Multimodal inverse kinematics significantly improves IMU-based biomechanical analyses

Area of Science:

• Biomechanics and Human Movement Analysis
• Computational Simulation and Modeling
• Sensor Fusion and Data Integration

Background:

• Inertial measurement units (IMUs) are widely used in musculoskeletal simulations for motion analysis.
• IMU-based methods often suffer from inaccuracies like joint angle drift and calibration errors, affecting kinematic and dynamic results.
• Existing IMU approaches struggle to provide precise outcomes due to inherent sensor limitations.

Purpose of the Study:

• To investigate the potential of integrating spatial reference information into IMU-driven inverse kinematics analyses.
• To systematically assess how spatial data can compensate for typical IMU errors in musculoskeletal simulations.
• To evaluate the accuracy improvements offered by a multimodal approach combining IMU and positional data.

Main Methods:

• Developed a simulation-based framework using synthetic inertial and positional data.
• Generated error-free kinematic and dynamic data from optical motion capture as a reference.
• Introduced various IMU error types (noise, drift, misalignment) into synthetic orientation and position data for systematic analysis.

Main Results:

• The multimodal inverse approach, integrating both IMU and positional data, significantly outperformed solely IMU-based analyses.
• Demonstrated substantial reductions in Root Mean Square Error (RMSE) for joint angles, joint torques, residual forces, and residual torques.
• Quantified the sensitivity of the multimodal approach to the spatial accuracy of the positional data.

Conclusions:

• Integrating spatial reference information into IMU-driven inverse kinematics is a viable strategy to enhance musculoskeletal simulation accuracy.
• The multimodal approach effectively mitigates common IMU errors, leading to more reliable kinematic and dynamic outcomes.
• Further validation with real-world measurement data is warranted to confirm the practical efficacy of this method.