Global sensitivity analyses for test planning with black-box models for Mars Sample Return.

Abstract

This work describes sensitivity analyses performed on complex black-box models used to support experimental test planning under limited resources in the context of the Mars Sample Return program, which aims at bringing to Earth rock, regolith, and atmospheric samples from Mars. We develop a systematic workflow that allows the analysts to simultaneously obtain quantitative insights on key drivers of uncertainty, the direction of impact, and the presence of interactions. We apply optimal transport-based global sensitivity measures to tackle the multivariate nature of the output and we rely on sensitivity measures that do not require independence between the model inputs for the univariate output case. On the modeling side, we apply multifidelity techniques that leverage low-fidelity models to speed up the calculations and make up for the limited amount of high-fidelity samples, while keeping the latter in the loop for accuracy guarantees. The sensitivity analysis reveals insights useful to understand the model's behavior and identify the factors to focus on during testing, in order to maximize the informational value extracted from these tests and ensure mission success even with limited resources.