Work

The Attention Enhanced Autoencoder for Balise Anomaly Detection in Potting Process

Year: 2025

Type: article

Abstract: Balise is a critical trackside component in railway systems, operating in harsh environments where silicone gel potting enhances reliability. However, the potting process is irreversible; defects lead to costly scrap and operational delays. For this reason, this study introduces an anomaly detection method for the transponder potting process for the first time.We propose an attention-enhanced autoencoder (AEA) for anomaly detection. Our approach integrates three core innovations: max-min normalisation to handle scaling discrepancies between input attributes, an attention mechanism that dynamically highlights key features, and empirically derived adaptive reconstruction error thresholds to discriminate anomalies. Experiments on real production data show that the AEA model achievesPrecisionof 97.10%,Recallof 100.00%, weightedF1−Scoreof 98.53%,Accuracyof 99.90%,ROC−AUCof 99.96%, andG−meansof 98.54%, outperforming all baselines. Compared to the best baseline AE model, the AEA model improvesPrecisionby 1.45%,Recallby 1.49%, weightedF1− Score by 1.47%, Accuracy by 0.09%,ROC − AUCby 0.03%, andG − meansby 1.47%. These results demonstrated that our model significantly outperforms existing well-regarded models in anomaly detection performance.

Source: IEEE Access

DOI: 10.1109/access.2025.3569756

Authors: Ke Ye, Jingpin Jiao, Fanghua Chen, Qing Xu, Gehui Liu, Jun Wang

Institutions: Beijing University of Technology, China Railway Signal & Communication (China)

Language: English

Topic: Anomaly Detection Techniques and Applications

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Industry, innovation and infrastructure

Open Access status: gold

Estimated APC paid: $2,075