Download or read book A Probabilistic Framework to Locate and Track Maneuvering Satellites written by Zachary Joseph Hall and published by . This book was released on 2021 with total page pages. Available in PDF, EPUB and Kindle. Book excerpt: This dissertation develops a novel probabilistic framework for space domain awareness (SDA) with a specific focus on tracking noncooperative maneuvering satellites while accurately accounting for uncertainties in maneuver parameters and sensor observations. The proposed framework consists of a series of interconnected modules to accomplish the tasks of model prediction, sensor tasking, and model-data fusion for accurate tracking and maneuver estimation in a data-sparse environment. An important feature of the developed framework is that it provides accurate confidence intervals along with the point estimates for orbit states and maneuver parameters. Target satellite maneuvers are defined stochastically, and a Taylor series-based method termed higher-order sensitivity matrix (HOSM) method is developed to determine the mapping of stochastic initial conditions, maneuvers, and model parameters to future states, i.e., reachability set. The reachability set represents all possible target states given known control bounds and defines a search space for the target satellite. A systematic reachability set search (RSS) method is developed using a maximum detection likelihood criterion such that a maneuvering target for which tracking custody has been lost may be searched for and located. A higher-order moment matching (HOMM) method is developed to curtail the curse of dimensionality associated with multiple maneuvers. The independence of orbit state and future maneuver parameter is exploited to maintain constant target state dimensionality throughout the developed framework. The HOSM, RSS, and HOMM methods are validated via extensive numerical simulations throughout the dissertation including tracking a maneuvering satellite in cislunar space. Simulation results provides a basis of optimism for the ideas described in this dissertation. The HOSM method is shown to provide an increasingly accurate mapping of stochastic inputs to target reachability set by increasing the order of included sensitivity matrices. The nominal error in HOSM approximation is dependent on simulation parameters like propagation time, dynamic model, and uncertainty bounds; however, in all test case the root mean square (RMS) error between linear sensitivity matrix and fourth-order sensitivity matrix approximation decreases by approximately four orders of magnitude. Implementing the RSS method illustrates a number of key characteristics, namely that sensor data where the target is not detected are used to improve the estimate of target state and maneuver parameters. The true target state and maneuver is always contained within the estimated probability density function (pdf) of the target, even without detecting the target. This means that sensor data observing vacant regions of space can be used to inform sensor tasking at future timesteps. Using the simulation parameters defined in each test case examined, and randomly selecting 250 realizations of true maneuver parameters, the target was detected within half an orbital period for $100\%$ of cases for a target making a single maneuver in Geosynchronous Equatorial Orbit (GEO), $96\%$ of cases for a target making two maneuvers in GEO, and $96.8\%$ of cases for a target in an $L_1$ cislunar Halo orbit. These results clearly show the utility of the developed framework for accurately tracking a maneuvering satellite in a data-sparse environment and different orbital regimes.