Skip to main navigation Skip to main content
  • E-Submission

JKSPE : Journal of the Korean Society for Precision Engineering

OPEN ACCESS
ABOUT
BROWSE ARTICLES
EDITORIAL POLICIES
FOR CONTRIBUTORS
REGULAR

에너지소모량 추정용 IMU 신호 기반 신경망 모델의 비교분석

Comparative Analysis between IMU Signal-based Neural Network Models for Energy Expenditure Estimation

Journal of the Korean Society for Precision Engineering 2024;41(3):191-198.
Published online: March 1, 2024

1 한경대학교 융합시스템공학과

2 한경대학교 ICT로봇기계공학부

1 Department of Integrated Systems Engineering, Hankyong National University

2 School of ICT, Robotic & Mechanical Engineering, Hankyong National University

#E-mail: jklee@hknu.ac.kr, TEL: +82-31-670-5112
• Received: October 20, 2023   • Revised: December 28, 2023   • Accepted: January 2, 2024

Copyright © The Korean Society for Precision Engineering

This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited.

  • 7 Views
  • 0 Download
  • 1 Crossref
prev next

Citations

Citations to this article as recorded by  Crossref logo
  • Development of a Novel Ventilation Estimation Model Based on Convolutional Neural Network (CNN)
    Jeongyeon Chu, Jaehyon Baik, Kangsu Jeong, Seungwon Jung, Youngjin Park, Hosu Lee
    Journal of Korea Robotics Society.2025; 20(1): 138.     CrossRef

Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:

Include:

Comparative Analysis between IMU Signal-based Neural Network Models for Energy Expenditure Estimation
J. Korean Soc. Precis. Eng.. 2024;41(3):191-198.   Published online March 1, 2024
Download Citation

Download a citation file in RIS format that can be imported by all major citation management software, including EndNote, ProCite, RefWorks, and Reference Manager.

Format:
Include:
Comparative Analysis between IMU Signal-based Neural Network Models for Energy Expenditure Estimation
J. Korean Soc. Precis. Eng.. 2024;41(3):191-198.   Published online March 1, 2024
Close

Figure

  • 0
  • 1
  • 2
  • 3
  • 4
Comparative Analysis between IMU Signal-based Neural Network Models for Energy Expenditure Estimation
Image Image Image Image Image
Fig. 1 Input type for Method 1 and 2
Fig. 2 Experimental setup
Fig. 3 Treadmill speed profile for each walking/running test
Fig. 4 Energy expenditure (EE) estimation results of Datasets 1 and 2 from the (a) M1 and (b) M2 based on the foot-attached IMU
Fig. 5 Energy expenditure (EE) estimation results of Datasets 1 and 2 from the (a) M1 and (b) M2 based on the shank-attached IMU
Comparative Analysis between IMU Signal-based Neural Network Models for Energy Expenditure Estimation

Averaged normalized root mean squared error, i.e., NRMSE in W/kg (with mean absolute percent error, i.e., MAPE in %) of energy expenditure from Method 1 (M1) and 2 (M2) for each training and test sets

Training set Test set Method Sensor locations
Chest Wrist Thigh Shank Foot
Dataset 1 M1 1.20 (16.24) 1.37 (19.08) 1.18 (17.99) 1.09 (15.46) 0.95 (13.42)
M2 1.11 (14.23) 1.86 (25.19) 1.42 (21.26) 1.20 (16.94) 1.02 (14.79)
Dataset 2 M1 0.88 (11.40) 1.06 (12.94) 1.13 (14.57) 0.93 (11.25) 0.87 (10.95)
M2 0.83 (10.57) 1.12 (12.94) 1.47 (18.26) 1.14 (14.98) 1.17 (15.21)
Dataset 3 Dataset 1 M1 1.26 (17.99) 1.55 (21.60) 1.14 (17.93) 1.01 (14.61) 0.86 (12.25)
M2 1.15 (15.61) 1.82 (24.21) 1.41 (21.46) 1.20 (18.28) 0.86 (12.53)
Dataset 2 M1 1.26 (16.46) 1.53 (20.04) 1.18 (16.64) 1.04 (12.99) 0.91 (11.64)
M2 1.28 (14.80) 1.82 (22.69) 1.80 (22.73) 1.37 (18.34) 1.47 (16.61)
Dataset 3 M1 1.25 (16.06) 1.51 (19.17) 1.11 (15.56) 1.05 (13.29) 0.89 (11.29)
M2 1.25 (14.98) 1.82 (22.00) 1.57 (20.41) 1.23 (16.50) 1.14 (13.57)

*Dataset 1: Level and inclined walking data, Dataset 2: Level walking and running data, Dataset 3: Total data

Table 1 Averaged normalized root mean squared error, i.e., NRMSE in W/kg (with mean absolute percent error, i.e., MAPE in %) of energy expenditure from Method 1 (M1) and 2 (M2) for each training and test sets

*Dataset 1: Level and inclined walking data, Dataset 2: Level walking and running data, Dataset 3: Total data