Related Experiment Video
Updated: Apr 19, 2026
08:45
Assessment of Physical Activity Intensity with Accelerometers and Oxygen Consumption
Published on: June 20, 2025
779
Energy Expenditure Prediction Using Raw Accelerometer Data in Simulated Free Living
Alexander H K Montoye1, Lanay M Mudd, Subir Biswas
11Human Performance Laboratory, Ball State University, Muncie, IN; 2Department of Kinesiology, Michigan State University, East Lansing, MI; and 3Department of Electrical and Computer Engineering, Michigan State University, East Lansing, MI.
Medicine and Science in Sports and Exercise
|December 11, 2014
Summary
A thigh-worn accelerometer accurately predicts energy expenditure (EE) using artificial neural networks (ANNs). Wrist and hip accelerometers also offer reliable EE prediction, with thigh placement showing superior accuracy.
Area of Science:
- Biomedical Engineering
- Wearable Technology
- Exercise Physiology
Background:
- Accurate estimation of energy expenditure (EE) is crucial for health monitoring and physical activity assessment.
- Wearable accelerometers offer a non-invasive method for tracking physical activity and estimating EE.
- Developing robust prediction models for EE using accelerometer data is an ongoing area of research.
Purpose of the Study:
- To develop, validate, and compare EE prediction models using accelerometers placed on the hip, thigh, and wrists.
- To investigate the utility of simple accelerometer features as input variables for EE prediction models.
- To determine the optimal accelerometer placement for accurate EE estimation.
Main Methods:
- Forty-four healthy adults participated in a simulated free-living activity protocol.
- Four accelerometers were worn on the hip, thigh, and wrists, with EE measured by a portable metabolic analyzer.
- Artificial neural networks (ANNs) were employed to predict EE from accelerometer data using a leave-one-out cross-validation approach.
Main Results:
- ANN models for all accelerometer placements achieved high accuracy in predicting EE (r > 0.80).
- The thigh accelerometer demonstrated the highest accuracy (r = 0.90) and lowest error.
- Predictive model accuracy differences were more significant with fewer input variables, favoring thigh placement.
Conclusions:
- A single accelerometer on the thigh provides the most accurate EE prediction.
- Accelerometers worn on the wrists or hip also yield high measurement accuracy for EE.
- The findings support the use of accelerometers for reliable energy expenditure monitoring.
