Lunar Surface exploration based on LCNS orbiters and Onboard Sensor observables

Lunar exploration is a strategic priority to develop and experiment technologies that will pave the way for the future missions to Mars and to other celestial bodies of the Solar System. Robots are expected to prepare the return of humans to the Moon by surveying landing sites, demonstrating in situ resource utilization (ISRU), and expanding our access capabilities to difficult areas, i.e., craters and caves. Succeeding in these challenging tasks requires reliable and efficient navigation and communication capabilities. Therefore, space agencies are encouraging the development of a Lunar Communication and Navigation Service (LCNS) to efficiently support lunar assets. A dedicated LCNS infrastructure would lead to unprecedent advantages in future missions by enabling a constant contact with Earth, even in case of Direct To Earth (DTE) link unavailability, e.g., on the far side of the Moon. To fulfil critical tasks, such as obstacle avoidance, instrument manoeuvring and reaching a precise location on the map, rover near real time positioning is a key requirement. Thus, in our work we investigate a method based on the Extended Kalman Filter (EKF) that implements a multi modal sensor fusion approach to estimate the rover's position and velocity by using observables collected by onboard sensors or provided by a LCNS constellation. We focus on a realistic mission scenario in the Moon's south polar region that includes a robotic vehicle hosting onboard sensors to estimate the travelled distances (Wheel Odometry, WO) and the heading variation (Inertial Measurement Unit, IMU). Furthermore, the LCNS orbiters are supposed to broadcast one-way radio signals that the rover user terminal can detect and exploit, providing GNSS-like functionalities. The rover's localization is accomplished through dead-reckoning during LCNS visibility gaps, by using IMU and WO data and accurate Digital Elevation Models (DEMs) of the lunar surface. Whenever pseudorange and pseudorange rate data are acquired by the rover LCNS terminal, these measurements are processed by the navigation filter in combination with IMU and WO datasets, while optimizing the position, velocity and timing (PVT) computation in terms of integrity, accuracy, and convergence time. The proposed method copes with highly varying LCNS visibility conditions and would significantly improve rover's navigation on the Moon's surface in regions where DTE is not achievable. Moreover, our results confirm that the LCNS would be a valuable source of information to be exploited in combination with onboard sensors to improve the accuracy of the reconstructed rover's traverse.