Keeping a Good Attitude: A Quaternion-Based Orientation Filter for IMUs and MARGs

Area of Science:

• Robotics and Control Systems
• Sensor Fusion and Estimation Theory
• Aerospace Engineering

Background:

• Accurate orientation estimation is critical for Micro Aerial Vehicles (MAVs) attitude control.
• Low-cost MicroElectroMechanical System (MEMS) inertial sensors present challenges due to low accuracy and noisy data.
• Existing methods struggle with magnetic disturbances affecting orientation estimation.

Purpose of the Study:

• To develop a novel quaternion-based orientation estimation method for MAVs.
• To address the impact of magnetic disturbances on roll and pitch orientation.
• To propose a new complementary filter fusing gyroscope, accelerometer, and magnetometer data.

Main Methods:

• Algebraic solution for orientation estimation using gravity and magnetic field observations.
• Separation of 'tilt' and 'heading' quaternion estimation to mitigate magnetic interference.
• Development of a novel complementary filter integrating gyroscope, accelerometer, and magnetometer data for IMU and MARG sensors.

Main Results:

• Analytical and empirical validation of the proposed orientation estimation method.
• Demonstrated significant performance improvement over common orientation estimation techniques.
• Successful evaluation using real-world flight data from a micro quadrotor helicopter.

Conclusions:

• The proposed method effectively estimates MAV orientation, even with low-cost sensors.
• The novel complementary filter provides superior performance compared to existing methods.
• This approach enhances MAV attitude control by providing reliable orientation feedback.