The objective of this research is to explore methods that model and quantify the uncertainty associated with recorded underground utility location data. Locational uncertainty is the error associated with the methods of mapping underground utilities, and will be modeled as three-dimensional geometric shapes that correspond to the probability that these shapes contain the utilities. Locational uncertainty is dependent on locational accuracy, which will be estimated from data lineage - the format and process used to collect data. The premise associated with this research is that locational uncertainty results from inaccurate locational information that, in turn, is a function of data lineage. The utility data and its quantified uncertainty will be integrated with excavation operations, and communicated to operators by blending the geometry and attribute information with operators' views of the work environment using video-see-through augmented reality. Field experiments will be conducted with industry partners to validate the proposed geospatial uncertainty model and evaluate its benefits of improving excavation safety.
If successful, the research results will lead to improvements in construction operations such as, excavation in the presence of buried utilities, drilling in the presence of hidden obstructions, and robotic construction in the absence of exact prior knowledge of the environment. The project will impact education by promoting subsurface utility engineering as an area of specialization. A diverse group of under-represented students will participate in the project through active collaboration with the Women in Science and Engineering, and the Minority Engineering Program offices at Purdue and the University of Michigan. In summary, the societal benefits of the project are expected to be the improved safety of excavation and other construction operations, and the multi-disciplinary education and training of current and future engineers in an emerging engineering specialty.
