Performance characterization of a non-conventional star tracker based on a hyper-hemispherical panoramic camera

This paper aims at characterizing the attitude determination performance of a non-conventional star tracker based on a hyper-hemispheric panoramic camera designed as a multi-functional sensor for applications onboard small/micro-satellites. The peculiar features of the optical system allow imaging a field of view of 360° in azimuth and up to 135° in elevation (zenith), thus being able to observe an extremely wide portion of the celestial sphere, though at the expense of limited capabilities in terms of detector resolution and sensitivity. For this reason, each mode of the multi-functional sensor requires ad-hoc, original algorithmic solutions. In this respect, the star tracker mode relies on an innovative approach for star identification, based on template matching and image registration concepts inherited from the computer vision and robotic research community, combined with a standard solution to the Wahba's problem to determine the spacecraft attitude parameters during the lost-in-space condition. A numerical simulation environment is developed to realistically reproduce the star pattern imaged by the sensor including all the major sources of noise (e.g., outliers, hot pixels). Hence, the performance of the proposed approach is evaluated in terms of attitude determination accuracy and reliability over a wide set of attitude states characterized by a uniform distribution of the camera pointing in the celestial sphere. Also, this is done considering highly-variable settings in terms of sensor's specifications and algorithm's operational parameters.