Imagine tracking your heartbeat without strapping on a smartwatch or touching any gadget, just by sitting within range of your Wi-Fi. This is made possible with Pulse-Fi, a system recently unveiled by researchers at the University of California (UC), Santa Cruz.
By turning ordinary Wi-Fi signals into a health sensor, Pulse-Fi can measure heart rate with near-clinical accuracy using super-cheap hardware. For years, scientists have been trying to achieve this feat, but a big hurdle they faced was that a heartbeat nudges radio waves only a tiny bit, and those tiny ripples are easily buried by everyday movement in the room.
Pulse-Fi overcame this problem by pairing low-cost transmitters with machine learning that isolates the faint, rhythmic heartbeat signature even when people change posture or stand a few meters away.
Our research successfully “shows how the signal from a household WiFi device can be used for this crucial health monitoring with state-of-the-art accuracy, without the need for a wearable,” the UC Santa Cruz team notes.
Catching heartbeat using WiFi
Wi-Fi devices constantly send radio waves. When those waves pass through or reflect off a person, they’re altered in subtle, measurable ways. So much so that you can use Wi-Fi signals to see people through walls, as a different group of researchers demonstrated in 2023. Pulse-Fi listens to these changes with a transmitter–receiver pair, then runs signal-processing and a trained model to detect the tiny, repeating pattern made by the heart even amid noise.
Earlier attempts involving different technologies failed as soon as the main device was moved a little farther away or changed position. However, in Pulse-Fi’s tests, distance and body position no longer broke the method, and this is what makes this approach stand out.
The Pulse-Fi system is built on ESP32 boards, low-cost microcontrollers with built-in WiFi and Bluetooth that cost only $5 to $10. However, there existed no public heart rate dataset for ESP32 devices, so the team created its own.
They set up an ESP32 Wi-Fi link in the UC Santa Cruz Science & Engineering Library to catch signal changes, and at the same time, measured people’s heart rate with a fingertip oximeter as the “ground truth” (actual heart rate data to validate Pulse-Fi’s readings).
“By combining the data from the Pulse-Fi setup with the ground truth data, they could teach a neural network, which changes in signals corresponded with heart rate,” the researchers said.
They also validated Pulse-Fi on a separate dataset from Brazil that had been gathered with a Raspberry Pi (a $30 low-cost computer popular in DIY projects), giving them a broader range of signals to train and test the system.
Finally, the researchers ran experiments with 118 participants, each measured across 17 body positions (sitting, standing, lying down, walking, and more). After only five seconds of monitoring, the system’s average error was about 0.5 beats per minute (BPM), with longer monitoring times improving accuracy further.
Plus, the method worked even when participants were up to three meters (~10 feet) away, with performance holding steady across positions and distances.
“What we found was that because of the machine learning model, the distance apart basically had no effect on performance, which was a very big struggle for past models,” Pranay Kocheta, one of the study authors and a visiting researcher at UC Santa Cruz, said.
In terms of hardware, the ESP32 boards gave strong results, the Raspberry Pi performed even better, and the team believes that commercial Wi-Fi routers could push the accuracy even higher.
Affordable health-monitoring for all
If refined and validated in real homes, Pulse-Fi could turn existing Wi-Fi gear into a hands-free, low-cost health monitor. That could help people who can’t afford smart wearables or don’t want to wear them all day. The system could also enable quiet, continuous checks that flag issues earlier, which is useful for stress tracking, hydration changes, or cardiovascular risks over time.
However, for now, Pulse-Fi is a proof of concept, not a medical device. Real homes are messy with multiple people moving around, pets, TVs, fans, and different room layouts, which can all disturb the signal. Therefore, larger, multi-home trials will be needed to prove the system’s reliability and confirm performance across a variety of buildings and routers.
The UC Santa Cruz group is currently working on extending Pulse-Fi to track breathing rate and screen for sleep apnea. They are also looking for ways to deploy the tech at scale for spaces with multiple people. The success of these efforts will determine how soon a Wi-Fi-based vital-sign monitor could move from lab demo to living room.
The study is published in the journal IEEE Xplore.
