# Thigh-Worn Accelerometry and Gait Event Detection: Axivity AX6 with loadsol Force Insole Ground Truth in 21 Adults (Lendt, Regensburg)

## Facts

- Provider: Claas Lendt, Martin Grimmer, Ingo Froböse, Tom Stewart (University of Regensburg, German Sport University Cologne); harmonized by Josef Heidler
- License: CC-BY-4.0 (https://zenodo.org/api/records/17629131)
- Size: 24 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/josefheidler/har_gait_adults_2023-lendt
- Access: public, readable without an account
- DOI: https://doi.org/10.5281/zenodo.17629131
- Category: Physiology
- Source: https://doi.org/10.1101/2025.11.18.25339671
- Page: https://scigantic.com/archive/8105089a-5c8b-5516-934f-98ba83ac9081
- Keywords: Physiology, Parquet, Hugging Face

## About

Thigh-worn accelerometry with insole-derived gait events, dataset C of the TWAGA gait pipeline by Claas Lendt and colleagues (University of Regensburg and German Sport University Cologne). Twenty-one healthy adults wore an Axivity AX6 accelerometer on the lateral mid-thigh of the dominant leg (100 Hz) together with a loadsol pro force-sensing insole on the same leg during four 2-minute treadmill walking trials at 2, 3, 4 and 5 km/h and one 5-minute outdoor walk at preferred speed. Every sample carries the insole vertical ground reaction force and initial-contact (heel strike) and final-contact (toe-off) flags, the reference for detecting steps, stride time and cadence from a single thigh sensor. The accelerometer signal was low-pass filtered near 20 Hz upstream. There are about 1.61 million samples and 4.5 labelled hours, plus participants.parquet giving each participant's dominant leg. This Hugging Face copy harmonizes the Zenodo release (DOI 10.5281/zenodo.17629131) into one Parquet file per participant under harmonized/; load with Hugging Face datasets or pandas. The Zenodo record is licensed CC BY 4.0. Cite Lendt, Grimmer, Froböse and Stewart, medRxiv 2025.11.18.25339671 (preprint).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("josefheidler/har_gait_adults_2023-lendt")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `harmonized/alert-aracari.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/8105089a-5c8b-5516-934f-98ba83ac9081
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/8105089a-5c8b-5516-934f-98ba83ac9081
- Over MCP: `get_schema_card` with id `8105089a-5c8b-5516-934f-98ba83ac9081` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Physiology datasets](https://scigantic.com/topics/physiology)
