A software package for evaluating the performance of a star sensor operation

TL;DR

This paper presents a MATLAB software package for evaluating star sensor algorithms' performance, including simulations with ideal and realistic conditions, to aid in designing and optimizing star sensors for small satellites.

Contribution

The software package enables comprehensive performance evaluation of star sensor algorithms under various simulated conditions, facilitating hardware-agnostic assessment and optimization.

Findings

Performance metrics like attitude accuracy and calculation time are quantified.

Simulation results highlight the impact of noise and hardware parameters on accuracy.

The package supports different hardware configurations for versatile evaluation.

Abstract

We have developed a low-cost off-the-shelf component star sensor (StarSense) for use in minisatellites and CubeSats to determine the attitude of a satellite in orbit. StarSense is an imaging camera with a limiting magnitude of 6.5, which extracts information from star patterns it records in the images. The star sensor implements a centroiding algorithm to find centroids of the stars in the image, a Geometric Voting algorithm for star pattern identification, and a QUEST algorithm for attitude quaternion calculation. Here, we describe the software package to evaluate the performance of these algorithms as a star sensor single operating system. We simulate the ideal case where sky background and instrument errors are omitted, and a more realistic case where noise and camera parameters are added to the simulated images. We evaluate such performance parameters of the algorithms as attitude accuracy, calculation time, required memory, star catalog size, sky coverage, etc., and estimate the errors introduced by each algorithm. This software package is written for use in MATLAB. The testing is parametrized for different hardware parameters, such as the focal length of the imaging setup, the field of view (FOV) of the camera, angle measurement accuracy, distortion effects, etc., and therefore, can be applied to evaluate the performance of such algorithms in any star sensor. For its hardware implementation on our StarSense, we are currently porting the codes in form of functions written in C. This is done keeping in view its easy implementation on any star sensor electronics hardware.