“Avoid the crater ahead” – Lunar navigation and AI – AI Daily

Original article was published on artificial intelligence

Getting lost on the Moon is a common problem, so it is reassuring to know that when you do miss the next lunar highway exit, AI software can guide you safely to where you want to go.

NASA’s Frontier Development Lab (FDL) seeks to “apply AI technologies to science to push the frontiers of research and develop new tools to solve some of the biggest problems that humanity faces”. Each year, the FDL highlights a series of problems that AI could be used to rectify – involving commercial partners (like Intel), research talent (those studying data science at a post doc level) and large data sets. This ensures that participants have the maximum available scope to innovate, explaining the high success rates of the programme.

One of the problems outlined at the 2018 FDL was “Localisation: Merging orbital maps with surface-perspective imagery”. Navigation on Earth is blessed by the aid of GPS – we know exactly where we are on Earth at any given time, but the same cannot be said for the lunar landscape. Hence, it is desirable to convert satellite images and surface data into a mapping tool so that astronauts on the lunar surface are able to know exactly where they are on the surface – to a high degree of precision.

This is where AI comes into play – a live-feed from cameras attached to the astronauts’ helmets provides real-time visual data which is fed into AI software that compares the data with an existing database of the lunar landscape – using a positive match to churn out the astronaut’s location. This sort of technology can greatly improve the efficiency of manned surface missions on different celestial bodies – the surface data we’re currently gathering on Mars may well be used to create navigation software for future astronauts exploring the Martian surface – all thanks to AI…

Thumbnail from Mashable India and image below from FDL themselves