Decoding Muscle Fatigue With Radar

Quantifying muscle fatigue during exercise remains a persistent challenge across sports science and rehabilitation. Existing approaches often rely on subjective reporting or contact-based sensors. While effective, these methods can be cumbersome and sensitive to placement. Researchers from Carnegie Mellon’s College of Engineering have developed a contactless sensing system that aims to estimate muscle force and fatigue by analyzing vibration patterns captured by mmWave radar.

Every time you lift a weight, your muscles fire in a complex, overlapping pattern. Individual muscle fibers activate asynchronously, producing tiny surface vibrations on the skin. These micro-movements are subtle and notoriously difficult to interpret. To most sensors, they appear as random noise. Drawing on principles from chaos theory, the researchers show that muscle vibrations, while noisy, follow hidden deterministic patterns.

“Our system decodes this data and makes sense of it,” explains Jiangyifei Zhu, an electrical and computer engineering Ph.D. student and lead author of the paper. “We are trying to measure how tired your muscles are during exercise, so you can train effectively without pushing yourself too far and risking injury.”

Using mmWave radar, the technology is directed at a person’s muscles during exercise. As the muscles contract and vibrate, the radar captures subtle changes in reflected signals. It can locate the active muscle group and isolate the relevant vibration signals. 

“We then apply chaos-inspired modeling to uncover structure, which translates those patterns into estimates of muscle force and fatigue,” says Justin Chan, assistant professor of electrical and computer engineering and advisor to the project.

The team will present their paper, “GigaFlex: Contactless Monitoring of Muscle Vibrations During Exercise with a Chaos-Inspired Radar,” at the ACM/IEEE International Conference on Embedded Artificial Intelligence and Sensing Systems.

In a study with 23 participants, the team showed they could estimate maximum muscle force with about 6 percent error, a level of accuracy comparable to traditional contact-based sensors. Additionally, the system could detect repetitions in reserve, a key metric in strength training that estimates how many repetitions a person has left before failure. 

The research shows that contactless radar, combined with advanced mathematical modeling, can reliably monitor muscle force and fatigue. 

“This could help people train more effectively and reduce injury risk, while also opening the door to tracking other complex biological signals,” says Swarun Kumar, Sathaye Family Foundation Professor of electrical and computer engineering and advisor to the project. “There’s a fine line between not lifting enough weight to see results, and lifting too much, which can cause injury. Our goal is to help people at any stage of their fitness journey.”

The paper was authored by Jiangyifei Zhu, Yuzhe Wang, Tao Qiang, Vu Phan, Zhixiong Li, Evy Meinders, Eni Halilaj, Justin Chan, and Swarun Kumar.