AI-Enhanced IoT System for Assessing Bridge Deflection in Drive-By Conditions

Abstract

The increasing traffic on roads poses a significant challenge to the structural integrity of bridges and viaducts. Indirect structural monitoring offers a cost-effective and efficient solution for monitoring multiple infrastructures. The presented work aims to explore new sensing strategies based on digital MEMS sensors integrated into an intelligent IoT infrastructure to predict the bridge deflection behaviour for indirect Bridge Structural Health Monitoring purposes. An experimental setup comprising a bridge model and vehicle equipped with a smart sensing node has been used to generate the dataset. Various models for bridge deflection estimation are deployed on the sensorized vehicle, exploiting edge AI capabilities of smart sensors. This study shows the potential of leveraging data-driven technologies to enhance the performance of low-cost sensors. Additionally, it demonstrates the viability of assessing static deflection shapes of bridges through indirect measurements on board vehicles, underlining the potential of this approach to make SHM more cost-effective and scalable.