Estimation improvement of space objects' orbital parameters through optical observations

During the last decades, the population of resident space objects has experienced a significant increase, and it is expected to grow in future years. New exploitation opportunities of the space environment arising with the new space economy, the deployment of satellite mega-constellations, as well as on-orbit fragmentations and collisions, are the main drivers of such overcrowding of near-Earth space. These circumstances demand new techniques and fresh approaches to improve the awareness on the space situation. The free, public orbit catalogue of Two-line elements (TLE) is being used by institutions from all over the world as a reference for orbit propagation and observational ephemerides generation. However, TLE orbits are characterized by modest accuracy, are not provided together with covariance information, and their availability is not guaranteed at regular intervals. Moreover, objects that are classified are not included in TLE catalogue. Operationally, a full orbit determination process typically requires many days of observations, depending on the object characteristics and orbital regime. Optical observations can provide accurate angular measurements at relatively low cost, but they are subject to constraints such as satellite and telescope illumination conditions and good weather. This thesis aims at developing a methodology capable of improving the orbital estimate of space objects with a limited time span of optical observations. In this thesis, an orbit determination algorithm has been developed, to provide orbital estimates in TLE format. Simulations have been performed to assess which orbital regimes and object characteristics may allow an improvement in the orbital parameters with short observational timespan. To weight the initial estimates of the public catalogue in the optimization process, a new technique has been developed for TLE covariance estimation through comparison of subsequent released orbits. Final results have been obtained by acquiring and processing real optical measurements taken through the Sapienza Optical Network and by processing measurements from Deimos Sky Survey (DeSS), and an improved method to estimate measurements error has been developed. This thesis shows that in a broad range of cases, it is possible to improve the knowledge of space object orbits through a very short arc of optical measurements, and the developed techniques can be used operationally to improve space situational awareness and decrease sensor busyness.