A5047917961- Time Series Analysis and Forecasting
- Anomaly Detection Techniques and Applications
- Icing and De-icing Technologies
- Stock Market Forecasting Methods
- Mobile Crowdsensing and Crowdsourcing
- Smart Materials for Construction
- Data Stream Mining Techniques
- Railway Engineering and Dynamics
- Gaussian Processes and Bayesian Inference
- Structural Health Monitoring Techniques
- Smart Grid Energy Management
- Forecasting Techniques and Applications
- Traffic Prediction and Management Techniques
- Wind and Air Flow Studies
- COVID-19 epidemiological studies
- Non-Invasive Vital Sign Monitoring
- Infection Control and Ventilation
Aalborg University
2023-2025
Norwegian University of Science and Technology
2022
Sichuan University
2020
Cyber-physical system sensors emit multivariate time series (MTS) that monitor physical processes. Such generally capture unknown numbers of states, each with a different duration, correspond to specific conditions, e.g., "walking" or "running" in human-activity monitoring. Unsupervised identification such states facilitates storage and processing subsequent data analyses, as well enhances result interpretability. Existing state-detection proposals face three challenges. First, they...
Correlated time series (CTS) forecasting plays an essential role in many practical applications, such as traffic management and server load control. Many deep learning models have been proposed to improve the accuracy of CTS forecasting. However, while become increasingly complex computationally intensive, they struggle accuracy. Pursuing a different direction, this study aims instead enable much more efficient, lightweight that preserve being able be deployed on resource-constrained...
Blade icing detection is critical to maintaining the health of wind turbines, especially in cold climates. Rapid and accurate allows proper control including shutting down clearing ice, thus ensuring turbine safety. This article presents a wavelet-driven multiscale graph convolutional network (MWGCN), which supervised deep learning model for blade detection. The proposed first uses wavelet decomposition capture multivariate information time frequency domains, then employs temporal (GCN)...
Imputation of Correlated Time Series (CTS) is essential in data preprocessing for many tasks, particularly when sensor often incomplete. Deep learning has enabled sophisticated models that improve CTS imputation by capturing temporal and spatial patterns. However, deep incur considerable consumption computational resources thus cannot be deployed resource-limited settings. This paper presents ReCTSi (Resource-efficient imputation), a method adopts new architecture decoupled pattern two...
Indoor air quality (IAQ) is an important parameter in protecting the occupants of indoor environment. Previous studies have shown that environment with poor ventilation increases airborne virus transmission. Existing research has concluded high rates can reduce risk individuals environments being infected. However, most existing systems are designed to be efficient under non-pandemic conditions. Ultimately, will become hotspots for transmission viruses. Current infection assessments estimate...