Your WiFi can now do more than stream movies; it can sense the beat of your heart.
Engineers at the University of California, Santa Cruz, have developed a system that turns everyday wireless signals into a medical tool.
The proof-of-concept research shows that a simple WiFi transmitter and receiver can measure heart rate with clinical-level accuracy without requiring any smartwatch, chest strap, or hospital monitor.
The technology, called Pulse-Fi, uses inexpensive hardware already found in homes and workplaces.
By applying machine learning algorithms to WiFi signals, the system detects faint changes caused by a heartbeat, filtering out background noise such as movement or environmental interference.
Researchers tested the system on 118 participants and achieved results nearly identical to those from traditional monitors. After just five seconds of signal processing, Pulse-Fi measured heart rate with an error margin of half a beat per minute.
Longer monitoring improved accuracy further, regardless of whether participants were sitting, standing, lying down, or walking.
WiFi gets a heartbeat
Heart rate is among the most basic measures of health, tied to stress, hydration, and fitness. Yet it typically requires wearables or clinical machines to monitor. Pulse-Fi suggests a future where WiFi routers double as non-intrusive health trackers.
“Our results show that this could work in everyday environments, without special positioning or expensive equipment,” said Nayan Bhatia, a Ph.D. student who co-led the project with computer science professor Katia Obraczka.
The team relied on ultra-low-cost ESP32 chips, which retail for $5 to $10, and Raspberry Pi boards costing around $30. Even at these price points, the system proved accurate. Using commercial-grade routers would likely push the performance even higher, the researchers said.
Second life for WiFi signals
The system works by analyzing how radio frequency waves behave as they pass through space. When WiFi signals encounter a human body, they are partially absorbed and scattered.
A heartbeat causes subtle but detectable variations in these signals. Pulse-Fi’s algorithms learned to recognize these variations by training on ground-truth data collected with standard oximeters.
To build a dataset, the researchers set up experiments inside UC Santa Cruz’s Science and Engineering library. They compared WiFi signal fluctuations against actual heart rate measurements, teaching their neural network to map the two. The group also tested their method against an existing dataset from Brazilian researchers who used Raspberry Pi devices, further validating its accuracy.
Beyond the pulse
The team is already working on expanding the system to track breathing rate, which could help in diagnosing conditions such as sleep apnea. Early, unpublished results show promise.
They also found Pulse-Fi could measure heart rate accurately from up to three meters—nearly 10 feet—away, and performance didn’t degrade with distance thanks to the machine learning model.
“What we found was that distance basically had no effect, which was a big challenge for earlier approaches,” said high school researcher Pranay Kocheta, who joined the project through UC Santa Cruz’s Science Internship Program.
The findings of the study have been published in the proceedings of the 2025 IEEE International Conference on Distributed Computing in Smart Systems and the Internet of Things (DCOSS-IoT).