Date of Award

Spring 1-1-2017

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Jay W. McMahon

Second Advisor

Penina Axelrad

Third Advisor

Brandon Jones

Fourth Advisor

Daniel Scheeres

Fifth Advisor

Behrouz Touri

Abstract

Very long baseline interferometry (VLBI) specifically same-beam interferometry (SBI), and dual-satellite geolocation are two fields of research not previously connected. This is due to the different application of each field, SBI is used for relative interplanetary navigation of two satellites while dual-satellite geolocation is used to locate the source of a radio frequency (RF) signal. In this dissertation however, we leverage both fields to create a novel method for multi-satellite orbit determination (OD) using time difference of arrival (TDOA) and frequency difference of arrival (FDOA) measurements. The measurements are double differenced between the satellites and the stations, in so doing, many of the common errors are canceled which can significantly improve measurement precision.

Provided with this novel OD technique, the observability is first analyzed to determine the benefits and limitations of this method. In all but a few scenarios the measurements successfully reduce the covariance when examining the Cramér-Rao Lower Bound (CRLB). Reduced observability is encountered with geostationary satellites as their motion with respect to the stations is limited, especially when only one baseline is used. However, when using satellite pairs with greater relative motion with respect to the stations, even satellites that are close to, but not exactly in a geostationary orbit can be estimated accurately. We find that in a strong majority of cases the OD technique provides lower uncertainties and solutions far more accurate than using conventional OD observables such as range and range-rate while also not being affected by common errors and biases. We specifically examine GEO-GEO, GEO-MEO, and GEO-LEO dual-satellite estimation cases. The work is further extended by developing a relative navigation scenario where the chief satellite is assumed to have perfect knowledge, or some small amount of uncertainty considered but not estimated, while estimating the deputy satellite state with respect to the chief. Once again the results demonstrate that the TDOA and FDOA OD results are favorable with faster dynamics over classical measurements.

This dissertation not only explores the OD side, but also gaps in geolocation research. First the mapping of ephemeris uncertainty to the geolocation covariance to provide a more realistic covariance was implemented. Furthermore, the geolocation solution was improved by appending a probabilistic altitude constraint to the posterior covariance, significantly reducing the projected geolocation uncertainty ellipse. The feasibility of using the geolocation setup to passively locate a LEO satellite was also considered. Finally the simulated results were verified using a long-arc of real data. The use of FDOA for small-body navigation and gravity recovery was also examined as an extended application.

Share

COinS