CSI Compression Input Type Signaling for Reliable AI Encoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing CSI reporting scenarios with AI/ML models lack clarity and consistency in determining the type of input information for CSI compression and encoding models, leading to inaccuracies and reduced reliability in CSI encoding and decoding.

Innovation Solution

A method and apparatus for determining input information for CSI compression and encoding models based on explicit or implicit indication information, model parameters, and protocol specifications, ensuring consistent understanding among devices and improving accuracy and reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If AI/ML models are introduced into CSI reporting scenarios, then CSI compression and encoding capability is improved, but clarity and consistency in determining input information type deteriorates

Engineering Contradiction:
ImproveCSI compression and encoding capabilityVSAvoidclarity and consistency in determining input information type
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent introduces an intermediary mechanism (indication information) that mediates between the AI/ML model and the input information. This indication information explicitly specifies the type of input information (e.g., channel information, precoding matrix information, rank information) that should be fed into the CSI compression and encoding model, thereby resolving the ambiguity while maintaining the advanced compression capability.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent segments the input information determination process into distinct types (channel information, precoding matrix information, rank information) and uses indication information to specify which segment should be used. This segmentation approach maintains clarity by clearly defining what goes into the AI/ML model while preserving the improved compression capability.

Inventive Principle:
Principle #1Segmentation

2Device complexity

If the type of input information for CSI compression and encoding model is not clearly determined, then device complexity is reduced, but accuracy and reliability of CSI encoding and decoding deteriorates

Engineering Contradiction:
Improvesimplicity of determining input information typeVSAvoidaccuracy and reliability of CSI encoding and decoding
Core Design Contradiction:
Device complexityVSReliability

Solution Approach 1:

The indication information acts as an intermediary that carries explicit type information without adding significant complexity to the devices. The indication information simply specifies the input type (channel information, precoding matrix information, or rank information), maintaining device simplicity while ensuring accurate and reliable CSI encoding and decoding through clear type determination.

Inventive Principle:
Principle #24Intermediary (Mediator)

3Reliability

If explicit indication information is used to specify input information type, then accuracy and reliability of CSI encoding and decoding is improved, but device complexity and signaling overhead increase

Engineering Contradiction:
Improveaccuracy and reliability of CSI encoding and decodingVSAvoidcomplexity of processing indication information
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent uses parameter changes in the indication information to specify different input information types. By encoding the input type as a parameter (e.g., indicating channel information, precoding matrix information, or rank information), the system achieves high reliability in CSI encoding and decoding while keeping the complexity manageable through efficient parameter representation.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS20260066974A1Communication method, apparatus and device, storage medium, chip, and program product
Publication Date: 2026.03.05 BEIJING XIAOMI MOBILE SOFTWARE CO LTD
  • US20260066974A1 patent drawing
  • US20260066974A1 patent drawing
  • US20260066974A1 patent drawing

AI summary

A communication method is provided. The method includes: determining, based on first information, input information for a channel state information (CSI) compression and encoding model, in which the first information is used to determine a type of the input information for the CSI compression and encoding model.