Analysis and Improvement of Star Identification Algorithm Based on Singular Value Decomposition for Star Sensor

Abstract

One of the attitude sensors for the spacecraft is the star sensor. The attitude calculation by the star sensor includes several steps: image processing, star identification, database search, and implementation of the attitude determination algorithm. This paper presents analysis and improvements on the star identification algorithm based on singular value decomposition. To improve the identification results, modifications have been made to the algorithm. Also, analysis and simulation are presented to investigate the effect of field of view dimensions and the number of stars used on the identification results, and the rate of duplicate sets in the database. In addition, identification has been considered a two-step process based on singular values and vectors. The results show the superiority of the improved algorithm in increasing the identification rate and reducing the rate of duplicate sets in the database. The identification rate of the improved algorithm in 〖10〗^°×〖10〗^° and 〖12〗^°×〖12〗^° fields of view dimension is always more the %97. The simulations were performed based on the Hipparcos star catalog for stars brighter than magnitude 6.5 and using MATLAB software.

Keywords

• star sensor
• star identification
• singular value decomposition
• lost-in-space identification algorithm