Self-Calibration for Star Sensors

Jingneng Fu,Ling Lin,Qiang Li

DOI: https://doi.org/10.3390/s24113698

IF: 3.9

2024-06-07

Sensors

Abstract:Aiming to address the chicken-and-egg problem in star identification and the intrinsic parameter determination processes of on-orbit star sensors, this study proposes an on-orbit self-calibration method for star sensors that does not depend on star identification. First, the self-calibration equations of a star sensor are derived based on the invariance of the interstar angle of a star pair between image frames, without any requirements for the true value of the interstar angle of the star pair. Then, a constant constraint of the optical path from the star spot to the center of the star sensor optical system is defined to reduce the biased estimation in self-calibration. Finally, a scaled nonlinear least square method is developed to solve the self-calibration equations, thus accelerating iteration convergence. Our simulation and analysis results show that the bias of the focal length estimation in on-orbit self-calibration with a constraint is two orders of magnitude smaller than that in on-orbit self-calibration without a constraint. In addition, it is shown that convergence can be achieved in 10 iterations when the scaled nonlinear least square method is used to solve the self-calibration equations. The calibrated intrinsic parameters obtained by the proposed method can be directly used in traditional star map identification methods.

engineering, electrical & electronic,chemistry, analytical,instruments & instrumentation

What problem does this paper attempt to address?

This paper mainly discusses the on-orbit self-calibration of star sensors. Star sensors are important components of high-precision satellite attitude determination and control systems, which need to be pre-calibrated in the laboratory. However, the launch process and space environment may cause deviations from the calibration state, so on-orbit calibration is needed to maintain measurement accuracy.

Currently, there are two main methods for on-orbit calibration, which are dependent on attitude information and independent of attitude information. The methods dependent on attitude information require prior information such as gyroscopes or landmarks, while the on-orbit calibration methods independent of attitude use the constant angle between stars pairs to avoid coupling between parameters. These methods are usually based on the invariance of star pair angles, but they require a large amount of data storage and are prone to overfitting.

This paper proposes a new on-orbit self-calibration method for star sensors that does not rely on star recognition. It reduces the estimation deviation in calibration by defining a constant optical path constraint from star points to the center of the sensor's optical system. In addition, it uses a scale nonlinear least squares method to accelerate iterative convergence. Simulation and analysis results show that the focal length estimation bias of the on-orbit self-calibration method with this constraint is two orders of magnitude smaller than that of the unconstrained method, and it can reach convergence within 10 iterations. The intrinsic parameters obtained by the proposed calibration method can be directly used for traditional star pattern recognition algorithms.

In summary, this paper aims to solve the "chicken and egg" problem in on-orbit self-calibration of star sensors, that is, the order of star recognition and intrinsic parameter determination. It improves the accuracy and efficiency of calibration through innovative calibration models and algorithms.