Dataset Overview

This dataset contains the biomechanical data, electromyography (EMG) data, and exoskeleton parameters collected during the evaluation of lower-limb exoskeletons across various ambulation modes.

Dataset was collected from 19 healthy subjects walking using a lower limb exoskeleton (X2, Fourier Intelligence) in five locomotion activities (Overground Walking, Ramps ascending and descending, Stairs ascending and descending). The institutional review board of Northwestern University (STU00212684) and the Ethical Board of the Hospital Universitario Severo Ochoa, (Madrid, Spain) approved this study, and all procedures were in accordance with the Declaration of Helsinki. The study was performed in parallel at two sites with similar infrastructure: The experimental setup in Chicago comprised six meters of overground walking, a staircase with six stairs of 150 mm height, and a 5 deg ramp of 4 meters in length. The experimental setup in Madrid comprised of 13 meters of overground walking, a staircase with six stairs of 170 mm height, and a 6 deg ramp of 3 meters in length.

Each participant's data is stored in a separate file, containing all activities performed in chronological order (with rest times removed). EMG information is available for five subjects (collected at Hospital Los Madroños) and is stored in separate files with corresponding subject identifiers.

Files: 

• AB68_ExoGait-Benchmark.csv
• AB69_ExoGait-Benchmark.csv
• AB70_ExoGait-Benchmark.csv
• AB71_ExoGait-Benchmark.csv
• AB72_ExoGait-Benchmark.csv
• AB73_ExoGait-Benchmark.csv
• AB75_ExoGait-Benchmark.csv
• AB76_ExoGait-Benchmark.csv
• AB77_ExoGait-Benchmark.csv
• AB78_ExoGait-Benchmark.csv
• AB79_ExoGait-Benchmark.csv
• AB80_ExoGait-Benchmark.csv
• AB81_ExoGait-Benchmark.csv
• AB82_ExoGait-Benchmark.csv
• HL01_ExoGait-Benchmark.csv
• HL02_ExoGait-Benchmark.csv
• HL03_ExoGait-Benchmark.csv
• HL04_ExoGait-Benchmark.csv
• HL05_ExoGait-Benchmark.csv

Abbreviations and Terminology

The following acronyms and abbreviations are used in the dataset:

Conditions:

• SM: State-Machine Controller
• TR: Transparent Mode

Ambulation Modes:

• W: Overground Walking
• RA: Ramp Ascent
• RD: Ramp Descent
• SA: Stair Ascent
• SD: Stair Descent

Data Dictionary: Exoskeleton Data

Note: Due to the number of sensors, related variables are aggregated below. Where directional or multi-joint data is collected, it includes left/right (l/r) metrics for the respective joints. The data was recorded and stored at 333 Hz.

Data Dictionary: EMG Data

Note: EMG data is stored in separate files. The muscles recorded include the Rectus Femoris, Vastus Lateralis, Biceps Femoris, Tibialis Anterior, and Medial Gastrocnemius.

Code and Usage

Included with the dataset are sample Python scripts to guide readers on how to load and use the data, available at the project's GitHub repository.

Two primary Jupyter Notebooks are provided for exploration:

1. Explore_specific_user_dataset.ipynb: Introduces the dataset structure, collected components, and allows for visualization of user-specific features for each stride.
2. Explore_entire_dataset.ipynb: Presents the overarching dataset structure and plots the metrics used in the study for all subjects, enabling the replication of the published plots.

For questions and additional informations contact the authors Lhoste and Vianello: clhost@ethz.ch; lvianello@sralab.org.

Human subjects data

All human subjects data included in this dataset have been fully de-identified in accordance with applicable legal and ethical guidelines. Explicit informed consent was obtained from all participants for the collection and public sharing of their de-identified data. De-identification procedures included the removal of all direct identifiers (e.g., names, contact information) and the aggregation or masking of indirect identifiers to minimize the risk of re-identification. No personally identifiable information is contained in this dataset.