Quaternion-Based Local Frame Alignment between an Inertial Measurement Unit and a Motion Capture System

Area of Science:

• Biomechanics
• Robotics
• Sensor Fusion

Background:

• Accurate local frame alignment between Inertial Measurement Unit (IMU) systems and Optical Motion Capture Systems (MCS) is crucial for motion analysis and IMU data validation.
• Existing alignment methods may have limitations in accuracy and ease of setup.

Purpose of the Study:

• To propose and evaluate a novel quaternion-based local frame alignment method for IMU and MCS integration.
• To compare the proposed method against existing techniques using various data conditions and investigate key influencing factors.

Main Methods:

• Developed a quaternion-based alignment method utilizing angular velocity transformation equations.
• Compared performance against three other methods, analyzing effects of transformation concept (angular velocity vs. angle), orientation representation (quaternion vs. Direction Cosine Matrix - DCM), and solvers (nonlinear least squares vs. pseudoinverse least squares).

Main Results:

• The proposed quaternion-based angular velocity transformation method demonstrated superior performance.
• Angular velocity transformation proved more effective than angle transformation.
• Quaternions were found to be more suitable than DCMs for this application.
• Solver choice had minimal impact on general estimation performance.

Conclusions:

• The novel quaternion-based method offers high accuracy and easy setup for IMU-MCS local frame alignment.
• Its effectiveness stems from the quaternion's fewer components and constraints compared to DCMs.
• This method significantly benefits motion analysis and IMU system validation.