Oracle for Guidance with Deep Neural Networks in Reusable Launch Vehicle Landing*

Abstract

Oracles are of paramount importance for Deep Neural Networks training. In this paper, an oracle developed for landing reusable launch vehicles is created from a linearizing feedback control law that can perform a prescribed landing trajectory tracking. The oracle is then used to train a Deep Neural Network that can be used as a guidance system for landing maneuvers. Verification is performed by Monte-Carlo.