An heuristic approach to the problem of initial orbit determination: A test case with GEO observations

The orbit determination of a satellite in a geostationary orbit is usually based on optical observations. Telescopes all around the Earth are observing the geostationary ring to obtain periodically new measurements in order to update the estimation of the objects′ orbits. One critical aspect of the orbit determination process is the initial orbit determination because no a priori information are available. This paper proposes a new method based on search algorithms with heuristic principles aiming to find the six independent orbital parameters that best fit the observations. A genetic algorithm makes the candidate solutions moving in a solution space of six dimensions. The problem is reduced into a two dimensional space with a Lambert's solver using the initial and final angular observations. This paper presents the results using a single image acquired by a Earth-based telescope, the acquired streak is correlated to the objects of the NORAD database.