Jesse Codling receives ACM BuildSys 2024 paper award for work on heart rate monitoring

Jesse Codling, doctoral student in Electrical and Computer Engineering (ECE), received a Best Paper Runner Up Award for his presentation FloHR: Human Heart Rate Detection at a Distance using Indirect Structural Vibration Sensing at the 2024 Association for Computing Machinery BuildSys Conference held in Hangzhou, China. His work aims to continuously and non-invasively measure heart rate as an indicator of overall health.

Manual methods of tracking heart rate, through periodic pulse readings or using a stethoscope, lack continuity and precision. Wearable devices like smart watches require users to wear them consistently, while ambient sensors require a line-of-sight and introduce privacy concerns. Codling, his PhD advisor, Pei Zhang, and their collaborators became curious whether there might be a way to measure the vibrations produced by heartbeats without being directly in contact with the subject.

“We wondered: if we listen to the floor, can we hear your heartbeat?” Codling said. “It would be great if we can.”

The team first tried to use an accelerometer to measure the tiny vibrations that travel from a seated person’s heart, down their chair, and, they expected, into the floor. Unfortunately, the accelerometer couldn’t pick up the vibrations with enough amplitude to measure heart rate. But Codling was prepared to try another instrument called a geophone, made to be extremely sensitive to vibrations in the earth—similar to the devices used to detect earthquakes.

To use a geophone to measure heart rate, the team had to increase the sensitivity of the device while decreasing the noise detected. Vibrations from footsteps, plumbing or HVAC systems in the building, even the subjects tapping on their smartphones, interrupted the heartbeat pattern. With a combination of hardware configuration and algorithms designed to reduce noise, filter the data, and match the periodicity of a human heartbeat, Codling demonstrated that he could measure heart rate from the floor.

Using 10 healthy subjects in their real households, the team reduced the error of the FloHR (pronounced “floor”) system by at least 25 percent, with a similar accuracy to wearable devices when placed on the chair or the floor at the base of the chair. They could also detect the heart rate from 2m away, with just a slight increase in error.

“FloHR shows promise for ubiquitous heart rate monitoring in home environments, where people spend the most time, while avoiding some of the pitfalls of existing sensing approaches,” Codling said. “From this initial work with healthy adults, we hope to eventually enable ubiquitous long-term heart rate monitoring for the most sensitive populations, including children and the elderly. Such monitoring may serve as an early warning sign for a wide variety of potential health concerns.”

Future work will explore solutions for increased error produced by heart arrhythmias such as bradycardia, tachycardia, and atrial fibrillation; scaling the system to measure the heart rates of multiple people in one space; and refining a method to customize for individual differences.

“It’s impressive that this is even possible,” said one award program committee member.

“This is a really cool project, and the prospect of a building that can sense heart rate without wearable devices or crowding the RF spectrum is really compelling,” added a second committee member.

The FloHR technology is open source. Codling co-authored the paper with Zhang, ECE master’s student Jeffery Shulkin, ECE doctoral students Jiale Zhang and Yen-Cheng Chang, and three additional collaborators at Stanford University and Cisco.