RAPID: Data Fusion for Structural Assessment of the Fern Hollow Bridge Replacement During Construction

The recent collapse of the Fern Hollow bridge in Pittsburgh, Pennsylvania, highlighted how proactive assessment and maintenance strategies are required to operate the large, interconnected infrastructure of the country. This Grant for Rapid Response Research (RAPID) award will focus on obtaining high-fidelity three-dimensional models of the bridge replacement throughout every major construction phase.

Each sequential scan will be collected in the same reference system, to allow for consecutive reconstructions to be compared quantitatively. Since the landscape will evolve rapidly, this unique data will allow for analyses to assess the fast-paced dynamic changes in structural response. The obtained data and the related analyses bring the potential to significantly aid further technology development towards design for rapid construction as well as efficient bridge monitoring and non-destructive assessment of structural systems.

This project will benefit society by creating novel science and technology for (i) the acquisition of high-resolution data from a bridge by means of different sensing techniques, (ii) data fusion approaches to interpret such information, and (iii) the translation of such information to structural assessment of the resisting elements over time. There is clear potential for this research to serve as a benchmark to improve the maintenance of aging infrastructure in a timely fashion, while also providing valuable information on the reconstruction of a collapsed bridge.

The specific goal of the research is to characterize the time-dependent evolution of the structural response in pre-stressed reinforced concrete bridges during construction. The data will be acquired by the use of both laser-based and camera sensors mounted on unmanned aerial systems. Simultaneous Localization and Mapping algorithms will be employed for the fusion of the two datasets and subsequent reconstruction of the three-dimensional models.

The research hypothesis is that the richer datasets obtained by means of the data fusion algorithms will enable more accurate semantic segmentation of the reconstructed scenes, by leveraging the high spatial resolution of the laser-based measurement and the color information obtained from the camera sensors. This project provides the opportunity to advance both the use and application of measurement technology and the combination of such technology with inverse analysis methods.

The data will be directly usable to create novel “scan-to-analysis” digital tools and enable remarkable advancements in structural health monitoring and prognosis of pre-stressed reinforced concrete bridges, with specific attention devoted to incorporating construction information in the system assessment.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.