Research/ SPACE AI & AUTONOMY
Lunar Descent and Landing Via Fusion of Deep Learning-Based Visual Navigation and Moonlight Signals
Michele Ceresoli
Perception & State EstimationOpenAlex topic: Inertial Sensor and Navigation · Learning-based MethodsOpenAlex keyword: deep learning · Guidance, Dynamics & ControlOpenAlex topic: Spacecraft Dynamics and Control
OpenAlexSpacecraft Dynamics and Control · Inertial Sensor and Navigation · GNSS positioning and interference
Abstract. The renewed momentum in lunar exploration has driven the development of dedicated lunar communication and navigation infrastructures such as the European Space Agency’s Moonlight Lunar Communication and Navigation System (LCNS). In this context, landing operations are key mission phases that enable humans and scientific experiments to reach the Moon surface. However, recent crashes of both public and private missions have highlighted how challenging and complex such operations are. As such, this work investigates the achievable navigation performance of a lunar lander targeting the Moon’s South Pole through the integration of One-Way Ranging signals from LCNS and Deep Learning–based Vision-Based Navigation measurements, showcasing the benefits that the fusion of these two distinct type of measures can bring.