Inertial measurement unit-based pose estimation: Analyzing and reducing sensitivity to sensor placement and body...

Area of Science:

• Biomechanics
• Clinical Biomechanics
• Rehabilitation Engineering

Background:

• Inertial measurement units (IMUs) offer potential for automated pose estimation and exercise monitoring in clinical settings.
• Extensive calibration procedures in existing methods limit their clinical applicability.
• Lower body pose estimation is crucial for assessing mobility during clinical tests.

Purpose of the Study:

• To adapt an IMU-based pose estimation method for lower body mobility tests.
• To investigate the sensitivity of pose estimation accuracy to kinematic and sensor parameters.
• To develop a simplified calibration protocol for improved clinical application.

Main Methods:

• Utilized an extended Kalman filter and kinematic chain modeling for lower body pose estimation.
• Performed sensitivity analysis on kinematic model and sensor placement parameters.
• Developed and applied a simple calibration protocol focusing on IMU orientation.

Main Results:

• Sensitivity analysis identified IMU orientation as critical for accuracy.
• The proposed simple calibration protocol significantly improved joint angle accuracy.
• Ankle, knee, and hip joint angle errors were reduced without additional calibration.

Conclusions:

• Pose estimation accuracy using body-worn IMUs depends on a limited set of parameters.
• A straightforward calibration for IMU orientation yields effective pose estimation.
• This approach enhances the clinical feasibility of IMU-based pose estimation.