WiFi Beamforming Feedback Reduction via SVD Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current beamforming feedback techniques in WLAN systems, particularly in WiFi 8, face challenges with increased antenna and stream dimensions and larger bandwidths, leading to significant impact on network throughput.
Innovation Solution
The method involves receiving a sounding packet with feedback tones, deriving channel coefficients, forming channel matrices, computing singular value decomposition (SVD) on a subset of feedback tones, and performing lossy compression on the resulting vectors to generate per-tone beamforming angle information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If beamforming feedback is transmitted for each tone in high-dimensional WiFi 8 systems, then beamforming performance is improved, but feedback overhead increases significantly
Solution Approach 1:
The feedback tones are divided into multiple groups, and SVD is performed on each group separately. This segmentation allows the system to process and compress feedback data in manageable chunks, reducing overall feedback overhead while maintaining performance across different tone groups
Solution Approach 2:
The patent applies SVD transformation to convert the original high-dimensional channel matrices into lower-dimensional singular value and singular vector representations. This dimensionality reduction transforms the feedback data from a high-dimensional space to a compressed representation that captures essential channel information with fewer bits
2Loss of information
If SVD is performed on all feedback tones, then compression efficiency is improved, but computational complexity increases
Solution Approach 1:
By dividing feedback tones into groups and performing SVD on each group independently, the patent reduces the computational burden of processing all tones simultaneously. Each group's SVD operates on a smaller subset of data, making the overall computation more manageable while still achieving comprehensive compression across all tones
Solution Approach 2:
The patent performs SVD on a subset of feedback tones rather than all tones, achieving sufficient compression efficiency for practical implementation. This partial action approach balances compression performance with computational feasibility, avoiding the excessive complexity of processing every single tone
3Quantity of substance
If lossy compression is applied to steering matrices, then feedback size is reduced, but feedback precision deteriorates
Solution Approach 1:
The patent transforms the channel representation from original channel coefficients to singular values and singular vectors through SVD. This parameter transformation enables lossy compression by allowing selective quantization and truncation of less significant singular values, thereby reducing feedback size while preserving the most important channel characteristics
Data Source
AI summary
One example discloses a method for beamforming feedback reduction in a WLAN (wireless local area network), including: generating a per-tone beamforming feedback matrix by vectorizing a set of steering vectors using vectorization based dimension reduction (DR-Vec); performing lossy compression on the vectors; reducing the per-tone beamforming feedback matrix to one vector; and generating per-tone beamforming angle information using only the one vector.