Article
Engineering, Mechanical
Peijie Zhang, Pu Ren, Yang Liu, Hao Sun
Summary: This paper presents two autoregressive-based matrix factorization methods for imputing missing sensor data and forecasting structural response in large-scale structural health monitoring. Experimental results demonstrate excellent performance of the proposed methods in accurately recovering missing values and predicting future responses.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Engineering, Multidisciplinary
Yuequan Bao, Hui Li
Summary: The conventional vibration-based methods face challenges in accurately detecting structural damages, thus necessitating the development of novel diagnosis and prognosis methods based on various monitoring data.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2021)
Article
Computer Science, Information Systems
Yen -Pin Chen, Chien-Hua Huang, Yuan-Hsun Lo, Yi-Ying Chen, Feipei Lai
Summary: In the development of predictive models, handling missing data is a critical issue. Traditional approaches require a two-step analysis to analyze missing patterns, select variables, impute missing values, and train models. However, these models have limitations in handling high missing rates and variable changes. To address this problem, the researchers propose an attention-based neural network combined with a novel real number representation. This algorithm requires less manual variable selection and can overlook missing data, eliminating the need for imputation. The results show that the proposed algorithm outperforms current approaches in predicting prolonged length of stay in the ICU.
INFORMATION SCIENCES
(2022)
Article
Green & Sustainable Science & Technology
Ge Liang, Zhenglin Ji, Qunhong Zhong, Yong Huang, Kun Han
Summary: The theory of compressive sampling has revolutionized data compression technology by exploiting the sparsity of a signal. Recent advancements in deep generative models have been used to reduce sample size for image data compressive sampling. However, compressive sampling for 1D time series data has not significantly benefited from this progress. This study investigates the application of different deep neural network architectures for time series data compression and proposes an efficient method based on block compressive sampling and the vector quantized-variational autoencoder model with a naive multitask paradigm. Comparative analysis against various methods demonstrates the superiority of the proposed method in both synthetic and real-world data.
Article
Agriculture, Multidisciplinary
Yifan Zhang, Peter J. Thorburn
Summary: Digital agriculture relies on accurate water quality data for efficient resource utilization. This paper proposes a Dual-SSIM model to impute missing time series data in sensor networks, improving the efficiency and reliability of digital agriculture applications. The model outperforms alternatives in water quality data imputation based on various evaluation metrics.
COMPUTERS AND ELECTRONICS IN AGRICULTURE
(2021)
Article
Engineering, Multidisciplinary
Jack Poole, Paul Gardner, Nikolaos Dervilis, Lawrence Bull, Keith Worden
Summary: This article discusses the limitation of labelled data in the practical application of structural health monitoring, and introduces transfer learning methods, specifically domain adaptation. By using statistic alignment, the performance degradation issue under class imbalance in traditional methods is addressed, and the effectiveness of statistic alignment is demonstrated in numerical and real-world scenarios.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2023)
Article
Construction & Building Technology
Islam M. Mantawy, Mohamed O. Mantawy
Summary: This paper explores the use of machine learning techniques for structural health monitoring and damage prediction of bridges. By encoding time-series data into images and training convolutional neural network models, it is possible to effectively predict the damage state of bridges.
STRUCTURAL CONTROL & HEALTH MONITORING
(2022)
Article
Computer Science, Interdisciplinary Applications
Miromar Jose de Lima, Cesar David Paredes Crovato, Rodrigo Ivan Goytia Mejia, Rodrigo da Rosa Righi, Gabriel de Oliveira Ramos, Cristiano Andre da Costa, Giovani Pesenti
Summary: This article introduces a novel approach called HealthMon that computes a health index for machines based on sensor measurements, using time-series decomposition to show the machine's health state over time in an unsupervised manner. The method advances the field of machine learning by providing a more direct and intuitive view of machine degradation, estimating the health index of machines without supervision, and being applicable to a wide range of industrial equipment.
COMPUTERS & INDUSTRIAL ENGINEERING
(2021)
Article
Computer Science, Artificial Intelligence
Le Fang, Wei Xiang, Yuan Zhou, Juan Fang, Lianhua Chi, Zongyuan Ge
Summary: This paper proposes a novel dual-branch cross-dimensional self-attention-based imputation model for multivariate time series. Through global and auxiliary cross-dimensional analyses, the model is capable of learning and utilizing correlations across the temporal and cross-variable dimensions more effectively.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Interdisciplinary Applications
Sven Festag, Cord Spreckelsen
Summary: In the fields of medical care and research as well as hospital management, time series are crucial for making high-quality decisions. The authors developed a system based on generative adversarial networks for imputing and forecasting time series data. It outperforms current state-of-the-art methods and offers flexibility and probabilistic predictions.
