Precoding Matrix Quantization for Lower Massive MIMO Feedback Overhead

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In Massive MIMO systems, reporting overheads are high due to the need to report combination coefficients in precoding matrices, leading to significant system performance loss.

Innovation Solution

A method to reduce reporting overheads by using the same amplitude quantization bit quantity and rule for all combination coefficients, grouping them based on amplitude values, and applying different phase quantization precision for different groups based on their impact on performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If high-precision codebook is used to perform linear combination on spatial domain beam base vectors and frequency-domain base vectors, then spectral efficiency is significantly improved, but reporting overheads become relatively high due to needing to report combination coefficients

Engineering Contradiction:
Improvespectral efficiencyVSAvoidreporting overheads
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent segments the combination coefficients into two independent parts: amplitude values and phase values. This segmentation allows different quantization strategies to be applied to each part, reducing the total number of bits required for reporting while maintaining spectral efficiency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different quantization precision to different parts of the precoding matrix based on their local importance. Specifically, it uses higher precision for amplitude values and lower precision for phase values, optimizing the balance between performance and overhead.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If combination coefficients are reported with high precision to minimize quantization error, then channel quantization accuracy is improved, but system performance loss increases due to higher reporting overheads

Engineering Contradiction:
Improvechannel quantization accuracyVSAvoidsystem performance
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent changes the parameter representation by separating amplitude and phase, and applies different quantization bit allocations to each. This parameter transformation maintains channel quantization accuracy while reducing the total overhead, thereby improving system performance.

Inventive Principle:
Principle #35Parameter changes

3Device complexity

If same amplitude quantization bit quantity and rule are used for all combination coefficients, then implementation complexity is reduced, but phase quantization overhead increases

Engineering Contradiction:
Improveimplementation complexityVSAvoidphase quantization overhead
Core Design Contradiction:
Device complexityVSLoss of information

Solution Approach 1:

The patent segments the quantization process into amplitude quantization and phase quantization stages. By using same amplitude quantization for all coefficients (simplifying implementation) and then applying differentiated phase quantization based on amplitude groups (reducing overhead), it resolves the contradiction between implementation simplicity and overhead reduction.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3902153B1Precoding matrix indication method and related device
Publication Date: 2025.12.24 HUAWEI TECH CO LTD
  • EP3902153B1 patent drawingFigure 1
  • EP3902153B1 patent drawingFigure 2
  • EP3902153B1 patent drawingFigure 3~5

AI summary

This application provides a precoding matrix indication method and a related device. In the method, a transmit end can group K combination coefficients based on an amplitude value of each of the K combination coefficients, to obtain Q combination coefficient groups, and determine a phase value of each combination coefficient in each combination coefficient group. The amplitude value of each combination coefficient is determined by using a same amplitude quantization bit quantity and a same amplitude quantization rule, so that the transmit end does not need to additionally indicate a grouping status of the K combination coefficients. In addition, phase quantization bit quantities and/or phase quantization rules used by at least two of the Q combination coefficient groups are different, thereby facilitating use of different phase quantization precision based on different degrees of impact of different combination coefficient groups on performance, and further helping minimize a system performance loss caused by quantization of the combination coefficients while reducing reporting overheads of the foregoing combination coefficients.