Skymark selection algorithm for a space-based navigational concept

Abstract

The angles-only navigational concept envisions using orbiting satellites and other space objects as a set of moving landmarks, or skymarks, to improve upon an observer's knowledge of its position in space. This is accomplished by taking line of sight, or angles-only, measurements from an observer to a skymark using a modified star tracker, and then making use of triangulation to estimate the observer's location. A Kalman filter is used to update the observer's knowledge of its state, and the uncertainty concomitant with that knowledge can then be determined. This thesis covers the development of an algorithm for selecting four successive skymarks to be viewed by an observer, with a goal of minimizing both the observer's error in its position estimate and the time required to make the selection. During the development process a number of different algorithms were developed and compared, both with one another and with an algorithm that determined the optimal answer, in order to determine the one that best met the above criteria. Factors taken into account by the various algorithms include the geometry of the four observed skymarks, the distance between observer and skymark, the errors associated with the observer's knowledge of the skymark's state, and the effects of the observer's motion on error propagation.