What does this research mean for the field?

An unsupervised deep-learning framework utilizing a convolutional block attention module-enhanced convolutional autoencoder (CBAM-CAE) can reliably detect, localize, and quantify bridge damage using raw acceleration data from random passing vehicles without requiring labeled data. Novelty: ClaimNovelty.METHODOLOGICAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

Develop an unsupervised deep-learning framework to detect, localize, and quantify bridge damage using raw acceleration data captured by passing vehicles.

September 14, 2026Advances in Engineering Software0 citationsOpen Access

Mobile Sensing of Bridge Damage Using Random Vehicle Accelerations and Deep Learning

Mobile sensing of bridge damage with raw acceleration data from random passing vehicles using unsupervised deep learning

View Full Paper

Ask AI

Bookmark

Authors

JZJunyong Zhou YYYuxiang Ye ZZZunian Zhou

Discussion

Member takes

Overview

Computational and laboratory study reveals reliable bridge damage localization from passing vehicle vibrations, highlighting cost-effective infrastructure monitoring without labeled damage data.

Key Points

Develop an unsupervised deep-learning framework to detect, localize, and quantify bridge damage using raw acceleration data captured by passing vehicles.
Integrated time–frequency signal processing with a convolutional block attention module-enhanced convolutional autoencoder (CBAM-CAE) to extract damage-sensitive representations.
Identified and quantified damage via reconstruction-error-based feature analysis without requiring labeled damage samples or explicit modal identification.
Validated the approach through numerical vehicle–bridge interaction simulations under varying road roughness and experimental testing on a laboratory scale model.
Numerical simulations reliably distinguished damaged from healthy bridge states and localized damage corresponding to 40% element stiffness loss even under rough road conditions.
Laboratory experiments on a vehicle–bridge interaction system successfully localized structural damage equivalent to a 3.69% mass addition using only vehicle acceleration responses.

Cite This Study

Zhou et al. (2026) studied this question.

synapsesocial.com/papers/6aa7b2b20926e14a848b13f7 https://doi.org/10.1016/j.advengsoft.2026.104305

View Full Paper

Ask AI

Bookmark

September 14, 2026Advances in Engineering Software0 citationsOpen Access

Mobile Sensing of Bridge Damage Using Random Vehicle Accelerations and Deep Learning

Mobile sensing of bridge damage with raw acceleration data from random passing vehicles using unsupervised deep learning

View Full Paper

Ask AI

Bookmark

Authors

JZJunyong Zhou YYYuxiang Ye ZZZunian Zhou

Discussion

Member takes

Overview

Computational and laboratory study reveals reliable bridge damage localization from passing vehicle vibrations, highlighting cost-effective infrastructure monitoring without labeled damage data.

Key Points

Develop an unsupervised deep-learning framework to detect, localize, and quantify bridge damage using raw acceleration data captured by passing vehicles.
Integrated time–frequency signal processing with a convolutional block attention module-enhanced convolutional autoencoder (CBAM-CAE) to extract damage-sensitive representations.
Identified and quantified damage via reconstruction-error-based feature analysis without requiring labeled damage samples or explicit modal identification.
Validated the approach through numerical vehicle–bridge interaction simulations under varying road roughness and experimental testing on a laboratory scale model.
Numerical simulations reliably distinguished damaged from healthy bridge states and localized damage corresponding to 40% element stiffness loss even under rough road conditions.
Laboratory experiments on a vehicle–bridge interaction system successfully localized structural damage equivalent to a 3.69% mass addition using only vehicle acceleration responses.

Cite This Study

Zhou et al. (2026) studied this question.

synapsesocial.com/papers/6aa7b2b20926e14a848b13f7 https://doi.org/10.1016/j.advengsoft.2026.104305

View Full Paper

Ask AI

Bookmark