Browsing by Subject "radanmääritys"

The thesis presents results for simulating an orbit determination process with least-squares fitting (LSF) of synthetic observations for non-cooperative satellites. Five satellites with different masses, sizes, and orbital inclinations are orbiting in Low Earth Orbit (LEO). Three different inclinations (53°, 85°, and 98°) are simulated using Orekit (Orbit Extrapolation Kit), an open-source astro- dynamical software library. Satellites are observed with four hypothetical radars. The thesis is a feasibility study. It addresses four main research questions: How reliable are predictions of the satellite position for 24 h, 48 h, 72 h, and 96 h after the first detection when using the observations from four consecutive overpasses and the LSF-estimated orbit? How accurately the Keplerian or- bital elements be modelled with LSF? How does the satellite’s mass affect the results’ reliability, and can the satellite’s drag, reflection, and absorption coefficients be estimated with Orekit? The simulation utilizes the Orekit software library in a Python environment, and the simulation incorporates several perturbing forces: Earth’s gravitational potential model EGM2008, an atmo- sphere model with space weather data NRLMSISE-00, tidal forces, and point-like masses of the Sun and the Moon. Synthetic observations were simulated by assuming Gaussian distribution for the uncertainties of the measurements. The initial estimate of the orbit is done with Gooding’s method, and the Levenberg–Marquardt algorithm is used for the LSF. The propagation of the actual satellite for 96 h was compared with the 96-h propagation of estimated orbital elements and with a satellite with estimated size and mass. Simulations were conducted with overestimated and underestimated initial masses. Each combination of a satellite, an observing radar, and an orbit was repeated 100 times. The results for reliability were promising. The likelihood of the sky-plane projected error of the position staying under the 3.0-km for up to 3 days was found to be 90% for two of the simulated radars, but only in cases of orbits having an inclination of 53°. Two other orbits had less successful results, but simulations revealed a potential increase in reliability by adjusting the observational strategy. There was a small bias in the results hinting that underestimating the satellite’s mass could result in less accurate orbit estimates than overestimating it. The satellite’s drag, absorption, and reflection coefficients could not be obtained due to the short observational time intervals simulated.