A two-year collaborative Eurostars project between the Icelandic health-tech company Driftline and the Swiss technology innovation center CSEM.
“FITSILVER lays the groundwork for next-generation metabolic monitoring solutions that support healthier aging and improved population health.”
— Willum Þór Þórsson, Former Minister of Health (2021–2024), Chairman of the Budget Committee, Member of Parliament
View Full Endorsement (PDF)Funded by the Eurostars program, a EUREKA initiative, as well as Rannís, the Icelandic research institute, and Innosuisse, the Swiss Innovation Agency.
Driftline traveled to Switzerland to meet with CSEM for the initial project meeting and planning. The kickoff included defining project goals, timelines, and deliverables, as well as establishing a collaborative framework for the two-year initiative.
A smartphone app that provides personalized fitness assessments and recommendations using heart rate and activity data was developed.
A wrist-worn prototype that accurately collects heart rate and activity data using optical sensors and motion tracking, tailored for Driftline’s metabolic models and elderly users, was built and tested by CSEM.
Activity testing and gas exchange measurements with the Vyntus metabolic system as part of the METFIT PhD project at the University of Iceland. Participants completed both CPX tests, max speed sprinting tests and Driftline protocol assessments on a treadmill, after which the results were analyzed and compared.
The Calo Tracker underwent initial pilot testing to validate and compare its measurements against the Cosmed K5 metabolic system. The study aimed to assess accuracy, reliability, and usability of the device in real-world conditions.
This master's study by Þórey Hákonardóttir at Reykjavík University validated the Driftline Walking Test by comparing it to the six-minute walking test (6MWT) in a group of 40 elderly participants (ages 70–90). Participants performed both tests and the study tracked heart rate, fitness parameters, and metabolic data.
This study, in cooperation with Sigma Sportslab, tested the metabolic response from a single meal intake under laboratory settings with the Cosmed K5 metabolic system as reference. The subjects also performed fitness testing and underwent blood sample analysis. This study provided important important proof of concept for Driftline's automatic calorie tracking.
A web app that uses Driftline's metabolic model to compute energy intake and expenditure based on heart rate data was developed.
The Calo tracker prototype was validated in a study with 28 subjects at the sports lab at CSEM's headquarters in Neuchatel. Calorie expenditure was measured with the Cosmed K5 under mixed activities and compared to the Calo estimated values. The Calo accuracy was outstanding, with only a 7.7% mean absolute error.
The final project report submitted to Eurostars, summarizing the project outcomes, findings, and recommendations for future work.
The following charts summarize the validation results for Driftline’s metabolic model. First, we compare measured energy expenditure (EE) from gold-standard gas exchange (CPET) to estimates from the model and a basic speed-based approach. The model showed exceptional accuracy, with a mean absolute error of just 5% during treadmill walking and running trials.
Building on these results, the METFIT feeding study tested whether the same model could accurately estimate calorie intake from heart rate signals. Results showed strong agreement with values measured via indirect calorimetry, validating the model’s ability to decompose heart rate signals into food-related and activity-related energy components.
In the final validation study led by CSEM, Driftline’s heart rate–based energy expenditure model was evaluated alongside CSEM’s own reference method combining data from the CALO prototype and motion sensors. The analysis included 20 participants across three protocol phases. As shown below, Driftline’s individualized heart rate model demonstrated strong performance, particularly during the exercise phase (B), where its mean absolute error (MAE) was 8.2% compared to 24.7% for the reference method. When averaged across all phases (ABC), the model maintained a low MAE of 5.8%, underscoring the potential of heart rate–based estimation in mixed activity contexts.
Building on the promising results of FITSILVER, future efforts will expand validation of heart rate–based energy intake and expenditure models in larger, more diverse populations under real-world conditions. Key priorities include refining food wave fitting algorithms, integrating additional physiological signals (e.g., accelerometry, skin temperature, glucose), and developing adaptive AI calibration systems. Hardware optimization will also focus on signal quality, energy efficiency, and user experience.
Following the successful CALO prototype tests, Driftline and CSEM intend to formalize their collaboration through a license agreement to further develop and commercialize the calorie tracking concept. Driftline will compensate CSEM for use of its IP should the technology reach production.
In parallel, the project will focus on scaling the FITSILVER platform, securing regulatory approvals, and preparing for commercialization — with the goal of advancing individualized metabolic monitoring to support healthy aging and preventive care.