UKF Magnetometer-Free Sensor Fusion for Pelvis Pose Estimation During Treadmill Walking

Area of Science:

• Biomechanics
• Robotics
• Sensor Fusion

Background:

• Inertial Measurement Units (IMUs) are crucial for estimating rigid body orientation.
• Wearable IMUs are widely used in clinical settings for motion analysis.
• Sensor fusion techniques are essential for accurate pose estimation.

Purpose of the Study:

• To evaluate the feasibility of a nonlinear Unscented Kalman Filter (UKF) for gyroscope/accelerometer sensor fusion.
• To estimate pelvis pose during treadmill walking without using IMU magnetometer data.
• To provide a robust pose estimation method suitable for magnetically disturbed environments.

Main Methods:

• Implemented a magnetometer-free Unscented Kalman Filter (UKF) for sensor fusion.
• Utilized gyroscope and accelerometer data from IMUs for orientation estimation.
• Validated the proposed filter against a gold standard optometric system for pelvis pose (yaw, roll, pitch).

Main Results:

• The proposed UKF approach achieved root mean square errors below 1 degree for pelvis heading, bank, and attitude angles.
• Orientation estimation showed high correlation (> 0.90) with the gold standard optometric system.
• The magnetometer-free method proved effective even in the absence of magnetic sensor data.

Conclusions:

• The magnetometer-free UKF approach is suitable for pelvis pose estimation during human treadmill walking.
• This method offers accurate orientation data without reliance on magnetometers, enhancing applicability in clinical environments.
• The findings support further estimation of center of mass displacement using this technique.