Star tracker orientation optimization using Non-dominated Sorting Genetic Algorithm (NSGA)

One of the devices used to determine the attitude of a satellite is the star tracker, whose principle of operation is based on star position measurements on a specific inertial frame, allowing precise attitude determination and control of the satellite. Due to the high sensitivity of star camera, bright objects like Sun, Earth or Moon must be avoided in the sensor's field of view. This characteristic imposes a design constraint that shall be satisfied simultaneously by all the star trackers used in the specific satellite. Considering only the Sun exclusion case, the goal of this work is to find the star tracker orientation that maximizes simultaneously the Sun exclusion angle for both sensors in a way to ensure the proper equipment operation during a typical Earth pointing satellite mission. For this optimization problem, the search space is defined by the azimuth and elevation of each star tracker in the body centered coordinate frame system. Since the engineering goal implies a vectorial objective function optimization, the method used in this work was the Non-dominated Sorting Genetic Algorithm (NSGA), which allows a multi-optimization problem solution without the scalarization approach, in order to get a few optimal solutions along the non-dominated region. In order to get diversity in the optimal solutions, simulations used six different dummy fitness functions and compared the final results.