An Information Theoretic Approach to Sample Acquisition and Perception in Planetary Robotics

TL;DR

This paper introduces an automated, information-theoretic method for planetary sample analysis that uses shape matching and eigenfaces to efficiently identify and prioritize interesting geological samples, even with sensor noise.

Contribution

It presents a novel shape matching approach combining the Superformula and eigenfaces for automated sample prioritization in planetary exploration.

Findings

Robustness to 20% sensor noise

Effective shape categorization and prioritization

Analysis of shape parameters on matching accuracy

Abstract

An important and emerging component of planetary exploration is sample retrieval and return to Earth. Obtaining and analyzing rock samples can provide unprecedented insight into the geology, geo-history and prospects for finding past life and water. Current methods of exploration rely on mission scientists to identify objects of interests and this presents major operational challenges. Finding objects of interests will require systematic and efficient methods to quickly and correctly evaluate the importance of hundreds if not thousands of samples so that the most interesting are saved for further analysis by the mission scientists. In this paper, we propose an automated information theoretic approach to identify shapes of interests using a library of predefined interesting shapes. These predefined shapes maybe human input or samples that are then extrapolated by the shape matching system using the Superformula to judge the importance of newly obtained objects. Shape samples are matched to a library of shapes using the eigenfaces approach enabling categorization and prioritization of the sample. The approach shows robustness to simulated sensor noise of up to 20%. The effect of shape parameters and rotational angle on shape matching accuracy has been analyzed. The approach shows significant promise and efforts are underway in testing the algorithm with real rock samples.