AI-Enabled TSN Bridge for Indoor Time Synchronization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Traditional TSN networks struggle with maintaining time synchronization accuracy in indoor industrial networks due to factors like signal interference, noise, physical obstructions, and device mobility, leading to decreased communication quality and network inefficiencies.

Innovation Solution

A computer-implemented method and TSN bridge utilizing an AI module that processes SINR, delay, and obstacle data to adjust packet sizes and time information, leveraging machine learning algorithms to enhance synchronization accuracy and network performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional TSN networks are used in indoor industrial environments, then basic time synchronization can be achieved, but time synchronization accuracy deteriorates due to signal interference, noise, physical obstructions, and device mobility

Engineering Contradiction:
Improvetime synchronization accuracyVSAvoidsignal interference and noise
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The system performs preliminary calibration by collecting time synchronization data under various known conditions (different SINR levels, obstacle positions, device mobilities) before actual operation. This pre-collected data is used to train machine learning models that can predict and compensate for synchronization errors in real-time, addressing the harmful effects before they degrade performance

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system continuously monitors actual time synchronization performance and feeds this information back to the machine learning model. The model uses this feedback to adjust its predictions and compensate for ongoing interference and noise effects, maintaining accuracy despite dynamic environmental conditions

Inventive Principle:
Principle #23Feedback

2Adaptability or versatility

If traditional TSN networks are used with mobile devices, then device mobility is supported, but time synchronization accuracy deteriorates due to changing network conditions

Engineering Contradiction:
Improvedevice mobility supportVSAvoidtime synchronization accuracy
Core Design Contradiction:
Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The system transitions from static TSN configuration to dynamic adaptation by using machine learning models that continuously adjust synchronization parameters based on real-time device mobility patterns. The model learns how different devices move through the environment and predicts synchronization requirements dynamically, maintaining accuracy despite mobility

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes synchronization parameters (such as timing advance values, packet transmission intervals, and reference signal frequencies) based on predictions from the machine learning model. These parameter adjustments are made in response to detected mobility patterns and environmental changes, maintaining synchronization accuracy across varying conditions

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If AI module with machine learning algorithms is added to TSN bridge, then time synchronization accuracy is improved, but device complexity increases

Engineering Contradiction:
Improvetime synchronization accuracyVSAvoidTSN bridge complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The TSN bridge with AI module performs self-calibration and self-optimization by automatically collecting data, training models, and adjusting parameters without external intervention. The system serves itself by continuously improving its own synchronization performance through embedded machine learning, reducing the need for complex external configuration and management infrastructure

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4475464B1A method and TSN bridge for optimizing performance of an indoor industrial network and an industrial network
Publication Date: 2025.08.06 CUMUCORE OY
  • EP4475464B1 patent drawingFigure 1
  • EP4475464B1 patent drawingFigure 2
  • EP4475464B1 patent drawingFigure 3A~3B

AI summary

Disclosed is a computer-implemented method (600) and a TSN bridge (202) for optimizing performance of an indoor industrial network (100), an indoor industrial network (100) therefor, and a method (600) for predicting network problems in an indoor industrial network (100). The method (600) optimizes the performance of an indoor industrial network (100) by utilizing an AI module (212) in a TSN bridge (202). The AI module (212) receives SINR and Delay information from a sync agent in a mobile router, environmental parameters from sensors, and obstacle data from cameras (136). The AI module (212) identifies potential time accuracy deterioration in device clocks based on the received data. This information is then communicated to a CNC (502) and a CUC (504) for adjustment of data packet size. Finally, a User Plane Function (UPF) sync agent (120) is instructed to send updated time information to the devices for clock adjustments.