Optimising filtering parameters for a 3D motion analysis system.

Area of Science:

• Biomechanics
• Motion Analysis
• Signal Processing

Background:

• Low-pass filters are commonly used to reduce noise in movement data analysis.
• The selection of cut-off frequencies for these filters is often arbitrary, potentially impacting data accuracy.

Purpose of the Study:

• To evaluate a novel method for determining the optimal cut-off frequency for filtering kinematic data.
• To utilize rigid marker clusters for assessing the dynamic precision of 3D motion analysis systems.

Main Methods:

• A model-based approach was employed, focusing on swimming kinematics where measurement noise is significant.
• Dynamic precision derived from marker clusters was used as the criterion for selecting the optimal cut-off frequency.
• The method was validated using both model data and experimental data from a video-based motion analysis system.

Main Results:

• Filtering kinematic data with a fixed 6Hz cut-off frequency sometimes reduced reconstruction accuracy compared to unfiltered data.
• Employing the cut-off frequency that maximized dynamic precision improved kinematic reconstruction accuracy by 29% compared to raw data.
• The optimal cut-off frequency varied based on marker cluster position and individual subjects, consistent across model and experimental data.

Conclusions:

• 3D motion analysis systems can achieve higher accuracy by optimizing filter cut-off frequencies based on dynamic precision.
• The proposed dynamic precision method is a valuable tool for evaluating the impact of different filtering techniques on kinematic data.