Resource-Efficient Technologies in Transport

Spectral information extracted from vibration measurements is the workhorse of operational bridge monitoring. While modal frequencies and damping ratios are widely used, many other frequency-domain descriptors—moments, band energies, spectral shape and entropy measures, and cross-spectral statistics—carry complementary sensitivity to stiffness loss, connection slippage, and boundary condition changes.

This review organizes the landscape of spectral features derived from measured power spectra and cross-spectra for bridges, from the estimator level to decision rules. Leveraging recent deployments and lessons learned in Vietnam and internationally, we (i) propose a feature taxonomy, (ii) provide equations, units, and physical interpretations, (iii) assess confounders and robust preprocessing, and (iv) outline benchmarking and reporting standards for reproducible SHM.

We close with research directions: moment-based damage indicators, multi-sensor coherence graphs, time–frequency spectra under nonstationary traffic, and learning frameworks that remain robust to environmental drift.

Highlights

• Presents a taxonomy of spectral features for vibration-based bridge SHM: modal peaks, band-power, spectral moments, shape descriptors, cross-spectral, and time–frequency features.
• Clarifies the estimation pipeline (sensing → PSD/CSD estimation → feature engineering → learning/decision) with good-practice settings and units.
• Analyzes robustness to environmental/operational variability (temperature, traffic, humidity, wind) and normalization strategies.
• Bridges Vietnam’s current practice (visual inspection–try loading–SHM) with global research trends (FEM vs measurement-driven/PSD).
• Provides a reproducible reporting checklist and future research agenda centered on spectral-moment indicators, cross-spectral coherence, and edge-deployable inference.

Future road transport will surely be primarily with electrical vehicles. These cars will quite soon be self-driving and autonomous.

That gives us extreme demands on safety, reliability of sensors and electronics when exposed to shock and vibration (not to mention the software reliability).

With electrical vehicles, the main source of shock and vibration is the road. It is therefore everything that has to do with interaction between road and vehicle is more important than ever.

Topics Covered in Keynote

• Road quality measurements
• Road modeling for simulation
• Virtual Test Tracks
• Modeling of vehicles running on road models
• Sensor reliability testing
• Methods to evaluate durability and reliability from measurements and test specifications