A quaternion-based indirect Gaussian particle filter for nonlinear attitude estimation.

Area of Science:

• Robotics
• Control Systems
• Aerospace Engineering

Background:

• Accurate attitude estimation is crucial for autonomous systems.
• Traditional methods struggle with nonlinearities and computational demands.
• Unit quaternions offer a singularity-free representation for attitude.

Purpose of the Study:

• To develop a novel nonlinear attitude estimation algorithm.
• To improve the efficiency and accuracy of quaternion-based estimation.
• To address limitations of existing particle filter (PF) methods.

Main Methods:

• A recursively indirect quaternion estimator utilizing a Gaussian particle filter (GPF).
• Estimation of an on-tangent-plane Gaussian distribution for quaternion manifold uncertainty.
• Global nonsingular attitude description in the prediction step and 3D attitude error estimation in the update step.

Main Results:

• The proposed GPF-based filter eliminates the need for resampling and regularization required by standard PFs.
• Experimental and theoretical validation demonstrates superior performance.
• Achieved faster convergence speed, lower complexity, and reduced computational cost.

Conclusions:

• The developed GPF-based quaternion estimator provides an efficient and accurate solution for nonlinear attitude estimation.
• It outperforms existing quaternion PF methods in terms of speed and computational resources while maintaining comparable accuracy.
• This method holds significant potential for applications in robotics, navigation, and aerospace.