Damage-Sensitive and Domain-Invariant Feature Extraction for Vehicle-Vibration-Based Bridge Health Monitoring

Liu, Jingxiao; Chen, Bingqing; Chen, Siheng; Bergés, Mario; Bielak, Jacobo; Noh, Haeyoung

doi:10.1109/icassp40776.2020.9053450

Cited by 13 publications

(12 citation statements)

References 18 publications

(38 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For city and area-scale systems, high-cost devices make it dicult to achieve high density or large coverage of the deployment. The state-of-the-art approaches including utilizing the physics model to enhance the data-driven estimation with limited high-resolution sensors [48], utilizing the mobility of the platform to enhance the coverage of the system with limited devices [49,50]. However, the challenges remain.…”

Section: (C4) System Cost and Data Qualitymentioning

confidence: 99%

Opportunities in the Cross-Scale Collaborative Human Sensing of 'Developing' Device-Free and Wearable Systems

Pan

Nguyen

2020

Proceedings of the 2nd ACM Workshop on Device-Free Human Sensing

View full text Add to dashboard Cite

This is a position paper that discusses the challenges of emerging new sensing modalities for both device-free and wearable sensing systems, as well as opportunities lying in the combination of them across multiple information scales. With the development of the Internet of Things (IoT), many devices with sensing-ability have entered people's life. These systems mainly fall into two categories: wearables and infrastructure sensing (device-free). In this paper, we rst briey summarize the state-of-the-art sensing modalities of these two categories, then we discuss the challenges faced by them. We envision a future of IoT human sensing systems that achieves seamless sensing across multiple scales through collaborative information inference by both categories of modalities. Finally, we discuss the opportunities to expand the boundaries of sensing modalities that lie in their collaborative adaptation.

show abstract

Section: (C4) System Cost and Data Qualitymentioning

confidence: 99%

Opportunities in the Cross-Scale Collaborative Human Sensing of 'Developing' Device-Free and Wearable Systems

Pan

Nguyen

2020

Proceedings of the 2nd ACM Workshop on Device-Free Human Sensing

View full text Add to dashboard Cite

show abstract

“…To improve scalability and efciency, many researchers recently proposed mobile sensing methods for BHM. For example, capturing visual and dynamic information by scanning the bridge using vehicles (e.g., cars and unmanned aerial vehicles) in a nondedicated manner [6][7][8][9][10]. Although such mobile sensing methods can capture high-spatial-resolution information of multiple bridges, they have limited temporal information at each coordinate of bridges due to their mobile sensing nature, restricting their ability to continuously infer and diagnose bridge conditions.…”

Section: Introductionmentioning

confidence: 99%

Turning Telecommunication Fiber-Optic Cables into Distributed Acoustic Sensors for Vibration-Based Bridge Health Monitoring

Liu

Yuan

Luo

et al. 2023

Structural Control and Health Monitoring

View full text Add to dashboard Cite

We introduce a nondedicated bridge health monitoring (BHM) system that turns pre-existing telecommunication fiber-optic cables into distributed acoustic sensors to collect bridge dynamic strain responses. Due to extensively installed telecommunication fiber cables in the cities, our telecommunication cable-based system enables efficient and low-cost BHM without the requirement of on-site sensor installation and maintenance; however, it is challenging to extract bridge damage-sensitive information (e.g., natural frequencies and mode shapes) from this nondedicated strain data as it has large measurement noise and error propagation. To overcome the challenge, we develop a physics-guided system identification method that models strain mode shapes based on physics-guided parametric mode shape functions derived from bridge dynamics. We then estimate the displacement mode shape function by analytically double-integrating the modeled strain mode shape. Our method improves the accuracy of estimating bridge damage-sensitive features and reduces error propagation by constraining strain and displacement mode shapes with bridge dynamics. We evaluated our system on a concrete continuous three-span bridge in San Jose, California. Our system successfully identified the first three natural frequencies and reconstructed strain and displacement mode shapes in a meter-scale resolution.

show abstract

“…These methods, however, rely on ML techniques used as signal processing tools for feature extraction. Since the extracted features from ML techniques are not often transferable to other bridges [ 33 ], one concern is that such approaches might bind the monitoring framework to a specific structure or system. Furthermore, most of the current methods are based on the frequency domain.…”

Section: Introductionmentioning

confidence: 99%

A Time-Domain Signal Processing Algorithm for Data-Driven Drive-by Inspection Methods: An Experimental Study

Lan

Lin

2023

Materials

View full text Add to dashboard Cite

Constructional material deterioration and member damage can cause changes in the dynamic characteristics of bridge structures, and such changes can be tracked in the responses of passing vehicles via the vehicle-bridge interaction (VBI). Though data-driven methods have shown promising results in damage inspection for drive-by methods, there is still much room for improvement in their performance. Given this background, this paper proposes a novel time-domain signal processing algorithm for the raw vehicle acceleration data of data-driven drive-by inspection methods. To achieve the best data processing performance, an optimizing strategy is designed to automatically search for the optimal parameters, tuning the algorithm. The proposed method intentionally overcomes the difficulties in the application of drive-by methods, such as measurement noise, speed variance, and enormous data volumes. Meanwhile, the use of this method can greatly improve the accuracy and efficiency of Machine Learning (ML) models in vehicle-based damage detection. It consists of a filtering process to denoise the data, a pooling process to reduce data redundancy, and an optimizing procedure to maximize algorithm performance. A dataset is obtained to validate the proposed algorithm through laboratory experiments with a scale truck model and a steel beam. The results show that, compared to using raw data, the present algorithm can increase the average accuracy by 12.2–15.0%, and the average efficiency by 35.7–96.7% for different damaged cases and ML models. Additionally, the functions of filtering and pooling operations, the influence of window function parameters, as well as the performance of different sensor locations, are also investigated in the paper. The goal is to present a signal processing algorithm for data-driven drive-by inspection methods to improve their detection performance of bridge damage caused by material deterioration or structural change.

show abstract

Damage-Sensitive and Domain-Invariant Feature Extraction for Vehicle-Vibration-Based Bridge Health Monitoring

Cited by 13 publications

References 18 publications

Opportunities in the Cross-Scale Collaborative Human Sensing of 'Developing' Device-Free and Wearable Systems

Opportunities in the Cross-Scale Collaborative Human Sensing of 'Developing' Device-Free and Wearable Systems

Turning Telecommunication Fiber-Optic Cables into Distributed Acoustic Sensors for Vibration-Based Bridge Health Monitoring

A Time-Domain Signal Processing Algorithm for Data-Driven Drive-by Inspection Methods: An Experimental Study

Contact Info

Product

Resources

About