Justin Kim, BAE Systems, Inc.; Naomi Owens Fahrner, BAE Systems, Inc.
Keywords: Cislunar, Architecture, Design, Optimization, Capacity
Abstract:
Cislunar Space Domain Awareness (SDA) has become an active area of research in recent years due to increased interest in the cislunar region amongst government, scientific, and commercial organizations. An effective cislunar SDA architecture will likely require some combination of collaborative in-space assets; however, the complex dynamics and expansive design space make this a challenging problem to solve. While the desire and need for SDA solutions in the cislunar region is clear, at present there is no standard approach to design, evaluate, and optimize cislunar architectures.
Prior work has framed this cislunar architecture design problem as a multi-objective optimization problem to maximize observability of a volume. Observability is the theoretical Field of Regard, subject to geometric and radiometric constraints. Solutions to this problem have been demonstrated with Monte Carlo Tree Search [1], branch & bound, particle swarm, simulated annealing, and other evolutionary algorithms. However, observability alone is not sufficient to assess mission performance and evaluate realistic architectures – as a result, work has been done that demonstrates architecture optimization for observability, with subsequent sensor tasking via capacity optimization. Capacity is a measure of architecture performance, defined by the volume that is observed by the architecture subject to sensor FOV, agility, and tasking constraints, in addition to the observability constraints. Incorporating these factors into cislunar architecture optimization is crucial to assess mission performance. A capacity-based cislunar SDA solution is explored by Fahrner et al. [2], in which the authors initially optimize and select architectures for observability, and then perform a capacity analysis on the best-in-class architectures. Using the same problem formulation and framework, we will compare these solutions with outputs from the unified optimization that is proposed here.
This paper will expand on such work by simultaneously optimizing for observability and capacity. Additional constraints and objectives are imposed on the multi-objective optimization problem, such as orbit stability, scheduling metrics, and a higher fidelity radiometric model. Challenges arise due to the three-body orbital dynamics, occlusion and exclusion viewing constraints, vast ranges, low SNR, and too-short-arc problems [3]. In this work we address such challenges and demonstrate the trades that must be considered in optimizing a cislunar SDA architecture.
As a mixed-integer, multi-objective, non-linear optimization problem, the architecture optimization is formulated as a Markov Decision Process (MDP) as in [1]. A Monte Carlo Tree Search solution was demonstrated which we expand upon by including schedule and tasking related objectives. Past studies have explored some these objectives as isolated pieces, which is already a challenging problem. In this work, the objective space of the problem is expanded to include metrics and constraints that are critical for assessing architectures and evaluating mission performance. The action space of the problem is dramatically expanded by including capacity optimization and sensor tasking. The capacity optimization problem is also formulated as an MDP, in which the sensor searches an RA-Dec Gridworld with an epsilon-greedy algorithm. By simultaneously optimizing for observability and capacity, we provide a single tool that reduces effort and mission knowledge required to evaluate cislunar mission architectures.
By expanding the optimization problem with critical objectives and constraints, we have developed a streamlined methodology that more accurately reflects the complex dynamics and expansive trade space of the problem. In this work we demonstrate a comprehensive, viable, and extensible framework for high-level conceptual design and evaluation of cislunar SDA architectures.
References
[1] Klonowski M, Holzinger M, Owens Fahrner N. 2022. Optimal Cislunar Architecture Design Using Monte Carlo Tree Search Methods
[2] Owens Fahrner N, Correa J, Wysack J. 2022. Capacity-based Cislunar Space Domain Awareness Architecture Optimization
[3] Bolden M, Craychee T, Griggs E. 2020. An Evaluation of Observing Constellation Orbit Stability, Low Signal-to-Noise, and the Too-Short-Arc Challenges in the Cislunar Domain
Date of Conference: September 17-20, 2024
Track: Cislunar SDA