A Safer Way to Monitor Vital Signs

The system could pave the way for safer, more comfortable care in settings ranging from neonatal units to in-home recovery. Pexels

UW–Madison computer scientists are pioneering a new approach to health monitoring: using radar to measure breathing and heart rate without physical contact to the patient. Their system could pave the way for safer, more comfortable care in settings ranging from neonatal units to in-home recovery.

When you think about monitoring heart and breathing rates, you likely picture a wearable device — a wristband, chest strap, or sticky patch connected to a maze of wires. But what if monitoring your breathing or heart rate didn’t require contact at all?

For UW–Madison computer sciences professor Suman Banerjee, that prospect is a near possibility. In collaboration with researchers at the Georgia Institute of Technology and with support from the National Science Foundation, his team is developing a system — called MEDUSA — that uses radar to monitor vital signs without touching patients at all.

Banerjee has long been interested in ways that contactless technology can support health care. “In the neonatal intensive care unit, the very devices that monitor fragile infants can also cause skin abrasions, introduce infection risks, or become tangled,” he notes. For adults, wearables can be uncomfortable or even inaccurate when poorly fitted.

Contactless sensing offers an appealing alternative. Because radar waves can detect small chest movements, they are able to infer vital signs such as breathing and heart rate without attaching anything to the body.

Existing radar systems struggle to detect vital signs outside of controlled lab settings because people naturally move around, turn away from the sensor, and change their posture throughout the day.

MEDUSA overcomes this problem by placing several radar units throughout a room, creating a multiview system that detects vital signs even when some sensors lose line of sight. Custom hardware combined with tightly integrated software separates vital signs from other movements. The result is a system that works in real-life patient settings.

The long-term goal is to make radar hardware more compact.

“We’ve shown that this distributed approach works,” Banerjee says. “Now we want to make it feasible for environments like the neonatal intensive care unit.”

Published in the Summer 2026 issue