Development of an EKF-based tracking algorithm for unresponsive objects in cislunar space and methods for its use in cislunar surveillance architecture design
Neel Puri, Michael J. Steffens, Dimitri N. Mavris · 2024
As humanity expands into cislunar space, operational saftey will necessitate a robust and reliable surveillance architecture. Such an architecture will need to be capable of tracking numerous objects through space with minimal and/or infrequent observations. This paper outlines an Extended Kalman Filter that can track unresponsive objects in cislunar orbit; and then explores the filter's behavior under relevant conditions. Three unique cislunar orbits are tested using two different experiments. The first experiment varies the number of observations. The second varies the "clustering" of the EKF's observations. Each experiment logs the EKF's tracking ability based on the converged covariance bounds. The results are then analyzed for relevant insight into how an EKF responds to conditions that a cislunar surveillance system is likely to encounter. The results produce three core conclusions. The first is that increasing the time between observations has a direct, potentially linear, relationship with EKF tracking performance with no clear tipping or drop off point. The second is that an EKF's ideal observation spread is highly dependent on the type of target orbit. The third is a series of issues that bring into question an EKF's effectiveness in this use case.