What question did this study set out to answer?

This research aims to improve gait-phase detection accuracy during rehabilitation and mobility assessment using machine learning.

February 2, 2026Applied SciencesOpen Access

Machine Learning Models for Reliable Gait Phase Detection Using Lower-Limb Wearable Sensor Data

View Full Paper

Ask AI

Bookmark

Authors

Demonstrates improved gait phase detection using machine learning on sensor data, suggesting enhanced rehabilitation applications.

This research aims to improve gait-phase detection accuracy during rehabilitation and mobility assessment using machine learning.
Utilized a multi-subject IMU dataset collected from lower-limb wearable sensors.
Segmented IMU recordings into 128-sample windows and flattened them into 2944-dimensional feature vectors.
Developed an automatic relabeling pipeline for accurate ground-truth labels with heel-strike and toe-off detection.
Trained multiple machine learning models, including Random Forests and XGBoost, with systematic hyperparameter tuning.
Assessed model performance using k-fold cross-validation.
Achieved gait-phase classification accuracy exceeding 97% across test sets.
Tree-based ensemble models demonstrated high reliability and performance in gait analysis.

Fiaz et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark