Autonomous rocket landing stands as a crucial milestone in aerospace engineering, pivotal for the realization of safe and cost-effective space missions. This paper introduces a pioneering approach that harnesses reinforcement learning methodologies to enhance the precision and efficiency of rocket landing procedures. Grounded on a realistic Falcon 9 model, the study integrates sophisticated control mechanisms including Thrust Vector Control (TVC) and Cold Gas Thrusters (CGT), ensuring agile propulsion and balance adjustments. Observational data, encompassing critical parameters like rocket position, orientation, and velocity, guide the reinforcement learning algorithm in making real-time decisions to optimize landing trajectories. Through the strategic implementation of curriculum learning strategies and the Proximal Policy Optimization (PPO) algorithm, the rocket agent undergoes iterative training, steadily improving its capabilities to execute soft landings on designated pads. Experimental results underscore the efficacy of the proposed methodology, exhibiting remarkable proficiency in achieving precise and controlled descents. This research represents a significant stride in the advancement of autonomous landing systems, poised to revolutionize space exploration missions and unlock new frontiers in commercial rocketry endeavors.
AI-Guided Rocket Landing: Navigating Precision Descent Strategies
Advances in intell. Systems Research
International Conference on Emerging Intelligent Systems for Sustainable Development ; 2024 ; Aflou, Algeria June 26, 2024 - June 27, 2024
Proceedings of the International Conference on Emerging Intelligent Systems for Sustainable Development (ICEIS 2024) ; Kapitel : 28 ; 371-386
30.08.2024
16 pages
Aufsatz/Kapitel (Buch)
Elektronische Ressource
Englisch
Mars Precision Landing Using Guided Parachutes
British Library Conference Proceedings | 2009
|