Exploring the Health Applications of Vibration Data from Buildings

Haeyoung Noh, assistant professor in civil and environmental engineering at Carnegie Mellon University, had been assessing structural integration in a building. Using vibration sensors, she and her team could characterize a building and gather data regarding structural integrity and safety. Unfortunately, people walking around the building generated their own vibrations and noise that needed to be screened out. But what if that data could be useful?

Noh discovered that monitoring a building for the vibrations caused by individuals can reveal highly detailed and valuable information about the flow of people through the structure. Not only that, it’s possible to focus on just one person at a time. “Individual people have different gaits that induce different floor vibrations,” explained Noh. A footfall is similar to a micro-earthquake; just as vibrations reveal the epicenter of an earthquake, they can be used to localize an individual within a building, based on the person’s unique gait. And remarkably, the characteristics that describe an individual’s gait—height, stride length, and stride frequency— are also related to gait balance, a health indicator.

Vincentian Home and Baptist Home Society. Noh and her researchers first spoke with physical therapists and facility managers about residents’ health status, challenges facing residents, and how CMU’s work could positively impact the facility. “Researchers and medical professionals have very different perspectives,” said Noh. “These conversations are invaluable.”

Vincentian Home A major challenge for elder care facilities and their residents is mobility. Movement is critical for residents to maintain health and an independent lifestyle, but they are often discouraged from walking due to safety and liability issues associated with falls. Using vibration-sensing systems and algorithms developed at CMU, Noh and her team are working toward being able to monitor gait patterns unobtrusively to look for signs of an impending fall.

“Ultimately, this research will lead to the ability to instantly identify when an individual is at risk of falling and create an alert that could be communicated to a nurse,” explained Noh. The technology could even be used to identify breaks in routine, a potential sign of illness or distress.