ESA GNC Conference Papers Repository

Future missions for robotic planetary surface exploration are becoming ever more ambitious in their science goals as they attempt to answer bigger questions relating to the possibility of life elsewhere in our solar system. To achieve this increased level of ambition, it is becoming much more important to be able to carry out a safe, pin-point landing at a site of high scientific interest. The truly interesting landing sites are rarely found in the relatively flat, hazard free terrain that current missions require for a safe landing, using the proven, but unsophisticated entry, descent and landing system design that has changed relatively little since the early days of space exploration. In this paper we present a self-tuning square root unscented Kalman filter based structure from motion algorithm that uses image measurements from a single camera to estimate the surface structure and produce a digital elevation model (DEM) of the terrain surrounding a planned landing site. Such a DEM could be used in a hazard detection system to enable a safe landing at, or as close as possible to the planned landing site. Results are presented from a segment of a simulated descent scenario using representative images of an artificial planetary surface, which demonstrates that the proposed method has strong potential to supply the sensing accuracy to achieve a safe landing in the presence of surface hazards.