Related Experiment Video
Updated: Aug 5, 2026
05:47
Simulation of a Scaled Assembly Process with Collaboration of a Robotic Arm and Monitoring through a Vision System for Quality Control
Published on: August 29, 2025
MGFR-ViT: A Multi-Scale Gated Feature Refinement Vision Transformer for Vibration-Based Fault Diagnosis
Yan Yan1, Ting Shang1, Kun Jia2
1School of Automation and Information Engineering, Xi'an University of Technology, Xi'an 710048, China.
Sensors (Basel, Switzerland)
|July 28, 2026
Summary
A new Multi-Scale Gated Feature Refinement Vision Transformer (MGFR-ViT) improves fault diagnosis for rolling bearings and gears. This method enhances local feature extraction and multi-scale information fusion, outperforming existing models.
Area of Science:
- Mechanical Engineering
- Artificial Intelligence
- Signal Processing
Background:
- Vibration-based fault diagnosis is crucial for rotating machinery.
- Existing methods struggle with local feature extraction, multi-scale information fusion, and noise interference in complex conditions.
Purpose of the Study:
- To propose an advanced fault diagnosis model for rolling bearings and gears.
- To enhance the extraction and fusion of fault features from vibration signals under complex operating conditions.
Main Methods:
- Reconstruction of 1D vibration signals into 2D matrices for Vision Transformer compatibility.
- Development of a locally enhanced Vision Transformer for improved local feature extraction.
- Introduction of a multi-scale gated feature refinement module for adaptive feature enhancement and noise suppression.
Main Results:
- The proposed MGFR-ViT model demonstrated superior diagnostic performance and feature separability on bearing and gear datasets.
- Ablation studies confirmed the effectiveness and complementary nature of the integrated modules.
- The model successfully addressed challenges in local feature extraction, multi-scale fusion, and noise reduction.
Conclusions:
- MGFR-ViT offers an effective feature-learning framework for vibration-based fault diagnosis.
- The proposed method significantly improves diagnostic accuracy in complex operating environments.
- This work advances the capabilities of AI in machinery health monitoring.
