Generalized q-Method Relative Pose Estimation for UAVs with Onboard Sensor Measurements.

Area of Science:

• Aerospace Engineering
• Robotics
• Computer Vision

Background:

• Accurate pose estimation (position and orientation) is critical for autonomous air vehicle navigation.
• Existing methods often rely on complex sensor fusion or make simplifying assumptions.
• Davenport's q-method is established for satellite attitude estimation but requires adaptation for broader applications.

Purpose of the Study:

• To develop and implement a generalized q-method for robust pose estimation in air vehicles.
• To extend existing q-methodology by incorporating inertial position and removing spacecraft-specific constraints.
• To validate the method's accuracy for Unmanned Aerial Vehicle (UAV) applications.

Main Methods:

• Utilized on-board measurement vectors of reference objects for pose calculation relative to an inertial frame.
• Solved for the quaternion eigenvalue solution to minimize error in the derived system of equations.
• Developed and implemented the pose estimation model for UAVs using onboard camera measurements in a controlled environment.

Main Results:

• The generalized q-method successfully estimated air vehicle position and orientation.
• Algorithm outputs were validated against truth data, demonstrating high accuracy.
• The method proved effective even in the presence of sensor errors.

Conclusions:

• The generalized q-method provides an accurate and robust solution for air vehicle pose estimation.
• This approach extends the applicability of q-methodology beyond satellite attitude determination.
• The developed model is suitable for UAV applications requiring precise navigation.