Artificial intelligence for spacecraft guidance, navigation, and control: a state-of-the-art

This article reviews state-of-the-art of the problem of using Artificial Intelligence (AI) for spacecraft Guidance, Navigation, and Control (GNC). Developing AI algorithms for GNC purposes with enhanced and robust performance depends heavily on the models used for GNC algorithms. This is because using AI algorithms may require the designer to train his algorithms using training examples (data sets) via a supervised learning (SL) algorithms. GNC algorithms could be used to provide AI algorithms with the required data sets of training. So, reviewing state-of-the-art GNC algorithms is also mandatory. Good and up-to-date understanding of the current state-of-the art of GNC algorithms enable us to build better survey of using AI algorithms in spacecraft GNC. Unsupervised learning, or reward-based learning algorithms could also be used also for GNC purposes. The current article focuses on GNC algorithms, AI algorithms, and the interaction among AI and GNC algorithms. In addition, a future vision of using AI techniques with GNC is also provided. © 2022, Shanghai Jiao Tong University.