Each year, the U.S. Department of Education’s Graduate Assistance in Areas of National Need (GAANN) program provides fellowship money for graduate students working in areas identified as critical to the advancement of society. GAANN awards are made to specific academic departments or other programs to support graduate fellowships. In the last grant cycle, the Department of Civil Engineering in the College of Engineering and Applied Sciences was among four Stony Brook University departments that received GAANN awards to further graduate student research.
For civil engineering, the funding will allow the department to train five GAANN fellows to assume educational and leadership roles in a convergent area — Smart Civil Infrastructure Systems (SCIS) enabled by intelligence methods such as Artificial Intelligence (AI), sensing, robotics and the Internet of Things.
“Civil infrastructure systems are critical to our daily lives and the nation’s economic development,” said Ruwen Qin, associate professor and graduate director in the Department of Civil Engineering. “These systems are built on foundational research in civil engineering. However, the increased complexity with the corresponding need to identify, define and solve problems across the boundaries of traditional disciplines necessitates changes in civil engineering. To continue serving our society effectively, civil engineers are expected to be master innovators and integrators of ideas and technology across public, private and academic sectors.”
The proposed GAANN project will strengthen civil engineering, an area of national need, by integrating the enabling technologies and scientific methodologies for SCIS into PhD training.
“The U.S. Highway Bridge Inventory currently has about 617,000 bridges, 42 percent of which are more than 50 years old, and 7.5 percent of which are structurally deficient,” she said. “To avoid a catastrophic incident, all these bridges are required to be inspected every two years. A traditional bridge inspection might require closing traffic and using heavy equipment. It requires a crew of inspectors working at the site for many hours; and some field operations are dangerous, like climbing up to high bridge columns.”
To make bridge inspection safer, faster, cheaper and less interruptive to the traffic, Qin said that future civil engineers will be assisted by unmanned aerial vehicles (UAVs) to collect inspection data and collaborate with AI algorithms to analyze the inspection data.
“How to translate big but unlabeled data into engineering information for infrastructure maintenance and preservation is a challenge,” said Qin. “Self-learning and active learning can be used to train AI models efficiently by leveraging the experiences of civil engineers into the models. Civil engineers can label a small amount of initial training data, and then an AI model can learn by itself. Civil engineers can further identify failures of the AI model and provide a small set of additional data to the AI model for improvement. This trained model will help engineers process big data efficiently.”
— Robert Emproto