The ever increasing number of space debris objects orbiting the Earth is becoming a larger threat to the present and future satellite operations and spaceflight. To ensure the safety of these missions, the orbits of these space debris objects need to be precisely known. At the moment this is not possible for small (1-10 cm) space debris objects in the Geostationary ring. This thesis attempts to solve this problem by optimizing a constellation of satellites using optical observations for the number and accuracy of the orbit determinations. This was done by simulating real-life optical observations and estimating the orbit using a combination of a simplex fusion algorithm and an unscented Kalman filter. The optimized design showed good performance for newer space debris objects, however, as the inclination of the space debris objects gradually increased during its lifetime the performance of the design dropped rapidly.