Pre-coding Matrix Construction Using Shared Orthogonal Vector Groups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The current Type II codebook in wireless communication systems supports only rank1 and rank2 codebooks, leading to increased feedback overhead and inter-data layer interference for higher rank values, as each layer's linear combination coefficients are independently determined and fed back, lacking guaranteed orthogonality between layers.
Innovation Solution
A method and device for feeding back channel state information to construct a pre-coding matrix using orthogonal vector groups, where linear combination coefficients are determined based on vectors in these groups, reducing feedback overhead by sharing coefficients across layers and ensuring orthogonality through phase adjustment factors, supporting higher rank codebooks while minimizing inter-layer interference.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If linear combination coefficients are independently determined for each layer, then channel quantization accuracy is improved, but feedback overhead increases
Solution Approach 1:
The patent merges the determination of linear combination coefficients across multiple layers by introducing a shared basis vector group. Instead of independently determining coefficients for each layer, the invention combines all layer coefficients into a unified linear combination model using common basis vectors, thereby reducing the total number of feedback parameters while maintaining channel representation accuracy.
Solution Approach 2:
The basis vector group serves as a universal component that functions across all data layers. The same set of basis vectors is used to represent channels for multiple layers simultaneously, making the coefficient determination process multi-functional and eliminating the need for separate independent coefficient sets for each layer.
2Measurement precision
If linear combination coefficients are independently determined for each layer, then channel representation accuracy is improved, but inter-layer interference increases
Solution Approach 1:
By merging the coefficient determination process across layers through shared basis vectors, the invention creates a coordinated representation system. This unified approach ensures that linear combination coefficients for different layers are derived from the same basis set, which inherently reduces inter-layer interference while preserving accurate channel representation for each layer.
3Measurement precision
If more orthogonal beam groups are used, then channel quantization accuracy is improved, but feedback overhead increases
Solution Approach 1:
The patent merges the functionality of multiple orthogonal beam groups into a single shared basis vector group that serves all data layers. This consolidation reduces the total number of beam groups that need to be independently configured and fed back, thereby reducing feedback overhead while maintaining the ability to accurately represent channels across multiple layers through the unified basis set.
Data Source
AI summary
Disclosed are a method and device for feeding back channel state information, and a method and device for determining a pre-coding matrix. In the present application, a terminal determines N orthogonal vector groups, one of the orthogonal vector groups being used for constructing pre-coding corresponding to M layers in a pre-coding matrix; the terminal determines, according to beams in the N orthogonal vector groups, linear combination coefficients for constructing the pre-coding matrix, the linear combination coefficients includes linear combination coefficients corresponding to the layers in the pre-coding matrix; the terminal feeds back channel state information to a base station, the channel state information includes indication information of the N orthogonal vector groups and linear combination coefficients, in the determined linear combination coefficients of the pre-coding matrix, corresponding to Y layers, wherein Y is less than or equal to the number of layers of the pre-coding matrix.