Title: Machine Learning-Based Structural Health Monitoring of Concrete Structures using Acoustic Signals
Committee:
Dr. Ying Zhang, Advisor
Dr. Moore, Chair
Dr. Ma
Abstract: The objective of the proposed research is to develop a generalized data-driven approach for the assessment of damage in concrete structures while minimizing the need for labeled data. The preliminary research includes a damage detection model which utilized an autoencoder for anomaly detection and a meta-learning approach which demonstrated the feasibility of a few-shot learning model for damage severity assessment of concrete structures. The proposed research focuses on combining probabilistic learning with meta-learning to provide a prediction of damage severity while taking into consideration the uncertainty of measurements.