UC Santa Cruz Tests Pulse-Fi Heart Monitoring on Low-Cost Wi-Fi Hardware

Key Takeaways

• Pulse-Fi estimates heart rate by analyzing small heartbeat-related changes in ordinary Wi-Fi signals.
• Tests involving more than 100 participants produced errors of less than 1.5 beats per minute.
• Multiuser monitoring, clinical validation, privacy controls, and regulatory positioning remain key commercialization questions.

Researchers at the University of California, Santa Cruz have developed Pulse-Fi, a contactless heart-rate-monitoring system that runs on inexpensive Wi-Fi hardware. Rather than requiring a smartwatch, chest strap, or camera, the system detects subtle changes in wireless signals caused by a person’s heartbeat.

The approach is led by a UC Santa Cruz professor, working with a postdoctoral researcher and a high-school intern. Their work was highlighted by IEEE Journal Watch and IEEE Xplore, following a study presented at the 2025 International Conference on Distributed Computing in Smart Systems and the Internet of Things.

“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 heartbeats,” the lead researcher said.

Wi-Fi sensing does not measure the heart electrically. Instead, Pulse-Fi filters background noise and looks for minute variations in signal amplitude associated with physical movement from the heartbeat. An AI model interprets those filtered patterns and estimates pulse rate in real time.

Importantly, that processing can happen on relatively modest edge hardware. The researchers ran Pulse-Fi on ESP32 microcontrollers and Raspberry Pi devices, avoiding the need to stream all raw sensing data to a large cloud system. The lead researcher noted that ESP32 chips cost about $5, while Raspberry Pi devices serve as another highly affordable edge option.

That hardware profile could matter for senior living facilities, remote-care programs, wellness monitoring, and other environments where outfitting each person with a wearable may be expensive or impractical. Wearables also depend on users remembering to charge, wear, and correctly position them. Cameras remove the contact requirement but introduce different concerns, including lighting sensitivity and the collection of identifiable imagery.

Pulse-Fi’s evaluation used two experiments. In the first, seven volunteers sat at distances of 1, 2, and 3 meters from two ESP32 microcontrollers. Pulse-Fi’s estimates were then compared with measurements from a pulse oximeter.

The second experiment was broader. Raspberry Pi devices monitored more than 100 participants while they walked, ran in place, sat down, and stood up. The researchers reported an error of less than 1.5 beats per minute, noting that performance remained sufficiently accurate across different postures and distances of up to 10 feet.

Can those results translate from controlled testing into busy care settings? That is the larger commercial question. Homes, clinics, and assisted-living facilities contain moving people, furniture, electronic interference, pets, and Wi-Fi traffic that may complicate signal interpretation. Still, the lead researcher stated the model performed well in a setting that had not been part of its training data, suggesting that it learned transferable patterns rather than simply memorizing one environment.

The current limitation is significant: testing has involved only one monitored person in a room at a time. Separating several overlapping physiological signatures from the same wireless environment will be a tougher signal-processing problem. The team has begun piloting Pulse-Fi in multiuser settings and is also exploring breathing-rate and sleep-apnea applications.

Commercial deployment would bring another layer of work. A product intended for general wellness faces a different validation and regulatory path from one used to support diagnosis or clinical decisions. Buyers would likely want evidence across larger and more diverse populations, along with documented performance during ordinary daily activity. They would also need clear policies governing consent, retention, access, and secondary use of physiological data.

The lead researcher noted the team plans to establish a company to commercialize Pulse-Fi. Its early results point to an intriguing edge-AI proposition: repurposing familiar wireless infrastructure as a low-cost health sensor. Whether that proposition becomes a dependable product will hinge on multiuser performance, real-world validation, and careful handling of the data that ordinary Wi-Fi signals can reveal.