominant eigenvalues and eigenvectors by performing eigenvalue decomposition. Instead of feedback the entire matrix, only the significant components (largest eigenvalues and their corresponding eigenvectors) are selected and fed back, achieving dimensionality reduction while preserving essential channel information.
2Measurement precision
If high-rank DFT vectors or many DFT vectors are linearly weighted and combined to improve CSI quantization precision, then the CSI feedback overhead becomes relatively large
Solution Approach 1:
The patent fundamentally changes the parameter representation from linear combinations of DFT vectors to eigenvalue decomposition results. This parameter transformation allows the system to represent the same channel information more efficiently, achieving high precision with fewer feedback parameters by exploiting the mathematical properties of eigenvalue decomposition.
Solution Approach 2:
The patent extracts only the essential components from the full CSI matrix - specifically the largest eigenvalues and their corresponding eigenvectors. By taking out only the dominant components that carry the most significant channel information, the system achieves high precision feedback with reduced overhead, discarding less significant components.
Data Source
AI summary
Provided are an information feedback method, terminal, base station, a storage medium, and an electronic device. The method includes: decomposing a channel state information (CSI) matrix H to obtain a matrix Ud and a matrix Vd, where Ud is a matrix having d columns, and every two column vectors are mutually orthogonal; and Vd is a matrix having d columns, and every two column vectors are mutually orthogonal; and feeding back amplitude and phase information of elements in Ud including d left eigenvectors and/or amplitude and phase information of elements in Vd including d right eigenvectors.