This project focuses on optimizing Earth-to-Mars transfers by integrating reachability set analysis and reinforcement learning (RL). Our objective is to develop a method for designing multiple impulse transfers to Mars, wherein reachability set analysis assists in identifying admissible states for RL exploration. The trajectory transfer design is formulated as a Markov process, with RL guiding the selection of position vectors (waypoints) along the transfer path. Lambert's arcs serve as the mechanism for accomplishing the transfers between states and automatically satisy constraints, a notoriuos major difficulty for RL-based algorithms. The entire project will be implemented using Matlab, providing a comprehensive framework for advancing the efficiency and reliability of Earth-to-Mars space missions, and similar interplanetary transfers.
Undergraduate
A Matlab software to optimize interplanetary transfers
Very good maths and programming skills.
Students should have selected AEROSPCE720 as elective.
Space Institute Lab (405.443, Lab)