(POSTER) Enhancing Structural Health Monitoring with On-device AI

This paper presents a service model for structural health monitoring (SHM) by leveraging on-device AI techniques. Unlike conventional SHM systems relying on server-based diagnostics, our approach incorporates recent advancements in lightweight AI platforms to enable real-time diagnostics directly on the sensor nodes. This integration enhances both real-time performance and accuracy while significantly reducing network load by minimizing the need for data transfer to centralized servers. Through the experimentation using real data collected from a pedestrian bridge, we demonstrate the feasibility and efficacy of the on-device AI approach for SHM. By seamlessly integrating on-device AI into IoT infrastructure, our approach not only offers improved real-time performance and accuracy but also showcases the potential of AI in optimizing SHM systems.