Star Tracker Attitude Estimation for an Indoor Ground-Based Spacecraft Simulator

This paper presents a study of star tracker attitude estimation algorithms and implementation on an indoor ground-based Three Axis Spacecraft Simulator (TASS). Angle, Planar Triangle, and Spherical Triangle algorithms are studied for star pattern recognition. Least squares, QUEST and TRIAD algorithms are studied for attitude determination. A star field image is suspended above TASS. The indoor laboratory environment restricts the placement of the star field to be in close proximity to TASS. This restriction adds some additional complication to the standard attitude determination problem. An iterative solution handles this complication. Experimental verification is also performed for the proposed iterative solution. I. Introduction HE objective is to develop a star tracker precision attitude estimation system for use on an indoor, ground-based spacecraft simulator. Star pattern recognition algorithms are studied with a focus on accuracy and algorithmic efficiency. Attitude determination algorithms are studied similarly. The Three Axis Spacecraft Simulator (TASS) is the testbed and it is equipped with a CCD camera to capture the star field image suspended above. The star pattern recognition algorithm takes the camera image of the star field, assigns to it a mathematical description of the pattern, and finds the unit vector to each imaged star. Stars in the CCD image appear distributed among multiple pixels and the centroid of this distribution must be found. Unit vectors from the focus point of the camera lens to the image plane of the CCD are then found for each star. These vectors are mapped from the camera reference frame to the spacecraft body reference frame. The pattern for the stars in the camera image is then found. The pattern is checked against a database of the entire sky to find a match. The pattern can be defined as simply as an angle. Triangles provide more information for more robust matching. 3,4 Novel methods such as grids can also be robust and efficient. 11 Algorithms used in this study are the triangle, planar triangle, and spherical triangle. Once the star vectors measured in the spacecraft frame are matched to the inertially referenced database of star patterns and star vectors all the necessary information is available to solve the attitude determination problem. Attitude determination algorithms determine the rotation of the star vectors from the inertial referenced frame to the spacecraft body frame. There are many algorithms for attitude determination, but three were studied in this report: least-squares 2 , Quaternion-Estimator (QUEST) 8 , and the TRIAD 8 algorithm.