Landmark Based Position Estimation for Pinpoint Landing on Mars

Abstract

Many of NASA's planned missions for the coming decade will require a pinpoint landing (PPL) capability, whether for sample acquisition and return or for precise insertion into hazardous but scientifically interesting terrain. Thus, a robust spacecraft-based position estimation system is a critical near term need. We present in this paper a vision-based system, which provides a low power, low cost, high accuracy solution with flight proven hardware. The key components of our image based localization approach are: (1) real-time detection of landmarks in descent imagery, (2) real-time matching of detected landmarks to a stored database, and (3) robust estimation of spacecraft state (position, attitude and velocity) from matched 2D-3D data. We present an analysis of the performance and noise sensitivity of our system and show that the suggested technology is able to deliver a spacecraft to within 100 meters of a pre-selected landing site in a typical Mars landing scenario.