Related Experiment Video
Updated: Mar 20, 2026
06:49
Evaluation of a Smartphone-based Human Activity Recognition System in a Daily Living Environment
Published on: December 11, 2015
9.4K
InclusiveHAR: A smartphone-based dataset for human activity recognition across diverse physical abilities
Seyed Reza Kamel Tabbakh1, Iman Naeimi1, Kosar Naghavi1
1Department of Computer Engineering, Ma.C., Islamic Azad University, Mashhad, Iran.
Data in Brief
|March 19, 2026
Summary
This study introduces InclusiveHAR, a new smartphone dataset for Human Activity Recognition (HAR) that includes individuals with and without disabilities. This inclusive dataset aims to improve AI model generalizability for diverse populations in healthcare applications.
Area of Science:
- Computer Science
- Biomedical Engineering
- Rehabilitation Science
Background:
- Human Activity Recognition (HAR) is crucial for healthcare, rehabilitation, and smart environments.
- Existing HAR datasets lack diversity, primarily featuring non-disabled individuals, limiting real-world applicability.
- This gap hinders the development of inclusive AI systems for diverse user groups.
Purpose of the Study:
- To introduce InclusiveHAR, a novel, smartphone-based Human Activity Recognition dataset.
- To address the limitations of current HAR datasets by including participants with diverse abilities.
- To facilitate the development of more robust and inclusive HAR systems.
Main Methods:
- Collected data from 20 participants (10 non-disabled, 10 with disabilities) performing six daily activities.
- Utilized an iPhone 14 Pro at a 50 Hz sampling rate, recorded via the SensorLog app.
- Provided baseline evaluations using MLP, K-NN, SVM, and XGBoost models.
Main Results:
- The InclusiveHAR dataset captures diverse movement patterns, highlighting variations in activity execution among individuals with disabilities.
- Baseline evaluations demonstrated the dataset's utility for training and comparing machine learning models.
- Performance metrics were evaluated across different training scenarios, showcasing model adaptability.
Conclusions:
- InclusiveHAR provides a valuable resource for advancing Human Activity Recognition research across diverse populations.
- The dataset supports the development of inclusive HAR systems for healthcare and assistive technologies.
- Detailed documentation ensures transparency, reproducibility, and facilitates future comparative studies.
