Towards an In-Orbit CubeSat Factory

Abstract

Current satellites must be designed to withstand the vibrations of a rocket launch and are manufactured long before their operational start in orbit, thus creating long lead times, high expenses, and mission inflexibility. In-orbit assembly represents a paradigm shift in the field of satellite production and deployment, vastly increasing flexibility. Last minute configuration changes can be performed and new business models such as customized production directly in space become possible. Our approach has the potential to revolutionize the way satellites are manufactured and deployed. In-orbit assembly is furthermore a necessity for ever larger space structures which cannot be launched in one piece. Our concept of an In-Orbit Factory aims to pioneer a solution, by providing an adaptive just-in-time production. Pivotal for its economic success is a framework for the transition from initial teleoperated assembly towards full autonomy. This allows for a quick start with improved throughput at full capacity. The market size for small satellites was estimated at $3 billion in 2022 and is expected to rise to $15 billion in 2032, opening opportunities for our concept. Our approach of a tight integration of a space-ready robot equipped with advanced control algorithms and factory software is central to the development of In-Space Servicing, Assembly and Manufacturing (ISAM) capabilities. The envisioned system consists of a robotic arm equipped with a gripper and a camera at the end effector, mounted on an assembly table containing modular CubeSat subsystems. Those subsystems are to be integrated into a chassis with a backplane, that is connected to a satellite development board to allow for automated testing of the CubeSat already during assembly. The robotic arm is controlled using a space qualified On Board Computer (OBC) executing the low-level robot control software as well as the higher-level application software. To this end, an adaptive approach of Virtual Fixtures is implemented, supporting both teleoperation as well as fully autonomous execution. Feedback obtained during the teleoperation phase also allows for a self-improvement of the approach. A Digital Process Twin (DPT) orchestrates the system at high level, enabling a holistic process mapping. With this process understanding, it is possible to detect errors at an early stage and to make automated, knowledge-based decisions for their solution. This paper will present the envisioned in-orbit factory concept and analyze the currently available solutions as well as the research gaps to realize the factory. While the implementation using readily available CubeSats will allow for an inexpensive evaluation in space, the solution can easily be scaled to the assembly of larger structures as the challenges remain largely similar.