Resilient and Sustainable Planning and Management of the Future Space Industry Infrastructure with On-Orbit Servicing – ReSuSpace

The ReSuSpace project aims to optimize on-orbit servicing (OOS) resources and services by integrating demand-side uncertainties, such as the number and location of satellites requiring services. It introduces quantitative resilience and sustainability metrics into a multi-criteria decision framework, beyond simple cost and service considerations. A general and efficient mathematical programming framework will be developed for the optimal planning and scheduling of OOS platforms. The project will explore suitable artificial intelligence methods for autonomous decision-making under uncertainty, integrating data via functional approximations to efficiently find solutions in continuous action spaces. In addition, the potential of Deep Reinforcement Learning (Deep RL) will be explored to solve complex decision problems, with validation against the developed mathematical optimization methods. These advances aim to improve the flexibility, resilience, and sustainability of the space industry, while reducing operating costs and risks associated with space debris, enabling a more sustainable and efficient space architecture.

/

/

The scientific results will be published in international journals and conferences in the fields of operations research, space engineering, and machine learning, such as EJOR, Acta Astronautica, and conferences such as EURO and the International Astronautical Congress (IAC). The demonstration platform will help promote scientific culture through teaching and outreach, with participation in events such as «Vive la recherche.« Scientific workshops will invite the scientific community, industrial stakeholders, and space agencies to demonstrate the importance of industrial engineering methodologies. At the university level, the knowledge will be integrated into courses for the MSc&T in Space Business Strategy. The project's results could be used to create deep-tech startups at CentraleSupélec.

The space industry is a key sector for the functioning of crucial industrial systems as it provides communication, navigation, and Earth observation services. Today, the space industry is undergoing a fundamental transformation with the access to space becoming easier and cheaper and the cost of developing and deploying space systems significantly decreasing. New industrial actors are rapidly developing the future space production systems: building mega-constellations (tens of thousands) of satellites to provide global services such as Internet-of-Things (IoT), and developing in-orbit manufacturing facilities to produce new mechanical and pharmaceutical products. Together, these satellites and in-orbit manufacturing facilities constitute the future space production infrastructure. Today, this space infrastructure is operating in one of the harshest environments, subject to a myriad of uncertainties and is largely inaccessible for inspection, maintenance, refuelling, upgrade, or end-of-life disposal. The current space industry operates within a “one-off” paradigm, where the only way to recover a failed space system or to update its existing capability is to replace the faulty system with a new one, e.g., construct and launch a new replacement satellite. This results in extremely low flexibility, high costs and unsustainable space environment where disposed satellites become space debris . To address this problem, an alternative paradigm has recently emerged based on On-Orbit Servicing (OOS), which consists of the deployment of robotic On-orbit Servicing Vehicles (OSVs) in space that can provide maintenance, inspection, repair, refuelling, recovery and upgrade services to those satellites, significantly reducing their cost of operation, improving their flexibility to new demands and their resilience towards failure.

The objective of the ReSuSpace project is to develop quantitative decision-aid tools grounded in advanced mathematical modelling, operations research, systems engineering, resilience and sustainability assessment, and artificial intelligence for the optimal planning and management of the lifecycle of the space industry of the future. Particularly, the aim is to develop modelling and solution methods for the optimal planning of robotic On-Orbit Servicing (OOS) systems to: (1) extend the lifecycle of space production systems (i.e. satellites) by providing life-extension services such as predictive maintenance; refuelling; upgrade and inspection that increases their resilience towards failure; and (2) ensure the industry’s sustainability by assisting with active debris removal; satellites collision avoidance to prevent the creation of new debris; and end-of-life de-orbiting of inactive satellites.