Mars rovers localization by matching local horizon to surface digital elevation models

In this work we have performed a sensitivity analysis of the Visual Position Estimator for Rover (VIPER) algorithm using data and images provided by NASA MER exploration rovers and NASA Mars Reconnaissance Orbiter. The algorithm retrieves the rover camera position and orientation relative to a Digital Elevation Model by comparing the skyline extracted from a panoramic image captured by the rover and a set of skylines simulated on a template positions grid over the DEM. This algorithm can be used to initialize the rover position after landing in a Mars Body-Fixed Frame and as verification of rover guidance and navigation outputs. In order to test the algorithm performances we have processed data and images provided by NASA Mars Exploration Rover PANCAM and DEM provided by NASA Mars Reconnaissance Orbiter HiRISE telescope. The sensitivity analysis has been performed by varying DEM resolution and template positions density. In the tested cases we show that this localization technique achieves an error up to 50 [m], thus it is possible to decrease the position uncertainty estimated with other localization techniques, like the Entry Descent and Landing estimation.