Carnegie Mellon University
March 25, 2013

CEE Researchers Use Sensor-Equipped Vehicles to Monitor Bridges

CEE Researchers Use Sensor-Equipped Vehicles to Monitor Bridges

Jacobo Bielak, University Professor of Civil and Environmental Engineering, and Jim Garrett, Thomas Lord Professor of CEE and Dean of the College of Engineering, have teamed up with researchers from other CMU departments and the University of Pittsburgh to explore the potential of using sensor-equipped vehicles to monitor the structural health of bridges. CEE researchers George Lederman (PhD candidate), Zihao Wang (Master’s candidate) and Dr. Fernando Cerda (recent alumnus), and ECE researcher Siheng Chen(Master’s candidate) are all playing key roles in the project, which is being funded by the National Science Foundation and by T-Set, the CMU DOT University Transportation Center.

As bridges age, they tend to develop cracks and other types of damage that may propagate and lead to the failure of the bridge. The team has constructed a scale model of a bridge to test the ability of accelerometers attached to scaled-down vehicles to detect changes in the model bridge’s vibration patterns, which can be a sign of structural damage. “As the vehicle travels over the bridge, it vibrates, but also records the way the bridge vibrates,” Bielak explained. “We then process the signals from this vibration to try to detect changes in the structure.”

If perfected, the team’s monitoring method could serve as a mobile, sustainable way to gather information about the structural health of bridges. It could also be used to complement existing methods such as physical inspections or sensors placed directly on bridges, which are often costly. “This has the potential to have a transformational effect on the way we detect emerging issues in bridges,” Garrett said. “Rather than putting a lot of investment in each individual bridge, we will be able to use vehicles that can collect data from a greater number of bridges without the expense and upkeep of existing methods.” 

The team has been able to draw on research from a number of disciplines. Much of the monitoring is being completed with the help of biomedical imaging tools developed by ProfessorJelena Kovačević of Carnegie Mellon’s Biomedical Engineering and Electrical & Computer Engineering departments. They use a novel classification framework that takes advantage of supervised multi-resolution classification and a semi-supervised learning algorithm, based on the accelerations recorded in the vehicle. “The idea of using a vehicle to identify changes in a bridge is significant in itself, but the fact that we can combine that concept with mathematical tools that were originally developed for a completely different application – the bio-imaging of human organs – is very exciting,” Bielak said. The University of Pittsburgh’s Piero Rizzo, Associate Professor of Civil & Environmental Engineering, has also brought his expertise in structural health monitoring to the project. 

“This project is a quintessential example of the interdisciplinary work that goes on here at Carnegie Mellon,” Garrett commented. “We have researchers from a number of departments and institutions who are really all needed to design an interesting and innovative approach to this problem.”

Having completed their initial experimentation with the scale model, the researchers are now testing their method in the Carnegie Mellon parking garage with the help of a robot developed by Dr. Christoph Mertz of the Carnegie Mellon Robotics Institute. The robot plays the role of a vehicle moving over a beam in the garage, and records vibrations as it moves. Once the team members complete their experiments on the CMU campus, they intend to test their method on full-size bridges. “Many people feel that the current techniques for measuring vibrations in structures cannot detect changes this small,” Bielak said. “If we can show that the tools we are developing allow us to detect these changes, that will be a big step forward.”

Pictured above: A representation of the scale bridge model used in the project. inset: The robot fitted with accelerometers for field tests.