Published December 12, 2024 | Version 1.0.0
Dataset Open

Triaxial accelerometer gait dataset: foot and lower back motion during normal and metronome walking

  • 1. ROR icon HES-SO Arc
  • 2. ROR icon HES-SO University of Applied Sciences and Arts Western Switzerland

Contributors

Data collector:

  • 1. ROR icon HES-SO Arc

Description

This dataset contains accelerometric data collected from young and older individuals walking in a controlled environment. The data were recorded using two triaxial accelerometers, one attached to the participant's lower back and the other attached to the foot. Participants were instructed to walk back and forth along a 205-meter corridor under two different conditions:

Normal walking: Participants walked at their preferred walking speed, reflecting their natural gait and pace.

Metronome walking: Participants synchronized their walking pace to a metronome set to their preferred walking cadence. This condition introduced a rhythmic element to the walking pattern, allowing for the study of gait changes when adhering to an external tempo.

Another condition was also measured to introduce a more variable and dynamic walking pattern that reflects everyday pedestrian movement in a real-world context.

Free outdoor walking: Older participants engaged in approximately 5 minutes of free walking in an urban environment, navigating city streets. During this activity, only the lumbar accelerometer was used to record data. 

Methods (English)

Data collection methodology

  • Measuring device. The dataset was collected using two compact and lightweight triaxial accelerometers (Physilog 6S, Gaitup, Switzerland), each measuring 42.2 x 31.6 x 15 mm and weighing 15 grams. These high-precision instruments have 16-bit resolution, an accuracy range of +/- 8g, and operate at a sampling rate of 256 Hz. For data acquisition, one accelerometer was attached to the instep of the right foot, and the other was positioned in the lumbar region, specifically at the L4-L5 spinal level.
  • Participant information. The dataset contains data from 60 older participants and 42 young participants. The mean age of the older participants is 76 years (standard deviation: 6 years), with an age range of 65 to 88 years. Regarding the sex distribution, 60% of the participants (n=36) are female and 40% (n=24) are male. The mean body mass is 74 kg (SD=16) and the mean height is 1.68 m (SD=0.08). Detailed individual older participant information can be found in the file _data_old.csv. In addition, 42 young adults, including 16 males (38%) and 26 females (62%), participated in the indoor walking test. Their mean age was 27 years (SD = 5.9), with a body mass of 68.9 kg (SD = 17.7) and height of 1.71 m (SD = 0.08). Detailed individual young participant information can be found in the file _data_young.csv.
  • Data collection protocol. Participants completed two different walking tasks in a 205 m corridor and in an urban outdoor environment. They wore they own comfortable, low-rise shoes. First, they walked the corridor for two laps (outbound and return) at their natural pace with each lap timed to determine their preferred walking speed. After a five-minute break, during which walking cadence was calculated from lumbar accelerometer data, they began another corridor walk. This time, they synchronized their gait with an electronic metronome set to their previously determined cadence. The metronome-synchronized walk was preceded by a short 30-second practice walk. Finally, participants completed a 5-minute free walk in an urban environment, wearing only the lumbar accelerometer.
  • Data processing. The raw accelerometric data were initially recorded in a proprietary binary format (.bin files) for both the lumbar and foot accelerometers. This encompassed continuous recordings for the outbound and return laps, as well as the outdoor walking session. These files were then imported into MATLAB using a script provided by the accelerometer manufacturer. Subsequently, under expert supervision, each file was inspected using MATLAB figures to identify the segments of constant walking. Portions of the data corresponding to the beginning, breaks, and ending of each session were carefully selected and removed. The continuous walking segments for the outbound and return laps were then saved separately in .csv format. As a result, the files provided in this dataset represent uninterrupted walking periods. It should be noted that the outdoor walk files may contain brief pauses, reflecting instances where participants had to momentarily stop.

Notes (English)

File Naming Scheme

Each file is in CSV (Comma-Separated Values) format and adheres to the following structure: 

[SubjectID]_[AccelerometerLocation]_[Condition]_[Direction].csv.

The components of the file name are as follows:

  • Subject ID: A three-digit identifier unique to each participant (e.g., 001, 002, ...).
  • Accelerometer Location: Indicates the accelerometer's placement during data collection.
    • FO: Foot accelerometer.
    • LB: Low back (lumbar region) accelerometer.
  • Condition: Specifies the walking condition under which the data were recorded.
    • N: Normal walking at the participant's preferred pace.
    • M: Walking synchronized to a metronome.
    • E: External or outdoor walking, without the corridor's constraints.
  • Direction (applicable only for corridor walking): Indicates the segment of the indoor walking test.
    • O: Outbound (the first half of the corridor walk).
    • R: Return (the second half of the corridor walk).

