Fifty thousand miles of electric power lines, 4.5 million commercial buildings, 600,000 bridges, and 4 million miles of roads—our nation’s infrastructure profoundly affects our lives and economy. Yet for as much as we rely on these systems, we haven’t maintained them, and the neglect is evident. From 2000 to 2008, 161 bridges failed, including the I-35W Mississippi River Bridge that collapsed, killing 13 people. Power outages cost up to $75 billion a year, and buildings account for 39% of all CO2 emissions. Estimates by the American Society of Civil Engineers estimate that an investment of $3.6 trillion is needed by 2020 to shore up U.S. infrastructure. In light of urbanization, climate change, and constricting budgets, if we want to remain a functioning society we must rethink our approach to infrastructure problems, and that is precisely what faculty and students in the College of Engineering are doing. They are creating sensor-driven solutions to the problems ailing the structures and systems that literally support modern life.


Smart systems make smart decisions

Burcu Akinci is taking the guesswork out of construction management of facilities and infrastructure systems with data-driven tactics that provide engineers with comprehensive, real-time information about structures and their operations.

Replacing outdated, reactionary practices with technologies that streamline decision-making will deliver big payoffs, contends Akinci, a professor of Civil and Environmental Engineering and co-director of the Carnegie Mellon Smart Infrastructure Institute.

Akinci augments building information models (BIMs) with sensing technologies to get realistic, operations-rich information about a building. BIMs are used during design and construction in the industry, and many time and cost savings are documented. However, more often than not, these models alone do not truly depict reality, especially as changes occur during the service lives of the facilities.

To complement BIMs, Akinci uses a variety of sensors to collect data from job sites, existing facilities and infrastructure systems. For example, building automation embedded in HVAC systems and laser scanners, on the ground or mounted on unmanned air vehicles, can provide temperature and spatial information respectively. Through her research, it is possible to feed such data into existing building information models and continuously update information about a facility to create what Akinci calls a “living information model” that provides accurate and timely information about what is happening in a facility. Through these models, it is possible to develop new and more proactive methods for operations and maintenance—soon, managers will look at living information models to virtually identify possible problems and prevent them before they occur.

Sensors provide copious amounts of data, and that is where challenges lie: Is the data accurate? Did we introduce calibration errors into the data? How do we process data so we can tell if we are looking at a wall or a window? How do we fuse data from multiple sources?

As is the nature of research, there are always obstacles to overcome; however, in this case, the rewards will be significant. Akinci points out that the late detection of defects contributes an extra 6% to 12% to the cost of construction projects, and $4.8 billion is spent annually on construction verification. Akinci’s research, which span across all aspects of the life-cycle of facilities and infrastructure systems, offers smart solutions at a time when we need them.


Positioning systems move indoors

Akin to the way global positioning systems (GPS) track objects in the vast outdoors, College of Engineering researchers are creating a powerful new technology to tame the great indoors. The National Science Foundation (NSF) awarded a Carnegie Mellon team $998,387 dollars to develop a cost-effective indoor location service that will incorporate Internet of Things (IoT) technologies to manage commercial facilities. The CMU team, partnering with Bosch RTC Pittsburgh and the Sports & Exhibition Authority, will test the service in the David L. Lawrence Convention Center in Pittsburgh.

Electrical and Computer Engineering professors Bruno Sinopoli and Anthony Rowe co-lead the interdisciplinary team that will reveal information about a building’s interior and track its occupants and content through the use of indoor-mounted beacons and building information models (BIMs).

“We intend to demonstrate that it is possible to track the evolution of a building and its occupants,” say Sinopoli, who along with Burcu Akinci co-directs the Smart Infrastructure Institute.

Novel technologies and data processing provide building information through an interface that’s accessed with a smartphone, tablet, or smart watch. Ultimately, users will customize the service to address building operations management, emergency response, building security, and other unique needs.

Indoor localization, or the ability to locate objects in buildings, promises to revolutionize how people interact with the world. Outdoor localization relies on GPS signals emitted from satellites to provide geographical locations. These signals, however, cannot penetrate walls. This predicament led the CMU team to devise a unique multi-prong approach for determining indoor positioning that will be extremely accurate, reliable, and installed at low cost.

