There is no GPS in the space. In case people are on their way to Mars or got lost on the lunar surface, they will be much screwed. NASA in association with Intel does not want this sort of things to happen in any scenario, so they came to AI with the expectation of optimum solution.
NASA FDL (Frontier Development Lab) recently ended its 8-week summer session with an event organized by Intel. The program has 9 teams co-relatively working on the same issue of space traveling and colonization in the extraterrestrial region. The goal of the program is to find out these knowledge gaps with the help of Artificial Intelligence.
The event involved the hunt for more exoplanets, mission to observe the sun from more close and some more topics about the existing and future spacefarers. In the demonstration, there were 2 presentations which were specifically focused on the issue of direct navigation in an extraterrestrial region for robots and humans both, and a method that can help us to deal with this issue with the help of deep learning.
The space travelers first ever found this issue of localization on the lunar surface. With the advent of new methods and technologies, we cannot face this kind of issue on Earth. We have access to physical maps, GPS, digital application, visual landmarks or ask anyone around you.
So, how can AI assist in overcoming these issues? Andrew Chung, Benjamin Wu, Philippe Ludvig and Ross Potter are the scientists who developed a framework for simulation of the lunar surface and compare it with the native environment. Along with this, they have developed a program to train AI about the appearance of Moon by feeding it with images and then finally with the help of a neural network to build a virtual moon.