Data Eye Analysis for Communication Link Stability Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing technologies struggle to maintain the reliability of data communication links in computing devices, particularly in mission-critical or safety-critical systems, due to noise and environmental stressors that distort the data eye, leading to inaccurate data sampling and potential system failures.

Innovation Solution

A method and system utilizing a convolutional neural network (CNN)-based controller to monitor and analyze the data eye of a data communication link, adjusting clock-data timing and reference voltage to maintain link reliability, and initiating fail-over actions if instability is detected.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If data link training is performed periodically to re-align clock edge with data eye center, then communication link stability is improved, but the system cannot detect severe eye distortion that requires maintenance intervention

Engineering Contradiction:
Improvecommunication link stabilityVSAvoideye distortion detection capability
Core Design Contradiction:
ReliabilityVSDifficulty of detecting and measuring

Solution Approach 1:

The patent introduces an intermediary detection mechanism that uses machine learning models to analyze data eye characteristics and generate stability scores. This intermediary layer between the physical signal and the control system enables early detection of degradation trends before they reach critical failure points, allowing maintenance intervention while preserving link stability through normal training operations.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces traditional mechanical/electrical eye alignment methods with a software-based machine learning approach. Instead of relying solely on hardware training sequences and analog signal analysis, the system uses trained neural networks to process data eye images and predict link stability, enabling more sophisticated detection capabilities without additional physical infrastructure.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If traditional data link training methods are used to maintain synchronization, then clock edge alignment is achieved, but the methods fail when the eye has become severely distorted

Engineering Contradiction:
Improveclock edge alignment precisionVSAvoidlink performance under severe distortion
Core Design Contradiction:
Measurement precisionVSReliability

Solution Approach 1:

The patent applies preliminary action by training machine learning models in advance using datasets that include severely distorted eye patterns. These pre-trained models can then recognize and compensate for severe distortions during operation, predicting stability outcomes before traditional training methods would fail. The system performs maintenance actions based on predicted stability scores, preventing complete link failure.

Inventive Principle:
Principle #10Preliminary action

3Productivity

If high-speed data communication interfaces are used to increase processing speed, then productivity is improved, but noise and environmental stressors distort the data eye leading to sampling errors

Engineering Contradiction:
Improvedata communication speedVSAvoidnoise and environmental stressor impact
Core Design Contradiction:
ProductivityVSObject-affected harmful factors

Solution Approach 1:

The patent implements feedback by continuously monitoring data eye characteristics through machine learning analysis and using the stability scores to trigger maintenance actions. The system feeds back information about link quality degradation and automatically initiates corrective training or maintenance procedures, creating a closed-loop system that maintains high-speed communication reliability despite noise and environmental stressors.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4196882B1Data link stability detection using computer vision-based data eye analysis
Publication Date: 2025.03.05 QUALCOMM INC
  • EP4196882B1 patent drawingFigure 1
  • EP4196882B1 patent drawingFigure 2
  • EP4196882B1 patent drawingFigure 3A~3B

AI summary

The reliability of a data communication link may be analyzed and otherwise maintained by collecting a two-dimensional array representing a functional data eye, and using a convolutional neural network to determine a score of the functional data eye. The determined score may be compared with a threshold, and an action may be initiated based on the result of the comparison.