CSI Precoder Quantization for Lower 5G Feedback Overhead
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The high feedback overhead in CSI reporting for beamforming in 5G networks, particularly due to the large number of beam-combining coefficients in the second-stage precoder, necessitates efficient quantization and reporting methods to reduce communication overhead.
Innovation Solution
Decompose each combining coefficient into multiple components, separately quantize and report amplitude and phase values, and use bitmap indexing to reduce feedback bits, along with configuring the strongest coefficients for reporting.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If dual-stage precoding with second-stage precoder F2 is used for beamforming in 5G networks, then beamforming performance and communication link adaptation are improved, but feedback overhead increases approximately linearly with the number of subbands
Solution Approach 1:
The patent segments the second-stage precoder F2 into two components: a wideband amplitude matrix FA that is common across all subbands, and a subband-specific phase matrix FB. This segmentation allows the amplitude information to be reported once for all subbands while only phase differences are reported per subband, reducing feedback overhead from O(S) to O(1) for the amplitude part where S is the number of subbands.
Solution Approach 2:
The patent extracts the wideband amplitude component FA from the subband-specific precoder F2, separating it from the subband-dependent phase component FB. By taking out the common amplitude information and reporting it separately, the patent eliminates redundant amplitude reporting for each subband, thereby reducing overall feedback overhead while maintaining beamforming performance.
2Measurement precision
If all beam-combining coefficients in F2 are quantized and reported, then CSI accuracy is improved, but feedback overhead becomes considerably large for large numbers of subbands
Solution Approach 1:
The patent applies different quantization and reporting strategies to different components of the precoder: wideband amplitudes use one quantization scheme while subband phases use another. Specifically, amplitudes are quantized with 3 bits and phases with 2 bits, optimizing the balance between CSI accuracy and feedback overhead for each parameter type based on their respective importance and variability.
Solution Approach 2:
Instead of directly quantizing and reporting all complex coefficients of F2, the patent inverts the approach by reporting only the differences between subband phases and the wideband amplitude. This inversion transforms the problem from reporting absolute values to reporting relative differences, significantly reducing the number of bits required while preserving CSI accuracy.
3Measurement precision
If phase values are quantized with more bits, then phase accuracy is improved, but feedback overhead increases
Solution Approach 1:
The patent changes the parameter representation by reporting phase differences relative to a reference rather than absolute phase values. This parameter transformation allows for more efficient quantization where only the differential phase information needs to be encoded, reducing the bits required per subband while maintaining phase accuracy through the differential encoding scheme.
Data Source
AI summary
The embodiments herein relate to method performed by a radio network node, a network node, a method performed by a UE and a UE for reducing feedback overhead. The method performed by the UE comprises at least: decomposing each entry corresponding to a (i, j)-th combining coefficient of a precoder matrix into at least two coefficients; quantizing, separately, each of said at least two coefficients with a least one bit, and reporting information related to at least one phase value or at least one amplitude value or at least one phase value and an amplitude value of said quantized coefficient.