Toward Building Resilience Through Reconnaissance: Artificial Intelligence Approaches for Structural Health Monitoring

Natural hazards that have been occurring in various forms, frequencies, and intensities are unavoidable. However, developing a resilient built environment for these natural hazards is achievable. Among multiple ways of achieving community resilience, the Structural Extreme Events Reconnaissance (StEER) Network focuses on “Building Resilience through Reconnaissance.” This approach uses data and observations after natural hazards to make a difference directly on the affected communities, to guide related research, and to inform policy, rather than the regular sequential cycle of these aspects. For data collection, a worldwide trend is the expansion of sensor installation on different elements of the built environment. The acquired big vibration and vision datasets from continuous monitoring and reconnaissance efforts are used to study the response of structures (e.g., buildings and bridges) and ultimately improve design code provisions and practices. The effective and accurate collection and use of data rely on advances in Structural Health Monitoring (SHM) to improve the sustainability and resilience of cities and communities. This paper focuses on the activities of the StEER Network and two frameworks that use different modalities of data for achieving resilience using SHM and reconnaissance: (1) The establishment of the “Bridge Rapid Assessment Center for Extreme Events (BRACE2)” for real-time and near real-time vibration-based SHM and operational decision-making of instrumented bridges and (2) The development of a hierarchical database, namely, “PEER Hub ImagNet” ( \(\phi\) -Net) for integrating multi-task deep learning mechanisms for vision-based SHM using images.