Automated near real-time validation and exploitation of optical sensor data for improved orbital safety

Abstract

Abstract The orbital safety of operational spacecraft, regardless of the mission, relies on timely and actionable observations to maintain so-called “custody” of all trackable Resident Space Objects (RSOs), including space debris, that might pose a hazard to safe, secure, and sustainable operations. For operations in and around the Geosynchronous Earth Orbit (GEO) regime, electro-optical (EO) observations are the most prevalent observation type available for tracking and determining RSO orbits. The quality (both accuracy and precision) of the data affects the inferable kinematic, physical, and other characteristics of RSOs and, in particular, measurement biases will result in inaccurate orbital trajectories and subsequent predictions. Physically meaningful conjunction assessments rely on not only accurate orbit state prediction, but also the “realistic” covariances associated with said predictions. In this paper we demonstrate an automated near real-time (NRT) assessment of measurement biases with an appropriately implemented Unscented Schmidt Kalman Filter (USKF) [14]. Hypothesized biases that are deterministic but statistically unobservable in the measurement data and cannot be estimated are accounted for as so-called “consider” parameters. The method presented herein is assessed and quantified using both simulated and actual measurement data. This method will enable the exploitation and mining of so-called “non-traditional” sensor data to maximize Space Situational Awareness (SSA) in a robust and timely fashion toward improvement of orbital safety. The ultimate goal is to provide decision-making evidence required solve problems preventing the space domain from being safe, secure, and sustainable.