Lunar Pit Exploration via Autonomous Micro-Rover

Abstract:

Lunar pits are compelling destinations for surface mission exploration. They could access lava tubes with prospects of haven from the extreme temperatures, radiation exposures and micrometeorite hazards of the surface. Their walls offer direct observation of the only geologic columns not obscured by regolith, and they are scientifically interesting in their own right. The required vantages and proximities for these inquiries are not possible from orbit and motivate robotic surface exploration.

This paper presents an autonomous micro-rover approach to exploration and mapping of lunar pits by in-situ modeling via imagery gathered from their rims. The mission concept incorporates a commercial lander, rover forays around the rim, autonomous acquisition of a vast image dataset, and incremental photogrammetric construction of a pit model from the imagery. We develop the requisite autonomy, modeling, computing, and mobility technologies and present results from a case study performed at a terrestrial analog pit. We demonstrate the coverage and accuracy achievable by this method and conclude with progress toward a fully integrated rover for this important new class of mission.
The rover incorporates telephoto lensing to achieve the requisite resolution for observing detail across vast pit dimensions. For frequent, affordable, commercial landing opportunities, the rover must be small and solar powered. To complete a long range mission before the onset of lunar night, the rover requires unprecedented autonomy, steep-slope mobility, high performance computing and state-of-the-art photogrammetry algorithms.
We present a case study applying micro-rover imaging, computing, and photogrammetry to a terrestrial pit. This acquires 10,000 high resolution, high dynamic range images from 26 poses around the rim of the West Desert Sinkhole which is a terrestrial pit with diameter, depth, and stratification analogous to known lunar mare pits. Results are compared to ground truth LIDAR with error below 10cm across 90% of the pit surface.

Progress to-date suggests that autonomous micro-rover modeling of planetary pits is feasible and useful. Beyond the technical developments, the paradigm, mission, system, and technologies have been advocated to the lunar science community’s PRISM program seeking a flight opportunity.