Examples

  • 001_FO_M_O.csv: Subject 001, data from the foot accelerometer, metronome walking condition, outbound segment of the corridor walk.
  • 002_LB_N_R.csv: Subject 002, data from the low back accelerometer, normal walking condition, return segment of the corridor walk.
  • 003_LB_E.csv: Subject 003, data from the low back accelerometer, external (outdoor) walking condition.

Notes (English)

Data Format

The dataset is presented in CSV (Comma-Separated Values) format, a standard text file format that uses a comma (,) as a delimiter to separate individual values. 

Columns Description

Each CSV file consists of three columns, corresponding to the three axes of the accelerometer used for the measurements. The orientation of these axes differs depending on the location of the accelerometer:

  • Low Back Accelerometer (Lumbar Region):
    • First Column: Anteroposterior Axis - This axis aligns with the forward and backward direction relative to the body's anatomical position.
    • Second Column: Vertical Axis - Represents the up and down movement, perpendicular to the ground.
    • Third Column: Mediolateral Axis - Corresponds to the side-to-side movement, parallel to the ground but perpendicular to the body's forward direction.
  • Foot Accelerometer: The axes for the foot accelerometer are not aligned with specific anatomical directions. Instead, they represent the accelerometer's internal coordinate system, marked as X, Y, and Z.
    • First Column (X-Axis): Represents the accelerometer's first axis, which varies depending on its orientation on the foot.
    • Second Column (Y-Axis): Corresponds to the accelerometer's second axis, orthogonal to the X-axis.
    • Third Column (Z-Axis): Aligns with the accelerometer's third axis, orthogonal to both the X and Y axes.

The data represented in each column consist of raw acceleration measurements expressed in Earth's gravitational units (g), captured at a sampling rate of 256 Hz

Notes (English)

Description of the Participant Information Files (_data_old.csv and _data_young.csv)

The dataset includes two CSV file, named _data_old.csv and _data_young.csv, which contain detailed demographic and experimental data for each participant. The file is structured with each row representing an individual participant and columns detailing specific attributes:

  1. ID: Participant Identification Number - A unique numeric identifier assigned to each participant.
  2. age: Age in years - The participant's age at the time of the study, expressed in full years.
  3. sex: Sex of the participant - Coded as 1 for male and 2 for female.
  4. mass: Body mass in kilograms (kg) - The weight of the participant.
  5. height: Body height in centimeters (cm) - The height of the participant.
  6. speed_N_O: Walking speed under normal condition, outbound lap - This column records the walking speed (measured in meters per second, m/s) during the outbound lap of the corridor walk under normal conditions.
  7. speed_N_R: Walking speed under normal condition, return lap - Similar to the previous, this column documents the walking speed during the return lap of the corridor walk under normal condition.
  8. speed_M_O: Walking speed under metronome condition, outbound lap - Here, the walking speed is recorded for the outbound lap under the metronome walking condition.
  9. speed_M_R: Walking speed under metronome condition, return Lap - This column contains the walking speed data for the return lap during the metronome walking condition.

Any missing data in the columns are indicated as 'NaN' (Not a Number), 

Notes (English)

Ethical considerations

The  protocol for this study was submitted to the ethics committee ("commission cantonale d'éthique de la recherche sur l'être humain du canton de Vaud", project ID 2021-01937) in October 2021 and received approval in November 2021. The procedures of the study were conducted in strict accordance with the Declaration of Helsinki and in compliance with the Swiss Human Research Act. All participants were informed of the objectives and procedures of the study and provided signed informed consent. They agreed that their data would be anonymized and made publicly available to ensure both the transparency of the research and their privacy.

Files

_data_old.zip

Files (809.8 MB)

Name Size Download all
md5:8c7b04ef924fc385e9fba737dbf3f78d
536.1 MB Preview Download
md5:bf83f8e8524971722aca68bf20ce98b0
273.7 MB Preview Download

Additional details

Related works

References
Conference paper: 10.1016/j.gaitpost.2022.07.102 (DOI)
Conference paper: 10.1016/j.gaitpost.2022.07.106 (DOI)

Dates

Created
2023-11-20
Upload of the first version
Updated
2023-11-22
Add two references
Updated
2024-12-12
Add younger participants