Real-time gait event detection for normal subjects from lower trunk accelerations.

Area of Science:

• Biomechanics
• Gait Analysis
• Wearable Sensors

Background:

• Accurate detection of gait events like initial contact (IC) and final contact (FC) is crucial for clinical assessments.
• Previous methods often rely on force platforms or have limitations in real-time processing of trunk-based sensor data.

Purpose of the Study:

• To develop and validate a novel algorithm for real-time detection and timing of IC and FC gait events.
• To process vertical and antero-posterior accelerations from the lower trunk (L3 vertebra).

Main Methods:

• Algorithm development based on heuristic rules derived from 1719 gait steps.
• Processing of vertical and antero-posterior acceleration data from the lower trunk.
• Validation against a Digimax force platform.

Main Results:

• The algorithm demonstrated small deviations from force platform data: 13+/-35 ms for IC and 9+/-54 ms for FC.
• Real-time detection delays were 117+/-39 ms for IC and 34+/-72 ms for FC.
• Novelty in precise temporal location of FC using lower trunk accelerations.

Conclusions:

• The developed algorithm provides accurate and real-time detection of IC and FC gait events.
• This method offers an improvement over existing techniques for gait event analysis using lower trunk accelerations.
• The findings have implications for non-invasive, real-time gait monitoring and analysis.