New Smartphone Tool Offers More Accurate Calorie Burn Measure

Harvard researchers develop open-source activity monitor that uses machine learning to interpret leg muscle activity and provide better calorie estimates than commercial fitness trackers.

Published on Feb. 28, 2026

Got story updates? Submit your updates here. ›

Though fitness trackers and smartwatches often display calorie burn estimates, these numbers can be surprisingly inaccurate, with error rates as high as 30-80%. Researchers at Harvard University have developed a new smartphone-based activity monitor called OpenMetabolics that uses machine learning to interpret leg muscle activity and provide more accurate calorie expenditure measurements. In a study with 30 participants, the Harvard device was found to be twice as accurate as commercial fitness trackers.

Why it matters

Accurately measuring physical activity and calorie burn is critical for managing various health conditions, but current wearable devices often fall short. The new smartphone-based system from Harvard could enable more reliable data for both personal health tracking and large-scale research studies, especially in areas where wearable devices are less common.

The details

The OpenMetabolics system uses the smartphone's gyroscope and accelerometer to continuously monitor leg motion, which is then interpreted through a machine learning model to estimate calorie expenditure. This approach is more accurate than the heart rate and wrist motion data used by most commercial fitness trackers. The researchers validated the system by having 30 participants of varying ages, sizes, and fitness levels perform activities like walking, biking, and stair-climbing while wearing the smartphone device and other commercial trackers. The Harvard system proved to be twice as accurate as the commercial devices.

  • The new study was published on February 20, 2026 in the journal Communications Engineering.

The players

Patrick Slade

Assistant professor of bioengineering at the John A. Paulson School of Engineering and Applied Sciences (SEAS) at Harvard University, and the senior author of the study.

Haedo Cho

Ph.D. student who led the development of the OpenMetabolics system and the study validating its accuracy.

OpenMetabolics

An open-source, smartphone-based activity monitor developed by the Harvard researchers that uses machine learning to interpret leg muscle activity and provide more accurate calorie burn estimates.

Got photos? Submit your photos here. ›

What they’re saying

“Physical activity is critical for management of many aspects of health. By relying on a smartphone-based system, this approach can be easily deployed for large-scale use and research studies, even in underserved areas.”

— Patrick Slade, Assistant professor of bioengineering (Communications Engineering)

“Many biomechanics studies that evaluate physical activity are performed in the lab on a treadmill … but this does not capture how people walk in everyday life. People vary their speed during the day. When I catch a bus, I might walk fast. If I'm grocery shopping at Trader Joe's, I might walk slowly. We emulated these types of scenarios through audio prompts.”

— Haedo Cho, Ph.D. student (Communications Engineering)

“I think we should do a better job on this, because between what people perceive, and what the devices tell them, there is probably some mismatch.”

— Haedo Cho, Ph.D. student (Communications Engineering)

What’s next

The researchers are actively exploring ways to use the OpenMetabolics technology to address global health challenges, supported by a Harvard Impact Labs Fellowship focused on understanding and addressing cardiovascular health risks in Latin America.

The takeaway

This new smartphone-based activity monitor from Harvard offers a more accurate and accessible way to measure calorie burn and physical activity, which could lead to better personal health tracking as well as higher-quality research on the health benefits of exercise, especially in underserved areas where wearable devices are less common.