A Multi-Modal Data Fusion Framework for Real-Time Infrastructure Health Monitoring in Critical Civil and Industrial Systems

Nowadays, there is a demand to design reliable procedures for auditing infrastructure objects of various types. For example, Structural Health Monitoring (SHM) is considered a critical utility for the functioning and maintenance of constructions, but the current practice has many shortcomings in analysing heterogeneous data from different types of sensors. The article describes an innovative Multi-Modal Data Fusion Framework for Real-Time Structural Health Monitoring, which eliminates all the shortcomings mentioned above, providing an efficient method for finding the most relevant displacements. The five-level processing pipeline involves a number of exclusive methods, including suppressing noise, multi-stage algorithm fusion, and distributing calculations between edge and cloud-based computing resources. As a result, a reliable mathematical model was found, which forms the basis for solving more significant problems of predictive diagnostics and reducing the number of false alarms in large-scale industrial applications.