Error Analysis and Calibration Method of a Multiple Field-of-View Navigation System

Area of Science:

• Spacecraft engineering
• Robotics and autonomous systems
• Computer vision and optical systems

Background:

• Autonomous navigation is critical for spacecraft operations within confined environments like space stations.
• Accurate calibration of navigation systems, including intrinsic and extrinsic camera parameters, is essential for achieving required performance.
• Existing calibration methods may not be suitable for the specific constraints of multi-camera systems in space.

Purpose of the Study:

• To introduce the Multiple Field-of-view Navigation System (MFNS) for autonomous spacecraft navigation inside the Tiangong Space Station.
• To propose and validate a novel calibration method, checkerboard-fixed post-processing calibration (CPC), for simultaneously determining camera intrinsic parameters and coordinate transformations.
• To assess the performance and accuracy of the MFNS after calibration.

Main Methods:

• Development of the MFNS architecture, mathematical modeling, and error analysis.
• Implementation of the checkerboard-fixed post-processing calibration (CPC) method using a two-axis turntable.
• Theoretical derivation and practical application of the CPC method for intrinsic and extrinsic parameter calibration.
• Conducting calibration experiments and navigation tests to verify system performance.

Main Results:

• The CPC method achieved extrinsic parameter accuracy of 0.1° for Euler angles and 0.6 mm for position vector components (1σ).
• Navigation experiments confirmed the MFNS's proper functioning with position vector accuracy of 1.82 mm and Euler angle accuracy of 0.17° (1σ).
• The proposed calibration method demonstrated convenience and potential for integration into toolkits.

Conclusions:

• The MFNS is a viable subsystem for autonomous spacecraft navigation within space stations.
• The novel CPC calibration method provides high accuracy for multi-camera systems, essential for space applications.
• The MFNS architecture and calibration approach offer a valuable reference for future multi-camera system designs.