The goal of this ESA-funded project is to investigate the feasibility of recent advances in control and AI for space missions. The first step is to review the state of the art in AI for the development and implementation of embedded guidance, navigation, and control (GNC) systems. The objective is to improve the GNC design process and manage system complexity through the vertical and horizontal integration of disciplines. Upon the outcome of the review, we will establish the functional and performance requirements applicable to an AI-assisted GNC design process and to an AI-augmented GNC system.
In the second phase of the project, we will perform a trade-off of suitable mathematical AI approaches compatible with the current GNC architectures and design processes (model-based approach), including complexity, effort, and expected benefits assessment. This will allow us to establish the AI techniques suitable to the modeling, control, and verification needs in the view of robust and explainable AI-supported GNC architectures and functions.
In the last stage, we will develop a prototype set of benchmark problems for AI-assisted GNC design and AI-augmented GNC systems as well as for AI-supported autonomy (using an in-orbit assembly scenario or precision landing scenario including handling of failures and degradations). We will then perform a detailed design and coding of the established AI techniques applied to AI-assisted GNC design and to the AI-augmented GNC system. Finally, we will assess the performance and robustness of the AI-assisted GNC system and define the way forward for AI-GNC system deployment.