Hip joint centre position estimation using a dual unscented Kalman filter for computer-assisted orthopaedic surgery

Area of Science:

• Biomedical Engineering
• Orthopedic Surgery
• Robotics

Background:

• Accurate femur center of rotation localization is crucial for computer-assisted knee surgery.
• Untracked pelvic movements during surgery can compromise the accuracy of hip joint center estimation.
• Existing methods rely solely on femoral pose, neglecting pelvic dynamics.

Purpose of the Study:

• To develop and validate a dual unscented Kalman filter algorithm for improved hip joint center estimation.
• To incorporate pelvic reference point data to compensate for un-tracked pelvic movements.
• To assess the algorithm's accuracy and robustness compared to existing methods under simulated surgical conditions.

Main Methods:

• Development of a dual unscented Kalman filter algorithm.
• Integration of a pelvic reference point acquired via a skin marker.
• In vitro experiments simulating surgical conditions, including soft tissue artifacts and femoral pivoting (planar and multi-planar movements).
• Comparative analysis of the proposed algorithm against other methods.

Main Results:

• The dual unscented Kalman filter demonstrated reduced sensitivity to pelvic displacements.
• Comparable accuracy was achieved for small hip joint center displacements (<6 mm) during planar movements.
• The algorithm showed a 30% improvement in accuracy for displacements exceeding 6 mm during multi-planar movements.

Conclusions:

• The dual unscented Kalman filter algorithm offers a non-invasive and accurate solution for hip joint center localization in computer-assisted knee surgery.
• The method effectively mitigates errors caused by pelvic motion, particularly in challenging multi-planar scenarios.
• This advancement has the potential to enhance the precision and reliability of orthopedic surgical navigation systems.