Novel Algorithm for Fast Accessing On-Board Star Catalog

Fast accessing On-board star catalog is a key research technique in star tracker, which is a high accuracy navigation device on satellite. A novel method to fast retrieve the star data from the On-board star catalog, which is based on version of regression Support Vector Machines (SVM) is presented in this paper. An index access to the star triplet catalog is formed with a curve fitting function using Support Vector Regression (SVR). Experimental results demonstrate that the proposed method is faster than traditional binary search and other search methods. It reduced the first search range from 9276 to nearly 70, and save 7 catalog accesses compared with binary search.