Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Estimation of Resistance Training RPE using Inertial Sensors and Electromyography

James Thomas, Johan Wahlström

TL;DR

This study addresses estimating resistance-training RPE from wearable sensors by leveraging inertial measurements and EMG-derived training labels. The approach releases a novel EMG-IMU dataset (69 sets, 1003 reps) and demonstrates that a Random Forest classifier using EMG-informed labels achieves the strongest RPE estimation performance, with an exact accuracy of 41.4% and an ±1 accuracy of 85.9%. EMG features provide modest, yet consistent improvements over IMU-only models, while RNNs underperform in this setup due to data constraints and augmentation effects. The work highlights key predictors, notably eccentric duration, and discusses challenges in data quality, sensor placement, and generalizability, pointing to larger, more diverse datasets and alternative EMG integration strategies for future improvements.

Abstract

Accurate estimation of rating of perceived exertion (RPE) can enhance resistance training through personalized feedback and injury prevention. This study investigates the application of machine learning models to estimate RPE during single-arm dumbbell bicep curls, using data from wearable inertial and electromyography (EMG) sensors. A custom dataset of 69 sets and over 1000 repetitions was collected, with statistical features extracted for model training. Among the models evaluated, a random forest classifier achieved the highest performance, with 41.4% exact accuracy and 85.9% $\pm1$ RPE accuracy. While the inclusion of EMG data slightly improved model accuracy over inertial sensors alone, its utility may have been limited by factors such as data quality and placement sensitivity. Feature analysis highlighted eccentric repetition time as the strongest RPE predictor. The results demonstrate the feasibility of wearable-sensor-based RPE estimation and identify key challenges for improving model generalizability.

Estimation of Resistance Training RPE using Inertial Sensors and Electromyography

TL;DR

This study addresses estimating resistance-training RPE from wearable sensors by leveraging inertial measurements and EMG-derived training labels. The approach releases a novel EMG-IMU dataset (69 sets, 1003 reps) and demonstrates that a Random Forest classifier using EMG-informed labels achieves the strongest RPE estimation performance, with an exact accuracy of 41.4% and an ±1 accuracy of 85.9%. EMG features provide modest, yet consistent improvements over IMU-only models, while RNNs underperform in this setup due to data constraints and augmentation effects. The work highlights key predictors, notably eccentric duration, and discusses challenges in data quality, sensor placement, and generalizability, pointing to larger, more diverse datasets and alternative EMG integration strategies for future improvements.

Abstract

Accurate estimation of rating of perceived exertion (RPE) can enhance resistance training through personalized feedback and injury prevention. This study investigates the application of machine learning models to estimate RPE during single-arm dumbbell bicep curls, using data from wearable inertial and electromyography (EMG) sensors. A custom dataset of 69 sets and over 1000 repetitions was collected, with statistical features extracted for model training. Among the models evaluated, a random forest classifier achieved the highest performance, with 41.4% exact accuracy and 85.9% RPE accuracy. While the inclusion of EMG data slightly improved model accuracy over inertial sensors alone, its utility may have been limited by factors such as data quality and placement sensitivity. Feature analysis highlighted eccentric repetition time as the strongest RPE predictor. The results demonstrate the feasibility of wearable-sensor-based RPE estimation and identify key challenges for improving model generalizability.

Paper Structure

This paper contains 28 sections, 11 figures, 9 tables.

Table of Contents

  1. Introduction
  2. Data Collection
  3. Methods
  4. Preprocessing
  5. Sampling Rate Matching
  6. Repetition Segmentation
  7. IMU Feature Extraction
  8. EMG Feature Extraction
  9. EMG Labeling
  10. Dimensionality Reduction
  11. Clustering Methods
  12. EMG Estimation Models
  13. RPE Estimation Models
  14. Data Processing Pipeline
  15. Model Evaluation Methodology
  16. ...and 13 more sections

Figures (11)

  • Figure 1: Delsys sensor units attached to a participant.
  • Figure 2: Repetition count per participant.
  • Figure 3: Distribution of collected RPE values.
  • Figure 4: Data processing pipeline.
  • Figure 5: Example of a set marked with repetition boundaries and midpoints.
  • ...and 6 more figures