Rowan engineering experts are developing AI tools to improve NJ’s bridge asset management strategies

Repairing bridges is much like a root canal–painful but sometimes necessary, says Islam Mantawy, Ph.D., an assistant professor in the Department of Civil & Environmental Engineering at the Henry M. Rowan College of Engineering. 

Before dentists recommend a root canal, they need data about the extent of the problem, which in turn drives decisions about the timing, scope and type of intervention. Similarly, bridge asset management needs information about the bridge’s condition, a forecast about its near-term deterioration potential and whether the fix can wait without jeopardizing safety. Such a delicate balance is necessary to optimize limited budgets. 

Mantawy recently received a two-year, $700,000 grant from the New Jersey Department of Transportation (NJDOT) to develop effective bridge asset management strategies. Using advanced artificial intelligence (AI) and machine learning algorithms, Mantawy will address the data management, predictive modeling and performance forecasting parts of the asset management equation, in collaboration with Nidhal Bouaynaya, Ph.D., associate vice chancellor for artificial intelligence and director of Rowan’s Machine, Artificial Intelligence and Virtual Reality Center (MAVRC). 

Such efficiencies come not a minute too soon. New Jersey has 6,827 bridges and maintaining up-to-date information on their condition is time- and resource-intensive. Engineers cull data on each bridge from extensive reports, tracking damage over time to determine when urgent repairs are needed. 

The research team will train a machine learning model with images of various kinds of damage, such as spalling (or fragmenting), cracks and corrosion, so it can learn what defects and problems look like. The trained model will then process inspection images and flag potential issues for engineers to review. Leaning on machine learning to do a first pass on the data and serve as an additional pair of eyes could save engineers a lot of time. 

Another efficiency Mantawy hopes to deliver: helping engineers comply with Federal Highway Administration mandates for Specifications for the National Bridge Inventory (SNBI) that require states to refresh bridge asset data inventory every few years. What used to be a couple dozen information fields to be filled out has now ballooned to approximately a hundred, Mantawy points out. 

“We hope to automate this data extraction and SNBI report documentation process, using generative AI to extract, summarize and organize key information, so that inspection technicians can focus on higher-value assessments rather than going bridge by bridge through lengthy documents,” says Mantawy, who is also affiliated with Rowan’s Center for Research & Education in Advanced Transportation Engineering Systems (CREATES). 

Adriana Trias Blanco, Ph.D., an assistant professor in the department, will collaborate with Mantawy to use Light Detection and Ranging (LiDAR) technology to create comprehensive “maps” of bridges so every problem can be spotted and tracked over time. LiDAR uses light beams that bounce off an object and measures the time it takes to bounce back. The technology creates large “point clouds,” where each point describes a spatial location of the asset and delivers better information about the severity and extent of damaged areas. AI will help extract relevant information from these datasets, too.

Collaborating with faculty and students from the Connected Cities with Smart Mobility Transportation (C2SMARTER) center at New York University and Rutgers University, the research team will also use Life Cycle Cost Analysis and deterioration models to make strategic decisions about when to replace bridge components and the long-term effectiveness of recommended solutions. 

“Our work is about using human and monetary resources more efficiently,” Mantawy says. “We want safe and reliable bridges while optimizing how we spend our resources today and minimizing future costs. AI-driven bridge asset management helps us do that.”