Wearable Sensor Data for Physical Therapy Exercises
Catalog Summary
"Five participants, eight physical-therapy exercises and three execution styles; 275,603 synchronized time points across five wearable sensor units."
Catalog Notes
Overview
Five participants performed eight exercises in correct, fast and low-amplitude styles while wearing five MTx units. Each unit provides accelerometer, gyroscope and magnetometer measurements on three axes at 25 Hz.
Measured structure
The archive contains 200 template_session.txt files, 200 test.txt files and 40 template_times.txt files. Measurement files are semicolon-delimited, with a header and 10 columns: time index plus nine channels for one sensor. The u1–u5 directories identify sensor units; s1–s5 identify participants.
Counting and labels
The five unit files in each session have matching lengths. Counting each synchronized instant once gives 55,325 template and 220,278 test time points, totaling 275,603. Simply stacking all units produces 1,378,015 sensor rows and repeats those instants. UCI’s headline count of 268,881 does not match this archive count. Catalog row_count uses synchronized time points. Selected template intervals are supplied separately; measurement rows do not contain a per-row exercise-quality target.
Source and provenance
Dataset DOI: https://doi.org/10.24432/C5JK60. Related physical-therapy methods paper: https://doi.org/10.1016/j.cmpb.2014.07.003.
View Data Structure
To explore column names, data types, and sample rows, visit the official dataset page on UCI Machine Learning Repository.
Preview on UCI Machine Learning RepositoryCite This Dataset
Yurtman, A., & Barshan, B. (2014). Physical Therapy Exercises [Data set]. UCI Machine Learning Repository. https://doi.org/10.24432/C5JK60
Source metadata: UCI Machine Learning Repository (2014) · DOI: 10.24432/C5JK60
Indexed by IoTDataset.com on Sep 14, 2026
Review the Source Record
Confirm the licence, version, access conditions, file format, and provenance at the source before use.