JOURNAL OF BIOMEDICAL INFORMATICS
(2023)
Article
Automation & Control Systems
Samuel Harford, Fazle Karim, Houshang Darabi
Summary: This study proposes a method for generating adversarial samples on multivariate time series classification models, combining adversarial autoencoders and gradient adversarial transformation networks. By utilizing adversarial attacks, the adversarial samples are improved by replacing the adversarial generator function with variational autoencoders.
IEEE-CAA JOURNAL OF AUTOMATICA SINICA
(2021)
Article
Computer Science, Artificial Intelligence
Fazle Karim, Somshubra Majumdar, Houshang Darabi
Summary: This paper proposes a method to attack time series classification models using adversarial samples, demonstrating attacks on 42 datasets. The proposed attack generates a larger fraction of successful adversarial black-box attacks compared to the Fast Gradient Sign Method, and a simple defense mechanism is successfully devised to reduce the success rate of adversarial samples. Future researchers are recommended to incorporate adversarial data samples into their training datasets to enhance resilience against adversarial samples.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2021)
Review
Construction & Building Technology
Donghui Xu, Xiang Xu, Michael C. Forde, Antonio Caballero
Summary: In recent years, there has been an increasing installation of Structural Health Monitoring (SHM) systems on bridges worldwide, providing crucial data for bridge assessment and maintenance. Machine Learning (ML) has gained popularity in SHM studies as it can detect damages and perform condition assessment on bridge structures caused by material deterioration. This article summarizes and discusses various ML applications in bridge SHM, providing detailed critiques of each application type, and presents recommendations for future research to fill current gaps.
CONSTRUCTION AND BUILDING MATERIALS
(2023)
Review
Engineering, Multidisciplinary
Meisam Gordan, Saeed-Reza Sabbagh-Yazdi, Zubaidah Ismail, Khaled Ghaedi, Paraic Carroll, Daniel McCrum, Bijan Samali
Summary: This paper classifies the applications of data mining techniques in structural health monitoring (SHM) and presents the development of artificial intelligence, machine learning, and statistical methods. It compares the most commonly used techniques and algorithms in SHM.
Article
Computer Science, Artificial Intelligence
Fabio Giampaolo, Federico Gatta, Edoardo Prezioso, Salvatore Cuomo, Mengchu Zhou, Giancarlo Fortino, Francesco Piccialli
Summary: This study proposes a novel ensemble approach for generating predictions in a multivariate framework. It reduces data dimensionality through an encoding technique, extracts useful information via single predictive procedures, and combines the processed data to produce the final forecast. Extensive experiments demonstrate the higher accuracy and robustness of the proposed ensemble compared to conventional methods and state-of-the-art strategies.
INFORMATION FUSION
(2023)
Article
Transportation
Linchao Li, Yi Lin, Bowen Du, Fan Yang, Bin Ran
Summary: A hybrid model is proposed in this study to tackle the issues of small sample sizes and imbalanced datasets in traffic incident detection. The model utilizes a generative adversarial network (GAN) to expand the sample size and balance the datasets, as well as a temporal and spatially stacked autoencoder (TSSAE) to extract temporal and spatial correlations for incident detection. Evaluations using real-world data show that the proposed model, considering both temporal and spatial variables, outperforms benchmark models and improves real-time detection capacity.
TRANSPORTMETRICA A-TRANSPORT SCIENCE
(2022)
Article
Engineering, Electrical & Electronic
Hanlin Liu, Linqiang Yang, Linchao Li
Summary: This study establishes a hybrid model combining climate factors and GNSS monitoring data to analyze the impact on long-term GNSS records. At a local scale, a stable reference frame PRVI19 was created to avoid bias from long-term plate motions. Results indicate wind as the most influential climate factor, with the model accurately predicting displacements using climate and GNSS records.
JOURNAL OF SENSORS
(2021)
Article
Engineering, Mechanical
Y. W. Wang, C. Zhang, Y. Q. Ni, X. Y. Xu
Summary: The study proposes a Bayesian probabilistic approach to evaluate the occupant comfort of high-rise structures by formulating a Bayesian regression model for wind-induced acceleration responses and using structural health monitoring data. This approach enables the consideration of uncertainties in monitored acceleration responses and quantifying uncertainties in modeling and prediction, leading to a probabilistic assessment of occupant comfort in wind-induced motion of the structure. The evaluation was demonstrated using field monitoring data acquired from a 600 m high supertall structure during six tropical cyclones.
MECHANICAL SYSTEMS AND SIGNAL PROCESSING
(2022)
Article
Computer Science, Information Systems
Liang Zou, Sisi Shu, Xiang Lin, Kaisheng Lin, Jiasong Zhu, Linchao Li
Summary: Bus passenger flow prediction is crucial for public traffic management, and previous studies have used machine learning models to improve prediction accuracy. However, interpretability of the influenced variables remains a challenge. This study proposes a method and implements an XGBoost model to improve prediction accuracy and interpret the results.
