An Acceleration Vector Variance based Method for Energy Expenditure Estimation in real-life environment with a Smartphone/Smartwatch integration

Authors: M. Duclos12,4, G. Fleury1, P. Lacomme1, R. Phan1, L. Ren3, S. Rousset2

1 Laboratoire d'Informatique (LIMOS, UMR CNRS 6158), Campus des Cézeaux, 63177 Aubière Cedex, France.
2 INRA, Unité de Nutrition Humaine UMR 1019, CRNH d'Auvergne, 63009 Clermont-Ferrand, France.
3 CRCGM, EA3849, 63000 Clermont-Ferrand, France.
4 CHU Clermont Ferrand, Serv Med Sport & Explorat Fonct, 63003 Clermont Ferrand, France.


Article published into Expert Systems With Applications

Reference.
Duclos M., G. Fleury, P. Lacomme, R. Phan, L. Ren and S. Rousset. An acceleration vector variance based method for energy expenditure estimation in real-life environment with a smartphone/smartwatch integration. Expert Systems with Applications. Vol. 63. pp. 435-449. 2016.


Acknowledgements : Special thanks to the LIMOS researchers providing constant support and help for data collection.

Numerical experiments

The numerical experiments is divided into 3parts:

  1. Population;
  2. Energy expenditure estimation : classification in categories;
  3. Data files.


1. Population

The current population is composed of users with high professional positions with office work and with an average of 33 years old.

Table 1: population of volunteers
Participants Sexe Age height weigh BMI
1 F 53 170 57.6 19.9
2 F 27 165 64.9 23.8
3 F 26 160 59 23
4 F 38 150 55.8 24.8
5 M 29 170 75 25.9
6 M 26 175 69.9 22.8
7 M 45 183 79.8 23.8
8 M 25 170 59.9 20.7
9 M 37 180 74.8 23.1
10 M 24 185 86.6 25.3



2. Energy expenditure estimation : classification in categories

Energy expenditure estimation:


Table 2: Estimation of the resting metabolic rate (W: weight in kg; H: height in cm; A: age in year)
Male (kcal.day-1) Female (kcal.day-1)
Equation 66.473 + 5.0033H +13.7516W – 6.755A 655.0955 + 1.8496H + 9.5634W – 4.6756A

Reference:
Harris J.A., Benedict F.G. A biometric study of basal metabolism in man. In Carnegie Institute of Washington, 1919. link


Table 3: Classification in categories according to our work
Categories Minimal value in MET Maximal value in MET
C1 : standing or sitting activities 0.9 2
C2 : light-intensity activities 2 3.5
C3 : moderate-intensity activities 3.5 6
C4 : vigorous-intensity activities 6 9


Table 4: Personalized category bounds for each participant
Participants ξ Category 1 Category 2 Category 3 Category 4
1 1.14 [1.03 - 2.28[ [2.28 - 3.99[ [3.99 - 6.84[ [6.84 - 10.27[
2 1.12 [1.01 - 2.25[ [2.25 - 3.93[ [3.93 - 6.75[ [6.75 - 10.12[
3 1.07 [0.96 - 2.13[ [2.13 - 3.73[ [3.73 - 6.40[ [6.40 - 9.60[
4 1.09 [0.98 - 2.18[ [2.18 - 3.82[ [3.82 - 6.55[ [6.55 - 9.82[
5 1.08 [0.97 - 2.16[ [2.16 - 3.77[ [3.77 - 6.47[ [6.47 - 9.71[
6 1.02 [0.92 - 2.04[ [2.04 - 3.57[ [3.57 - 6.12[ [6.12 - 9.18[
7 1.13 [1.02 - 2.27[ [2.27 - 3.96[ [3.96 - 6.80[ [6.80 - 10.19[
8 0.96 [0.86 - 1.92[ [1.92 - 3.36[ [3.36 - 5.76[ [5.76 - 8.64[
9 1.08 [0.97 - 2.16[ [2.16 - 3.78[ [3.78 - 6.48[ [6.48 - 9.72[
10 1.08 [0.97 - 2.16[ [2.16 - 3.78[ [3.78 - 6.48[ [6.48 - 9.72[


3. Data files

Armband data file and Smartphone accelerometers data file for each participant

Table 5: Data file for each participant
Participants Armband Data File Smartphone Data File Smartwatch Data File
1 Download Download Download
2 Download Download Download
3 Download Download Download
4 Download Download Download
5 Download Download Download
6 Download Download Download
7 Download Download Download
8 Download Download Download
9 Download Download Download
10 Download Download Download


All Data Files : Download     

© 2015 - LIMOS