John Iannamorelli, Purdue University; Keith LeGrand, Purdue University
Keywords: Cislunar, Filtering, Tracking, Space Domain Awareness
Abstract:
Accurate knowledge of a space object’s orbit and its associated uncertainty is essential to cislunar space domain awareness. The probabilistic approach to cislunar space object (CSO) tracking involves finding complete statistical descriptions of state uncertainty conditioned on available noisy observations. Limited object visibility leads to long observation gaps, during which CSO state uncertainty grows rapidly and exhibits a highly non-Gaussian structure driven by the underlying nonlinear and chaotic multi-body dynamics, even when the CSO is assumed to be ballistic. Tracking a maneuvering CSO is even more challenging, as the unknown time and magnitude of CSO maneuvers accelerate uncertainty growth. This paper presents a novel adaptive Gaussian mixture filter for tracking noncooperative maneuvering objects in chaotic dynamic systems. The maneuvering CSO state estimation problem is formulated as a nonlinear jump Markov system (JMS) where the object modality is unknown and subject to random switching. The proposed filter automatically and strategically adapts the mixture resolution to capture nonlinear effects from dynamics, measurements, negative information, and known constraints. The effectiveness of the new filter is demonstrated in a challenging angles-only tracking problem, where a noncooperative CSO performs multiple unknown maneuvers to achieve an Artemis I-like trajectory. Synthetic angle measurements are generated based on a notional network of cislunar space-based observers flying in a 1:1 resonant orbit.
Date of Conference: September 19-22, 2023
Best Paper Award Winner 2023
Track: Cislunar SDA