A Vision Aided Initial Alignment Method of Strapdown Inertial Navigation Systems in Polar Regions

Area of Science:

• Robotics and Navigation
• Geophysics and Polar Research

Background:

• Strapdown inertial navigation systems (SINS) face challenges in high-latitude initial alignment due to polar longitude convergence.
• Autonomous underwater vehicles (AUV) operating in polar regions require accurate initial alignment for reliable navigation.

Purpose of the Study:

• To develop a novel vision-aided initial alignment method for SINS in polar regions.
• To address the limitations of traditional alignment algorithms in high-latitude environments.

Main Methods:

• Redesigned the initial alignment model using inertial navigation mechanization equations in a transverse coordinate system (TCS).
• Integrated visual measurement information from an onboard camera.
• Analyzed observability under various swing models.
• Employed an extended Kalman filter for information fusion.

Main Results:

• The proposed vision-aided method significantly improves the accuracy of SINS initial alignment in polar regions.
• Estimation accuracy of the deviation angle remained consistent across uniaxial, biaxial, and triaxial swing modes.
• Simulation results validated the method's effectiveness and alignment with observability analysis.

Conclusions:

• The novel vision-aided approach offers a robust solution for SINS initial alignment in polar regions for AUVs.
• The method demonstrates improved accuracy and consistent performance under different dynamic conditions.