CubeSat-Based Lunar Map Refinement Utilizing Surface Beacons and a Monocular Camera

SpaceX, Blue Origin, NASA, and others have recently proposed autonomous missions in preparation for new manned missions to the moon. Traditional approaches based solely on inertial navigation are not accurate enough to autonomously land a vehicle on hazardous lunar terrain, therefore Terrain Relative Navigation is being explored to supplement inertial navigation. Terrain Relative Navigation (TRN) is a capability that uses images of local terrain captured with a camera and/or imaging LIDAR to estimate the position and/or velocity of a spacecraft. Many TRN methods estimate the craft's absolute position by comparing sensor imagery to a crater/landmark database, global map, or another similar reference set. Because of the limited availability of high resolution lunar maps, the global accuracy of lunar TRN is currently limited to approximately 100 m. One compelling solution to improving global map resolution is to utilize an array of low cost CubeSats to image the lunar surface and refine existing maps. This paper explores the effectiveness of such a mission. In particular, the objective of this analysis is to determine the sensitivity of mapping uncertainty to sensor errors.