Boosting-inspired online learning with transfer for railway maintenance
Diogo Risca, Afonso Lourenço, Goreti Marreiros
TL;DR
The paper addresses predictive maintenance for railway wheel-rail interfaces under nonstationary operating conditions. It introduces BOLT-RM, a boosting-inspired online learning framework with transfer that grows capacity across domains via a shared feature generator and domain-specific classifiers. The method uses Markov transition field representations of time-series sensor data and a boosting-style training loop with constrained experience replay to enable continual adaptation. On a multi-domain wheel-rail simulation benchmark, BOLT-RM achieves average domain accuracy $0.93$ (vs $0.54$ for an isolated model), forward transfer $0.73$, and backward transfer $3.47 \times 10^{-4}$, with Friedman test $p$-value $9.2 \times 10^{-4}$ supporting significant domain differences and the approach's robustness.
Abstract
The integration of advanced sensor technologies with deep learning algorithms has revolutionized fault diagnosis in railway systems, particularly at the wheel-track interface. Although numerous models have been proposed to detect irregularities such as wheel out-of-roundness, they often fall short in real-world applications due to the dynamic and nonstationary nature of railway operations. This paper introduces BOLT-RM (Boosting-inspired Online Learning with Transfer for Railway Maintenance), a model designed to address these challenges using continual learning for predictive maintenance. By allowing the model to continuously learn and adapt as new data become available, BOLT-RM overcomes the issue of catastrophic forgetting that often plagues traditional models. It retains past knowledge while improving predictive accuracy with each new learning episode, using a boosting-like knowledge sharing mechanism to adapt to evolving operational conditions such as changes in speed, load, and track irregularities. The methodology is validated through comprehensive multi-domain simulations of train-track dynamic interactions, which capture realistic railway operating conditions. The proposed BOLT-RM model demonstrates significant improvements in identifying wheel anomalies, establishing a reliable sequence for maintenance interventions.