Human Energy Expenditure Models: Beyond State-of-the-Art Commercialized Embedded Algorithms
- Cite this paper as:
- Delgado-Gonzalo R. et al. (2014) Human Energy Expenditure Models: Beyond State-of-the-Art Commercialized Embedded Algorithms. In: Duffy V.G. (eds) Digital Human Modeling. Applications in Health, Safety, Ergonomics and Risk Management. DHM 2014. Lecture Notes in Computer Science, vol 8529. Springer, Cham
In the present study, we propose three new energy expenditure (EE) methods and evaluate their accuracy against state-of-the-art EE estimation commercialized devices. To this end, we used several sensors on 8 subjects to simultaneously record acceleration forces from wrist-located sensors and bio-potentials estimated from chest-located ECG devices. These subjects followed a protocol that included a wide range of intensities in a given set of activities, ranging from sedentary to vigorous. The results of the proposed human EE models were compared to indirect calorimetry EE estimated values (kcal/kg/h). The speed-based, heart rate-based and hybrid-based models are characterized by an RMSE of 1.22 ± 0.34 kcal/min, 1.53 ± 0.48 kcal/min and 1.03 ± 0.35 kcal/min, respectively. Based on the presented results, the proposed models provide a significant improvement over the state-of-the-art.
Keywordsenergy expenditure walking/running speed human model physical activity monitoring
Unable to display preview. Download preview PDF.