‘System of systems’: Key for future space mission success
Space exploration is vital, because of the innate propensity of humans to explore the unknown, as well as the tremendous commercial potential. But space discovery is a risky proposition, and, increasingly, mission success depends on a lot of very complex and interrelated things going right. This requires a “system of systems” (SoS) approach — recognizing that the entirety of the undertaking, the “space exploration architecture,” is a system that encompasses numerous other systems, each linked intricately, in a dynamic interaction with one another to fashion an effective, reliable whole.
A space exploration architecture is this comprehensive collection of elements (space systems, spacecraft, satellites, ground stations, astronauts, logistic support, etc.); their interactions, exchanges and interdependencies; and the operations and missions they perform. This wasn’t so much an issue in the past, when missions were limited in scope, and conducted by design to avoid interaction with other nations’ (or companies’) systems. The most complex missions (such as the International Space Station) involved like-minded collaborators that were willing to do the hard work of producing robust systems and proper interfaces (hardware and software).
However, the complexity and number of missions, and the number of systems involved, are growing and likely will grow further, inevitably producing more interactions and complexity.
An example of an issue that highlights this intricacy and interdependence is “space traffic management.” As with air traffic control for our aviation system, in space, we need to know where spacecraft are, where they are going, and what their health status is. But right now, there are few rules — and a dearth of policies — about how this information is provided for spacecraft. Furthermore, when things collide in space, a tremendous number of smaller fragments are produced, which themselves become dangerous objects. While there have been impressive improvements in tracking and classification technology, if policies are not put in place that prevent more space debris from being produced, our ability to track (and avoid) dangerous debris will disappear.
The objective of our work at Purdue is to analyze and compare architectures for human and robotic space exploration, accounting for the impact of interdependencies between and among technologies, policies, and operational concepts. This effort involves identifying relevant items in the architectures, at various levels of abstraction, and modeling and analyzing them. One example might be analyzing space habitats, based on the operational dependency of many systems, including the primary support structure, docking subsystem, crew wireless communication, environmental monitoring, water and waste management, thermal control, food preparation subsystem, and internal vehicle lighting.
Our team uses a suite of specialized systems analysis tools to uncover both hidden barriers to success and emergent opportunities for success. We evaluate such aspects of space architectures, as large-scale requirements, performance, robustness, resilience, technology prioritization, and flexibility. We use technologies like machine learning; deep learning with Bayesian networks (a method of statistical inference to determine probability, named after the 18th-century English mathematician Thomas Bayes); and stochastic (probabilistic) analysis.
For instance, in a project sponsored by the NASA Marshall Space Flight Center, we compared architectures for the human exploration of Mars based on analysis of the impact of interdependencies. The approach involved identifying items in human Mars exploration architectures, and modeling and analyzing them using SoS analysis tools. This work included analysis of alternative propulsion systems for interplanetary cargo and crew transfer, modeling and analysis of Lunar Gateway Habitation Module and its subsystems, and an assessment of priority for developing cryogenic fluid management technologies.
Our SoS tools address high-level, hierarchical features of complex space systems, such as multiple stakeholders, uncertain objectives, developmental/operational risks, and flexibility of space mission architectures. These tools constitute our SoS Analytic Workbench (AWB) at Purdue — a framework for the quantitative evaluation of criticalities, and the impact of interactions between systems and capabilities. The workbench lets us evaluate the risks and robustness of different technological choices and their effects on the entire architecture, to support trade-offs around factors like cost, risk and performance.
We’re also training the next generation. For example, the Purdue Modeling Architectures and Parametrization for Spacecraft (MAPS) environment was constructed in-house by students. The environment, coded in the MATLAB programming language, enables nontraditional analysis of spacecraft and space architectures. Multiple scenarios and architecture designs can be evaluated, and users can perform trade-off studies based on such metrics as complexity and cost.
A challenge continues to be building better and better predictive models — for both mission performance and optimal technology selection — for cases that have never happened, meaning data is scarce. Ideally, all stakeholders will embrace and enable large-scale, continuous “learning models” and research platforms that both help them fulfill their private goals and help the community realize scientific system integration. This approach will avoid negative behaviors and outcomes like accidents, technological delay, and outright conflict in space operations.
Daniel DeLaurentis, PhD
Prof., School of Aeronautics and Astronautics; Dir., Center for Integrated Systems in Aerospace; and Faculty Council Member, Purdue Engineering Initiative in Cislunar Space, College of Engineering; Dir., Inst. for Global Security and Defense Innovation (i-GSDI); Chief Scientist, Dept. of Defense Systems Engineering Research Center (SERC); Purdue University
Cesare Guariniello, PhD
Research Scientist
School of Aeronautics and Astronautics, and Center for Integrated Systems in Aerospace
College of Engineering
Purdue University
