Taking Your First Crack at Geolocation Using ODTK

ODTK 7.4 introduces the Initial Geolocation capability. You may be asking, “How is this different from my usual geolocation process, where I use a filter to refine the location of a facility marked for estimation?”

First, that’s uncanny, because I was just about to answer that. Second, let’s consider the filtering process you described. You need a facility location to serve as a seed to the estimation process, but what if you don’t have that yet? What if you have an incredibly sparse set of measurements to locate your facility and all you need is a rough solution? In those cases, the new Initial Geolocation object is for you.

Think of Initial Geolocation as a close relative to Initial Orbit Determination (IOD). With IOD you want to produce an approximate satellite state directly from raw observations. That coarse solution typically serves as an initial seed that is later refined using the filter-smoother process. Initial Geolocation works the same way, but instead finds an approximate solution for a facility location. Just like IOD, Initial Geolocation uses a small set of measurements to apply geometric constraints to the problem space and ultimately locate the facility. The exact geometric constraints applied depend on the geolocation method being used (and are described to a greater degree in the ODTK help), but they generally boil down to measurement-dependent isogram curves and Earth-surface constraints.

You can see an example of what is going on above. In this scenario, an earth-based facility of unknown location emits a single signal. Two satellites passing overhead pick up that signal, and each satellite processes it into a distinct measurement. Along with an assumed Earth-surface constraint, each measurement’s implied geometric constraints can then be used to produce curves called isograms. It’s the isograms, visually displayed here as two deformed circles on the Earth’s surface, that will ultimately identify the facility’s location.

For a facility to generate the measured signal associated with the isogram, the facility is required to exist somewhere on that isogram. With only one isogram, we have an infinite set of solutions along the path of the curve and we will not be able to pinpoint the facility. Once we consider the second isogram though, we can reduce the problem space down to the isogram intersections and leave only two possible facility locations. From here, we can use a process of elimination (eg. we know the facility is not in the ocean) to identify the correct solution, or we can feed both solutions into the filter-based geolocation process to find which initial seed is more agreeable within our larger ODTK scenario.

Currently, Initial Geolocation supports the use of SB TDOA measurements from configurations of either three or four satellites. Both approaches require the measurements to be derived from the same source signal. If this new functionality piques your interest but does not reflect your measurement system, be sure to keep an eye out for future updates from the ODTK team! We are already cooking up new functionality for Initial Geolocation as we speak.