LiDAR-Inertial-Based Absolute Positioning With Sparse DEM for Accurate Lunar Landing

Accurate determination of absolute positions for lunar landing holds a significant role in accomplishing diverse scientific and engineering mission objectives. In this article, we propose a LiDAR-inertial-based absolute positioning system concept for the approach phase of the lunar landing sequence. During the approach phase, since the lander is very close to the ground, the onboard LiDAR can acquire detailed terrain. On the other hand, the terrain maps constructed from data obtained by high-altitude reconnaissance orbiters are inherently sparse. Therefore, the resolution disparity makes it difficult to determine the absolute position by matching the two data, which is why existing systems primarily rely on relative navigation near the lunar surface. To address this challenge, this study employs an ensemble Kalman filter (EnKF) to mitigate the solution ambiguity caused by the sparseness of initial maps. Furthermore, we present a fractal terrain model and a two-step filtering framework to leverage the characteristics of terrains and EnKF effectively. The validity and accuracy of the proposed algorithm are evaluated through numerical simulations, while feasibility is confirmed through a scaled-down experiment. Both numerical simulation and experimental results, employing a map with a resolution of 60 m, demonstrate the potential to achieve absolute position determination with an error of approximately 10 m (1$\sigma$).