Reliability of the step phase detection using inertial measurement units: pilot study

Area of Science:

• Biomechanics
• Human Locomotion Analysis
• Wearable Sensor Technology

Background:

• Gait event detection is crucial for evaluating human locomotion.
• Previous research highlights gyroscope reliability over accelerometers/magnetometers for normal walking, but lacks clinical parameter linkage and optimal sensor placement verification.
• Existing studies haven't fully explored sensor performance across diverse walking patterns, surfaces, and movement constraints.

Purpose of the Study:

• To compare the gait event detection quality of inertial sensors against video ground truth, using Gait and Clinical Movement Analysis Society parameters.
• To evaluate sensor performance on the foot versus the shank.
• To determine optimal sensor placement under various walking conditions (normal, abnormal, different surfaces, movement constraints).

Main Methods:

• Utilized inertial sensors (gyroscopes, accelerometers, magnetometers) for gait event detection.
• Employed video analysis as the ground truth for comparison.
• Assessed sensor performance based on parameters from the Gait and Clinical Movement Analysis Society.
• Compared sensor placement on the foot and shank.
• Tested under normal walking, gait deviations, varied ground surfaces, and with/without movement constraints.

Main Results:

• The proposed methodology enables reliable gait event detection, even with abnormal footfalls and on slippery surfaces.
• Preliminary findings indicate that the shank is the optimal location for sensor placement.
• Gyroscopes demonstrated higher reliability for event detection compared to accelerometers and magnetometers in normal walking scenarios.

Conclusions:

• Inertial sensors, particularly gyroscopes placed on the shank, offer a reliable method for gait event detection in long-distance walking.
• The study provides evidence for optimal sensor placement and validates the methodology across diverse and challenging walking conditions.
• This approach supports accurate temporal parameter assessment in clinical gait analysis.