Neera Jain, Purdue University; Mariel Borowitz, Sam Nunn School of International Affairs, Georgia Institute of Technology
Keywords: space traffic management, policy and technology co-design, conjunction analysis, human operation of satellites, satellite autonomy, system-of-systems methodologies
Abstract:
With increasing space traffic, particularly in Low Earth Orbit (LEO), the likelihood of collisions between resident space objects is also increasing. This in turn has put more pressure on policy-making entities to develop rules, guidelines, or policies that govern how commercial or state actors respond to conjunction warnings. At the same time, increasing levels of autonomy are being developed and implemented on satellites to aid in maneuvering when a conjunction is likely. SpaceX, for example, has stated that satellites within their Starlink constellation use autonomous maneuvering capability to avoid collisions. However, the nature and extent of this capability is generally unknown. The current best practice for collision avoidance involves direct communication between the owners of satellites with a probable conjunction, and manual coordination of maneuvers to avoid said conjunction. However, there is no requirement for action or for coordination between operators or agencies. While increasing autonomy could alleviate the need for additional human satellite operators to plan and execute maneuvers to manage space traffic, it likewise poses several challenges, including the potential for unintended behaviors resulting from a lack of coordination or communication between independently-designed algorithms for collision avoidance. Additional potential for unpredictable behavior may occur when a human operator remains in- or on-the-loop.
These possibilities suggest a complex network of decision-making within the domain of space traffic management that occurs on very different timescales and involves different amounts of uncertainty. For example, decision-making meant to be influenced by domestic or international policy evolves on a very slow time scale, and enforcement and compliance may not be guaranteed. With respect to human operator behavior, variance among operators is extremely high, and the decisions made by these humans is a complex function of their own experience, their understanding of a given situation, their tolerance for risk, as well as interpersonal dynamics that surround a given situation or context. The decisions made by any autonomous system or algorithm(s) designed to avoid collisions will occur on the fastest timescale where the certainty is a function of the verification or certification process involved in approving deployment of the technology and how the autonomy on one satellite interacts with neighboring autonomy. Characterizing interdependencies between decisions made across these timescales, and the associated uncertainty, is not trivial.
Fortunately, the problem of space traffic management (STM) can be considered a system-of-systems (SOS) as its constituent elements (individual satellites or satellite constellations) embody operational and managerial independence and are geographically distributed. Additionally, satellite capabilities and operations, including relevant policies and stakeholders, change over time, and emergent behavior can be expected as a result of the interactions between RSOs, related operations, and policies. Therefore, a SOS methodology can be used to define and abstract a modeling framework for STM that captures how these decisions influence one another. SOS methodologies have been used successfully for modeling a wide range of complex systems, including air traffic management and adoption of powertrain technologies in freight transportation networks. For example, in the context of air transportation, an SOS approach was used to examine how to transform the air transportation system to a state that could satisfy growing demand efficiently; model analysis revealed how the desired system-level behavior could be achieved given the individual decisions of airlines to grow and restructure routes, and the inherent interactions that would evolve as a consequence, conditioned on specific variations in infrastructure. Similarly, in the context of regional Class 8 truck freight transportation, an SOS approach revealed that to achieve widespread adoption of battery electric powertrains for Class 8 trucks, not only did the payload capacity need to meet that of diesel vehicles, but charging times would need to be less than 2 hours. These types of findings are useful because they can provide guidelines not only for policy makers, but also provide specifications for technology developers as well. SOS methodologies provide tools for decomposing the broader system across multiple dimensions—hierarchical and temporal—while also considering the resources, operations, policy, and stakeholders of relevance across these dimensions. Once this multi-dimensional SOS is defined, a modeling framework comprised of the appropriate inputs and outputs can be abstracted and then motivate the appropriate mathematical tools for ultimately building and simulating the model.
In this study, the definition and abstraction of a SOS model for STM is presented, along with a roadmap for design and parameterization of a computational model. Such a model could be used to answer questions coupling technology and policy, such as “How will differing levels of autonomous capability across satellites affect STM?” and “Is there a minimum standard for autonomous maneuverability that should be mandated in order to minimize collisions?” The proposed SOS definition will also provide a framework for multiple stakeholders interested in STM to use to guide their own modeling and analysis efforts.
Date of Conference: September 19-22, 2023
Track: Space Domain Awareness