Throughout the convention center, the researchers installed beacons that employ light, sound, wireless local area networks (WLAN), and radio frequency (RF) technologies and techniques. On this front, Rowe is contributing his expertise in large-scale embedded systems and their supporting technologies. By incorporating these technologies with building information models (BIMs), which is Akinci’s area, greater accuracy and reliability can be achieved. This system will piggyback off of the convention center’s existing lighting and communication capabilities and keep installations and maintenance costs in check. Smartphone technologies will also contribute to the service’s heightened accuracy. Bosch, a longtime collaborator with Carnegie Mellon, will contribute sensors and other equipment for the project.

Stakeholders from the convention center will help maintain the beacons and offer input on the design and use of the management applications. Understanding how users interact with the service will help the engineering team refine the service so that it can serve as a platform for other commercial endeavors.

Drawing parallels to GPS, Sinopoli explains that when GPS became available, new apps followed and he predicts that the same will happen with indoor positioning systems. “In the future, it will be up to people to figure out how they want to use this service,” he says.

In addition to Rowe and Akinci, Anind Dey, from the School of Computer Science, is a co-principal investigator on this project. He is designing the interface that will operate on users’ devices.

Mario Bergés, associate professor of Civil and Environmental Engineering, is using sensing and data analysis to make built environments more adaptive and efficient by, for example, detecting anomalous behavior in embankment dams before they occur, or detecting and localizing structural damages in pipelines using acoustic waves. The majority of his present work, however, involves collecting and integrating data about buildings and using it to make better-informed decisions about the operation of these facilities under resource and occupant comfort constraints. One of the biggest challenges in this space, he says, is that “none of the software solutions we have for managing and/or controlling buildings systems scales well across the entire building stock.” In other words, software applications are generally custom developed for each facility. “There is no Windows or OS X for buildings,” says Bergés.

Along with other faculty members in the Smarter Infrastructure Institute at CMU, he has been tackling this problem by developing a common software platform that can be used to integrate not just building systems, but all the various internet-connected devices present in large built environments. This “operating system” will improve our ability to monitor and control our infrastructure and will contribute significantly to energy savings.

Civil and Environmental Engineering Assistant Professor Sean Qian uses data collected from transportation agencies and private sector to manage aging and overcrowded transportation infrastructure systems. The various agencies that manage highway, public transit, and parking systems don’t communicate with each other. By integrating all their data, Qian is learning how people travel—their routes and parking choices, modes of transportation, and departure times. By understanding human behavior he can help agencies better manage passenger and vehicle flow. In one research project, he is studying how roadway reconstruction on the I-95 and the Center City Philadelphia’s Vine Street Expressway affects traffic flow. He is working with the Pennsylvania Department of Transportation to find the best ways to detour traffic by using signaling, signage, and social media to mitigate congestion in the real time.


By the numbers

Source: Carnegie Mellon University

By the numbers

Matteo Pozzi, assistant professor in Civil and Environmental Engineering, focuses on probabilistic risk analysis and decision optimization related to civil infrastructures. He uses various types of sensors to measure strain and vibrations in an effort to mitigate risks and extend the life-spans of civil infrastructure systems.

In one of his projects, the new Scott Hall Building is a testbed for developing and testing a new platform for wireless sensor monitoring. “It is an interesting structure: a cantilever suspended on slender steel columns,” says Pozzi. The monitoring system can collect measurements for the condition assessment of those columns, of the horizontal steel girders, and of the overall performance of the building. During the building’s construction, fiber optic strain sensors were permanently installed and they provide continual information about the condition of the columns. The monitoring of Scott Hall is intended as pilot study for extensive monitoring of other buildings on campus.

Hae Young Noh takes a different approach in her use of sensing technologies—she wants to turn “structures into sensors.” Instead of using cameras and sensors to monitor people or environments, she explores how buildings themselves provide clues about human activity or environmental conditions. People moving through a building create vibrations, while temperature and other factors affect how buildings vibrate, too. Noh, an assistant professor Civil and Environmental Engineering, believes that she can use fewer sensors that measure vibrations to reveal where people are in a building and what they are doing. Her work will make outfitting buildings with sensors less expensive because there will be fewer sensors and reduced installation and maintenance costs.

Noh is collaborating with Electrical and Computer Engineering Associate Research Professor Pei Zhang on a project in a Californian nursing home in which sensors are used to prevent elderly patients from falling. By monitoring vibrations, the sensors detect if a person is fatigued or about to fall. The system alerts caretakers if a patient is moving erratically because the individual may be tired or dizzy from medication.