The NACOB multi-surface walking dataset.

Area of Science:

• Biomechanics
• Human Movement Analysis
• Wearable Sensor Technology

Background:

• Gait research often uses controlled lab environments, limiting real-world applicability.
• Existing irregular surface datasets are frequently small or lack detailed biomechanical data.
• Understanding gait adaptations on uneven terrain is critical for mobility and injury prevention.

Purpose of the Study:

• To introduce a comprehensive dataset of human gait on irregular surfaces.
• To provide a valuable resource for studying biomechanical adaptations during walking.
• To enable the development of advanced gait analysis algorithms.

Main Methods:

• Collected gait data from 134 participants walking on surfaces with varying irregularity.
• Utilized inertial measurement unit (IMU) sensors on the trunk and lower right limb (foot, shank, thigh).
• Developed and validated a machine learning model to classify walking surfaces based on gait data.

Main Results:

• The dataset captures detailed biomechanical gait data from diverse irregular surfaces.
• A machine learning model achieved 95.8% accuracy in predicting the walking surface.
• The results demonstrate the distinct biomechanical signatures associated with different irregular terrains.

Conclusions:

• The new dataset offers a robust resource for biomechanics and gait analysis research.
• Biomechanical gait data effectively differentiates between various irregular walking surfaces.
• This work facilitates advancements in understanding and analyzing human locomotion on challenging terrains.