Paul Billings, KBR / Pacific Defense Solutions; Jason Baldwin, Complex Futures LLC; Charles J. Wetterer, KBR; John Gaebler, KBR; Christopher Craft, KBR; Micah Dilley, KBR; Jill Bruer, AFRL
Keywords: Cislunar, IOD, Constrained Admissible Region
Abstract:
High levels of nonlinearity inherent in the cislunar orbital regime present stressing challenges for the estimation and custody of cislunar objects. Despite this, effective orbital determination strategies using Unscented Kalman Filters (UKFs), Gaussian Mixture Models (GMMs)and, in the most stressing cases, Adaptive Entropy-based Gaussian-mixture Information Synthesis (AEGIS)have been demonstrated. Before these track maintenance functions can be accomplished, however, initial orbit determination (IOD) must be achieved. The Constrained Admissible Region Multiple Hypothesis Filter (CAR-MHF) was previously developed by AFRL for IOD, employing an iterative filter-smoother for estimate refinement and the use of a multi-hypothesis, joint probabilistic data association (MH-JPDA) framework for multi-target data association. In this paper, simulated Earth-based and space-based angles-only observations of cislunar objects in various orbital families are utilized to evaluate the ability of CAR-MHF to perform IOD successfully in the cislunar regime. Cold-start catalog generation scenarios are considered in the presence of ballistic objects across a diverse set of orbit families, constellations, object separation events, and break-up events. Of critical importance in the cislunar regime is the use of a constrained admissible region to bound the landscape of considered orbits given the presence of a single short-arc tracklet of angles-only measurements. Significant treatment is given to the trade-offs involved in bounding the scope of admissible orbits with a focus on uncertainty evolution, sparseness of the hypothesis sampling, compute load, and other considerations. CAR-MHF shows promising results across the spectrum of evaluated scenarios and opens the door to information-based optimization of sensor tasking.
Date of Conference: September 19-22, 2023
Track: Cislunar SDA