WaveletInception Networks for on-board Vibration-Based Infrastructure Health Monitoring

TL;DR

This work tackles on-board vibration-based infrastructure health monitoring by introducing WaveletInception-BiGRU, a fully automated end-to-end framework that analyzes time-domain vibration signals without manual preprocessing. It combines a Learnable Wavelet Packet Transform (LWPT) for multi-resolution spectral decomposition with 1D Inception-ResNet blocks to learn high-level features, and BiGRU modules to fuse operational conditions (notably measurement speed) and model bidirectional temporal dependencies, enabling component-level health estimation aligned with the structure’s layout. The method is validated on two railway case studies: (i) track stiffness estimation, predicting parameters $k_p$ and $k_b$, and (ii) transition-zone identification, achieving high classification accuracy under real-world conditions; the framework consistently outperforms state-of-the-art baselines and ablations in both accuracy and computational efficiency. The results demonstrate that end-to-end, speed-aware vibration analysis can provide high-resolution, automated health profiles suitable for on-board monitoring and localized maintenance planning, with potential applicability to broader railway and civil infrastructure systems.

Abstract

This paper presents a deep learning framework for analyzing on board vibration response signals in infrastructure health monitoring. The proposed WaveletInception-BiGRU network uses a Learnable Wavelet Packet Transform (LWPT) for early spectral feature extraction, followed by one-dimensional Inception-Residual Network (1D Inception-ResNet) modules for multi-scale, high-level feature learning. Bidirectional Gated Recurrent Unit (BiGRU) modules then integrate temporal dependencies and incorporate operational conditions, such as the measurement speed. This approach enables effective analysis of vibration signals recorded at varying speeds, eliminating the need for explicit signal preprocessing. The sequential estimation head further leverages bidirectional temporal information to produce an accurate, localized assessment of infrastructure health. Ultimately, the framework generates high-resolution health profiles spatially mapped to the physical layout of the infrastructure. Case studies involving track stiffness regression and transition zone classification using real-world measurements demonstrate that the proposed framework significantly outperforms state-of-the-art methods, underscoring its potential for accurate, localized, and automated on-board infrastructure health monitoring.

Figures (11)

• Figure 1: Overview of the WaveletInception-BiGRU architecture, including WaveletInception vibration signal feature extraction, operational conditions feature fusion, and BiGRU sequential health conditions estimation.
• Figure 2: Tree-like structure of the Wavelet Packet Transform (WPT).
• Figure 3: 1D Inception block structure.
• Figure 4: Overview of the WaveletInception-BiGRU architecture, including WaveletInception vibration signal feature extraction, operational conditions feature fusion, and BiGRU sequential health conditions estimation.
• Figure 5: The layout of the 10-sleeper track segment.