Hierarchical information fusion for global displacement estimation in microsensor motion capture

Area of Science:

• Biomechanics
• Sensor Fusion
• Human Motion Analysis

Background:

• Accurate estimation of human global displacement is crucial for understanding locomotion.
• Existing methods often struggle with drift and accuracy across various gait patterns.
• Body-worn inertial and magnetic measurement units offer a portable solution for motion tracking.

Purpose of the Study:

• To develop and validate a novel hierarchical information fusion algorithm for precise human global displacement estimation.
• To address orientation and displacement errors in sensor-based locomotion analysis.
• To enable accurate tracking of diverse gait patterns like walking, running, and hopping.

Main Methods:

• A two-level hierarchical fusion approach employing complementary Kalman filters (CKF).
• First-level sensor fusion estimates segment orientation and foot displacement with zero velocity updates.
• Second-level geometric fusion integrates estimates from both lower limbs for enhanced accuracy.

Main Results:

• The proposed algorithm accurately estimates human global displacement for walking, running, and hopping.
• The hierarchical fusion effectively compensates for sensor errors and reduces drift.
• Results show high agreement with optical motion tracking systems.

Conclusions:

• The novel hierarchical information fusion algorithm provides a robust and accurate method for human locomotion analysis.
• This approach offers improved global displacement estimation compared to traditional methods.
• The system demonstrates effectiveness across multiple gait patterns using body-worn sensors.