A Multi-Modal Gait Database of Natural Everyday-Walk in an Urban Environment

Area of Science:

• Biomechanics
• Human Motion Analysis
• Wearable Sensor Technology

Background:

• Traditional human gait analysis often occurs in controlled lab settings, limiting real-world applicability.
• Existing datasets may not capture the complexity of everyday walking environments and transitional movements.

Purpose of the Study:

• To present a comprehensive dataset of human gait in natural urban environments.
• To enable machine-learning-based analysis of everyday walking scenarios.
• To investigate the interaction between gait and gaze during natural locomotion.

Main Methods:

• Collected synchronized data from 20 healthy participants using 17 inertial measurement unit (IMU) sensors, 8 pressure sensing cells per foot (FSR), and a mobile eye tracker.
• Participants completed diverse walking courses in an urban environment, including ramps, stairs, and pavements.
• Detailed data annotation was performed to facilitate machine learning applications.

Main Results:

• A rich, multi-modal dataset of human gait in naturalistic settings was successfully acquired.
• The dataset includes synchronized IMU, FSR, and gaze data, offering a holistic view of walking behavior.
• Detailed annotations support advanced data analysis and predictive modeling.

Conclusions:

• The presented dataset provides a valuable foundation for research on natural human locomotion.
• It facilitates the study of transitional motions and the interplay between gait and gaze in everyday scenarios.
• This resource is expected to advance the development of intelligent systems for gait analysis and human-computer interaction.