.png)
1 month ago
5
This article is part of our exclusive IEEE Journal Watch series in partnership with IEEE Xplore.
Wi-Fi signals today primarily transmit data. But these signals can also be used for other innovative purposes. For instance, one California-based team has proposed using ambient Wi-Fi signals to monitor a person’s heart rate.
The new approach, called Pulse-Fi, offers advantages over existing heart rate monitoring methods. It’s low-cost and easily deployable, and it sidesteps the need for people to strap a device to their body.
Katia Obraczka is a professor at the University of California, Santa Cruz, who led the development of Pulse-Fi. She notes that continuous tracking of vital signs, including heat rate, can help flag health concerns such as stress, dehydration, cardiac disease, and other illnesses. “But using wearables to monitor vitals can be uncomfortable, have weak adherence, and have limited accessibility due to cost,” she says.
Camera-based methods are one option for remote, contactless tracking of people’s heart rate without a wearable device. However, these approaches can be compromised in poor lighting conditions, and also raise privacy concerns.
In the search for a better option, Obraczka, along with postdoctoral student Nayan Sanjay Bhatia and high school intern Pranay Kocheta working in her lab, sought to create Pulse-Fi. “Pulse-Fi uses ordinary Wi-Fi signals to monitor your heartbeat without touching you. It captures tiny changes in the Wi-Fi signal waves caused by heart beats,” says Obraczka.
How Can Wi-Fi Signals Measure Someone’s Pulse?
Specifically, the team designed Pulse-Fi to filter out background noise and detect the changes in signal amplitude brought about by heartbeats. They developed an AI model—capable of running on a simple computing device, such as a Raspberry Pi—which then reads the filtered signals and estimates heart rate in real time.
The team tested their approach in two different experiments, which are described in a study published in August at the 2025 International Conference on Distributed Computing in Smart Systems and the Internet of Things.
First, the researchers had seven volunteers sit in a chair at various distances of 1, 2 and 3 meters from two ESP32 microcontrollers that used Pulse-Fi to estimate the volunteers’ heart rates, comparing these data to heart rate measurements taken by a pulse oximeter. In the second experiment, Pulse-Fi was used on a Rasberry Pi devices to monitor the heart rates of more than 100 participants in different positions, including walking, running in place, sitting down, and standing up.
The results show that the system performs on par with other reference sensors, and Pulse-Fi’s less than 1.5 beats-per-minute error rate compares favorably to other vital-sign monitoring technologies. Pulse-Fi also maintained sufficient accuracy despite the person’s posture (e.g., sitting, walking) or distance from the recording device (up to 10 feet away). Based on these results, Obraczka says the team plans to establish a company to commercialize the technology.
She adds that Pulse-Fi can work in new environments that its underlying AI model hadn’t trained for. “The model generalized well in [a new] setting, showing it’s not just memorizing, but actually learning patterns that transfer to new situations,” she says.
Obraczka also notes that the devices that Pulse-Fi runs on are affordable–with ESP32 chips costing about $5–$10 USD, with Raspberry Pis costing about $30 USD.
As yet, the researchers have only tested Pulse-Fi on a single user in the room at a time. The team is now beginning to pilot the approach on multiple users simultaneously. “In addition to working on multi-user environments, we are also exploring other wellness and healthcare applications for Pulse-Fi,” Obraczka says, citing sleep apnea and breathing rates as examples.
