Star Pattern Identification Using Discrete Attitude Variation Technique

Two star pattern identification methods that use a discrete attitude variation algorithm are presented. When a spacecraft is lost in space, these methods determine the inertial attitude of the platform iteratively, starting with a hypothesis corresponding to a pair of the brightest measured and reference stars. In the first method, the spurious match of the pairs is avoided by using the magnitude of the stars in conjunction with an angular separation database. in the second, a novel spiral search on an index to the catalog star magnitude replaces the separation database. Attitude is confirmed with the identification of the remaining star vectors transformed from the spacecraft body to the inertial reference frame with the discrete attitude. These procedures are subject to numerical simulation tests along with another contemporary star identification algorithm. The new methods result in a successful star identification rate close to 100% with a process time average of a few milliseconds for reasonable perturbation of measurements. The first method shows an overall improvement compared to the reference algorithm, whereas the second one is faster only at the lower brightness noise of the star tracker.