Deep learning techniques for micro-launchers branching trajectories optimization

This thesis was created with the purpose of investigating about the nature of branching trajectories for the class of micro-launchers. The results generated are the outcome of different methods of approach for which both the theoretical concepts and the practical results are reported. The reference model used is Electron produced by RocketLab, however IRATO (Intelligent Rocket Ascent Trajectory Optimizer) is scalable to any class of multi-stage launcher. The proposed goal is to present to the reader a tool that is able to provide robust results based on unconventional optimization techniques and artificial intelligence elements.