A Norm-Minimization Algorithm for Solving the Cold-Start Problem with XNAV

Abstract: An algorithm is presented for solving the cold-start problem using observations of X-ray pulsars. Using a norm-minimization-based approach, the algorithm extends Lohan's banded-error intersection model to 3-dimensional space while reducing compute time by an order of magnitude. Higher-fidelity X-ray pulsar signal models, including the parallax effect, Shapiro delay, time dilation, and higher-order pulsar timing models, are considered. The feasibility of solving the cold-start problem using X-ray pulsar navigation is revisited with the improved models and prior knowledge requirements are discussed. Monte Carlo simulations are used to establish upper bounds on uncertainty and determine the accuracy of the algorithm. Results indicate that it is necessary to account for the parallax effect, time dilation, and higher-order pulsar timing models in order to successfully determine the position of the spacecraft in a cold-start scenario. The algorithm can uniquely identify a candidate spacecraft position within a 10 AU $\times$ 10 AU $\times$ 0.01 AU spheroid domain by observing eight to nine pulsars. The median position error of the algorithm is on the order of 15 km. Prior knowledge of spacecraft position is technically required, but only to an accuracy of 100 AU, making it practically unnecessary for navigation within the Solar System. Results further indicate that choosing lower-frequency pulsars increases the maximum domain size but also increases position error.