Autonomous State Estimation for Vision-Based Navigation in Lunar Landing

Abstract

Since the dawn of space exploration, the Moon has been the primary target of numerous missions. In recent years, this interest has been renewed, as evidenced by the growing number of initiatives focused on lunar exploration led by space agencies and private companies. The main ambition is to establish a sustained human and robotic presence on the lunar surface, enabling the construction of permanent infrastructure and the exploration of scientifically valuable regions such as the lunar south pole. To achieve these objectives, onboard navigation systems must be both autonomous and precise to enable safe and accurate landings. While the long-term solution may involve the deployment of a lunar constellation providing positioning and communication services, in the near term this challenge can be addressed through crater-based optical navigation. This work assesses the integration of Crater Navigation (CNav), a crater identification and matching algorithm developed by the German Aerospace Center (DLR), into a navigation system designed for absolute navigation in the lunar environment. The implemented system is based on a sensor fusion scheme that combines measurements from an inertial measurement unit, a star tracker and CNav. In particular, an error-state extended Kalman filter is employed to estimate the spacecraft’s position, velocity, and attitude with respect to a Moon-Centered Moon-Fixed reference frame. Simulation results demonstrate that the filter consistently estimates the state variables and their associated uncertainties in both orbital and descent phases. In particular, promising performance is achieved in position and attitude estimation, showing that CNav-based optical navigation is a promising candidate for future autonomous lunar navigation systems. This work highlights the potential of crater-based navigation systems as a key component for upcoming lunar exploration campaigns.