Robust Inertial Measurement Unit-Based Attitude Determination Kalman Filter for Kinematically Constrained Links

Area of Science:

• Robotics and Control Systems
• Navigation and Guidance
• Sensor Fusion

Background:

• External acceleration introduces significant uncertainty in attitude determination using Inertial Measurement Unit (IMU) signals, degrading accuracy.
• Existing acceleration-compensating methods offer incomplete solutions to this underdetermined problem.

Purpose of the Study:

• To develop a novel constraint-augmented Kalman filter (KF) for robust IMU-based attitude determination.
• To eliminate acceleration-induced uncertainty in attitude estimation for IMUs attached to constrained links.

Main Methods:

• Implementation of an acceleration-level kinematic constraint derived from a ball joint mechanism.
• Development of a constraint-augmented Kalman filter (KF) to integrate this constraint.

Main Results:

• The proposed constrained KF demonstrated superior performance over the conventional unconstrained KF.
• Average accuracy improvement of 1.88°, with a maximum improvement of 4.18° observed.
• The constrained KF's accuracy was independent of test acceleration conditions, unlike the conventional KF.

Conclusions:

• The novel constraint-augmented KF effectively eliminates acceleration-induced uncertainty for robust IMU-based attitude determination.
• This method offers significant accuracy improvements and robustness, making it suitable for applications requiring precise attitude estimation under dynamic conditions.