Astrobiology Leverages Advanced Vision and Artificial Intelligence to Seek Life on Other Planets

Leveraging Advanced Vision and AI to Enhance Life Science Studies in Space

Astrobiology is a multidisciplinary science that studies life in the universe, including origin and evolution. Astrobiology research today is concerned with studying exoplanets (i.e.: planets that orbit stars other than the sun) to find and identify habitable zone planets and moons. Advanced vision technologies and artificial intelligence (AI) are being leveraged to enhance instrument capabilities and expand research possibilities.

Spacecraft, for example, such as NASA's New Horizons interplanetary space probe, are equipped with sophisticated imaging telescopes and instruments that help to meet a mission's science goals. The New Horizon spacecraft was equipped to study Pluto and its icy moon Charon. The spacecraft's science payload includes seven instruments, among which are a visible and infrared imager / spectrometer that provides color, composition and thermal maps; a telescopic camera used to obtain encounter data at long distances, to map Pluto's far side, and to provide high resolution geologic data; and an ultraviolet imaging spectrometer to analyze composition and structure of Pluto's atmosphere and to look for atmospheres around Charon and Kuiper Belt Objects (KBOs). The Kupier Belt object nicknamed Ultima Thule is the most distant object ever explored. Images were taken by New Horizon's telescopic camera on approach some 17,000 miles away!

So how do astrobiology researchers handle all the data that their space exploration provides? Artificial intelligence. AI is valued for its ability to interpret large amounts of imaging data. To advance the interpretation of data, the Frontier Development Lab (FDL), an applied artificial intelligence research accelerator program, was established to maximize emerging AI technologies and apply them to challenges in space science. The FDL has recently used AI to tackle the following astrobiology challenges:

1. Understanding what is universally possible for life by using machine learning to determine the biochemistry of a biosphere and ecological signatures that may suggest life, but ‘not as we know it'
2. Generating 'biohints' about extraterrestrial environments that may have coevolved a range of alternative life processes different from those we observe on earth

As a way of predicting the possibility of life on other planets, the University of Plymouth in the United Kingdom, a world-leading center in the field of computer science, cognitive robotics and neural computation, is conducting a study that uses artificial neural networks (ANNs) to classify planets into five types that are potentially the most habitable for humans. The classification identifies planets that are most like 1) the Earth today; 2) the Earth in its early days; 3) Mars; 4) Venus; or 5) Saturn's moon Titan. 

Life is currently only known to exist on Earth; however, astrobiology, and space exploration in general, are successfully leveraging advanced vision technologies and artificial intelligence to tackle tough challenges to search for life on other planets.