space

Where do the buffalo roam exactly? The answer to that puzzle raised by an old cowboy song may signal a new era of radical transparency as AI is increasingly used in conjunction with space satellites to monitor home ranges of all kinds. AI, meanwhile, also is poised to assume a greater role in lunar space missions and may create new space powers able to leverage AI into leading roles in the space economy.

A best guess is that around 22,000 buffalo roam free “Home On The Range” in a remote part of Australia called Arnhems Land, but what is certain is that the buffalo destroy environmentally and culturally sensitive spots whenever they show up.  A mission called Space Cows, with typically Aussie-style humor, is underway to track the buffalo in a large scale herd management program that utilizes GPS and AI to track the animals. Once tagged with solar-powered GPS tags, a buffalo’s location is transmitted to an orbiting satellite. A digital twin of the natural topography created with Microsoft tech and the use of AI algorithms allows rangers to anticipate buffalo movements so important areas can be fenced off for protection.

The Space Cows project is just one example of how AI is being utilized in a wide range of space missions. But perhaps generative AI’s biggest role may be making sense of the overwhelming amount of data now generated by Earth-observing satellites. Weather satellites alone reportedly generate 12 terabytes of data daily. AI promises to reveal important discoveries previously hidden amongst large volumes of data simply by asking a question.

Another key role for AI is to combine satellite and terrestrial data to yield faster assessments in an emergency. AI identification of space weather patterns will improve forecasting for events like solar flares that may adversely affect satellite operations. After a hurricane or an earthquake, for example, AI could be used to quickly show where damage to communications networks and infrastructure has occurred to advise emergency services accordingly. In many instances, speed is of the essence, so on-board processing is becoming more desirable and practical as the required radiation-hardened CPUs shrink in size. The European Space Agency (ESA), which has already funded several such projects, calls the AI-equipped orbiters “cognitive satellites.” AI on-board also is being seen as a way to thwart cyberattacks on satellites. The goal for many in the space community is to link disparate sensing systems that can be analyzed by AI to create what amounts to a central nervous system for the planet.

AI also may have a role in a burgeoning space business: Using satellites to take pictures of other satellites in order to keep track of orbiting assets in an increasingly crowded “Where’s Waldo?” environment. Ground management also is using AI to avoid collisions amongst increasingly large constellations of satellites.

AI is likely to enable the rise of space powers not previously thought of as such, especially if they can gather information not controlled by “Big Data” from their own satellites. This year’s merger of two local firms into an entity called Space42 promises to be the Middle East and North Africa (MENA) region’s “first AI-powered space technology champion” with a listing on the Abu Dhabi Securities Exchange being a potential case in point. New satellite launches from countries like North Korea (widely believed to have had a Russian assist) also add a level of uncertainty as to the role of AI in these instances.

Generative AI analysis of satellite data also can help urban planners identify problem areas when creating new towns, for example. Along those same lines, IBM is working with NASA and its satellite imagery capabilities to use AI to identify areas at risk for flooding and wildfire. Scientists in South Korea also think AI can analyze daytime satellite imagery combined with sources like street-level and aerial photography to estimate levels of poverty in countries from which data is not readily available. Generative AI analysis isn’t without its risks, however. AI might “hallucinate” fake landscapes when it needs to account for data gaps, for instance, so effective guardrails need to be in place.

Generative AI will also have a role when NASA’s Viper (Volatiles Investigating Polar Exploration Rover) lands near the lunar South Pole, a mission scheduled for late 2024. A system called SHERPA (System Health Enabled Real-Time Planning Advisor) helps map routes for the rover. While NASA won’t be handing the keys over to AI completely, AI is expected to be a big help riding shotgun.

“AI allows Viper to be more adaptable, flexible, resilient and efficient,” says Edward Balaban, Viper’s lead for strategic planning at NASA’s Ames Research Center. “It’s a tool that allows us to use change as a strength.”

Longer term, NASA and ESA are examining whether AI-powered “companions” may make a years-long space mission to Mars, for example, more socially palatable for astronauts tired of their fellow crew members jokes and stories. On a grander scale, NASA wants to equip its lunar Gateway space station with a talking ChatGPT-style AI that would help astronauts manage complex systems without having to delve into a manual.

Sending an engineer into space every time something goes offline or software develops a problem just isn’t practical. The idea brings to mind the rather idiosyncratic ship “minds” of sci-fi authors like Iain Banks (an Elon Musk favorite) as well as the inevitable comparison to the film 2001: A Space Odyssey. Note to NASA: just don’t call it HAL.