Researchers at the University of California, Santa Cruz have developed a new method for measuring heart rate using household WiFi signals, eliminating the need for wearable devices. The system, called “Pulse-Fi,” was designed by Professor of Computer Science and Engineering Katia Obraczka, Ph.D. student Nayan Bhatia, and visiting high school researcher Pranay Kocheta at the Baskin School of Engineering.
Traditional heart rate monitoring typically relies on wearables such as smartwatches or medical equipment. However, this new approach uses low-cost WiFi devices in combination with a machine learning algorithm to detect changes in radio frequency waves caused by a person’s heartbeat.
WiFi transmitters emit radio waves that are partially absorbed by objects they pass through, including human bodies. The Pulse-Fi system processes these signal variations to isolate those linked specifically to heartbeats. “The signal is very sensitive to the environment, so we have to select the right filters to remove all the unnecessary noise,” said Bhatia.
Experiments involving 118 participants demonstrated that Pulse-Fi could measure heart rate with clinical-level accuracy after only five seconds of signal processing. The margin of error was about half a beat per minute and improved further with longer monitoring periods. The technology worked regardless of where people were positioned in a room or their posture—whether sitting, standing, lying down, or walking.
Tests used affordable ESP32 chips priced between $5 and $10 and Raspberry Pi chips costing around $30. Results from Raspberry Pi-based experiments showed even better performance than those using ESP32 chips. More expensive commercial WiFi routers may further enhance accuracy.
Kocheta noted that distance between the person and device did not affect measurement accuracy: “What we found was that because of the machine learning model, that distance apart basically had no effect on performance, which was a very big struggle for past models.” He added: “The other thing was position—all the different things you encounter in day to day life, we wanted to make sure we were robust to however a person is living.”
To train their machine learning model, researchers created their own dataset by collecting data from both an ESP32-based setup and standard oximeter readings in UC Santa Cruz’s Science and Engineering library. They also validated their approach using an existing dataset produced by Brazilian researchers employing Raspberry Pi devices.
Future work aims to expand Pulse-Fi’s capabilities to monitor breathing rates as well as heart rates—a development that could aid in detecting conditions like sleep apnea. Unpublished results suggest strong potential for accurate detection of both breathing rate and apnea events.
Individuals interested in commercial applications are encouraged to contact Marc Oettinger, Assistant Director of Innovation Transfer at UC Santa Cruz: marc.oettinger@ucsc.edu.