WIRELESS COMMUNICATIONS & MOBILE COMPUTING
(2022)
Article
Computer Science, Information Systems
Linchao Li, Xiang Lin, Hanlin Liu, Wenqi Lu, Baoding Zhou, Jiasong Zhu
Summary: This article presents a data-driven and high-dimensional gap-imputation method, Tucker decomposition with L2 regularization, to recover missing displacement data. The results show that considering multiple temporal correlations can improve the accuracy and ability of data recovery.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Computer Science, Information Systems
Bowen Du, Chunming Lin, Leilei Sun, Yangping Zhao, Linchao Li
Summary: This article proposes a heterogeneous structural response prediction (HSRP) framework based on a deep learning model to improve the performance of machine learning models in mining structural health monitoring data. The experimental results show that the proposed model outperforms benchmark models in prediction accuracy and demonstrates good sensitivity and robustness.
IEEE INTERNET OF THINGS JOURNAL
(2022)
Article
Engineering, Electrical & Electronic
Jianfan Chen, Baoding Zhou, Shaoqian Bao, Xu Liu, Zhining Gu, Linchao Li, Yangping Zhao, Jiasong Zhu, Qingquan Li
Summary: This paper proposes a fusion framework that combines data-driven inertial navigation with BLE-based localization using a particle filter. The framework effectively addresses the limitations of single technology localization systems and improves positioning accuracy.
IEEE SENSORS JOURNAL
(2022)
Article
Engineering, Multidisciplinary
Bowen Du, Liyu Wu, Leilei Sun, Fei Xu, Linchao Li
Summary: This study presents a heterogeneous structural response recovery method based on multi-modal fusion auto-encoder, which considers temporal, spatial, and heterogeneous correlations simultaneously. Experimental results demonstrate that the proposed method achieves the best imputation performance under different missing scenarios, and performs better when the missing rate is high.
STRUCTURAL HEALTH MONITORING-AN INTERNATIONAL JOURNAL
(2023)
Article
Environmental Sciences
Hanlin Liu, Linchao Li
Summary: This paper proposes a method called TSHMF that considers both temporal and spatial correlation in monitoring GNSS time series, effectively addressing the issue of missing data. The method outperforms benchmark methods according to the experimental results.
Article
Engineering, Geological
Xingyu Li, Chaodong Zhang, Yue Zheng, Ning Zhang
Summary: This study proposes a novel constrained unscented Kalman filter (UKF) method for updating structural parameters and identifying unknown external excitations in strongly nonlinear structures. The unknown excitation is estimated by a recursive nonlinear least-square algorithm, while a specially developed Matlab-OpenSees recursion platform is used to implement the algorithm. The applicability of the method is verified through investigations on a steel frame and a reinforced concrete bridge.
SOIL DYNAMICS AND EARTHQUAKE ENGINEERING
(2022)
Article
Engineering, Electrical & Electronic
Baoding Zhou, Zhining Gu, Fuqiang Gu, Peng Wu, Chengjing Yang, Xu Liu, Linchao Li, Yan Li, Qingquan Li
Summary: This paper proposes a Deep Learning-based Vehicle Indoor Positioning (DeepVIP) approach using smartphone built-in sensors, and experiments show that this method outperforms existing methods.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2022)
Article
Physics, Multidisciplinary
Ziyi Liao, Minghui Liu, Bowen Du, Haijun Zhou, Linchao Li
Summary: A pressure prediction method based on spatial-temporal neural network is proposed in this study, which considers the spatial and temporal correlations of the pipeline network. The results show that this method achieves the highest accuracy among all tested methods, especially for multiple steps prediction.
PHYSICA A-STATISTICAL MECHANICS AND ITS APPLICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Linchao Li, Junnan Yi, Fei Xu, Hanlin Liu
Summary: Global Navigation Satellite System (GNSS) is widely used in critical infrastructures to provide timely structural settlement information. However, the accurate assessment, measurement, and evaluation of the status are affected by the inevitable noise. This study proposes a new denoising method based on truncated high-order singular value decomposition to effectively reduce the noise in multivariate GNSS signals.
IEEE SENSORS JOURNAL
(2023)
Article
Engineering, Electrical & Electronic
Hanlin Liu, Linchao Li
Summary: Anomaly detection in high-frequency sensing data is challenging due to the large volume of data and limited time. This paper proposes a four-stage model for quick detection using fine-tuned CNNs and compares it with other algorithms. The results demonstrate the effectiveness of ensemble methods in improving overall accuracy as well as accuracy for minor and outlier classes.
IEEE SENSORS JOURNAL
(2023)