Abstract
Space Surveillance and Tracking (SST) is an increasingly necessary field of research due to the exponential growth of debris. During this thesis several space environments are going to be simulated while surveillance and tracking is performed. First, a brief opening of the SST history and main problems is presented.
Then, in the methodology part, two methods of orbit determination are depicted: Initial Orbit Determination (IOD) and Statistical Orbit Determination (SOD) in order to solve a debris tracking problem, deepening in several of their different mathematical algorithms and architectures. For the IOD techniques, Gauss, Gibbs and Herrick-Gibbs methods are going to be studied and analysed while for SOD; Joint Probabilistic Data Association (JPDA), Global Nearest Neighbour (GNN) and Track Oriented Multi-Hypothesis Tracker (TOMHT) trackers will be developed and analysed, to fit the space scene simulated.
Afterwards, the software simulation and space environment modelling is deeply explained, by generating several scenarios with different configurations to study the proposed algorithms and their performances.
Finally, in the last part of the thesis, all these algorithms and their parameters are compared through several usual metrics: Root Mean Squared Error (RMSE) or Average Normalized-Estimation Error Squared (ANNES); and some special tracking metrics: Generalized Optimal Subpattern Assignment (OSPA/GOSPA); leading to a similar Multi Target Tracking (MTT) configuration for orbit determination using the Unscented Kalman Filter (UKF) to predict and correct the nonlinear debris' dynamics, the JPDA algorithm to correlate measurements with existing tracks; a gating and track maintenance structure to minimize the metrics quoted before.
