To Jupiter and beyond: AI is transforming space exploration
Artificial intelligence can be a vital resource helping humanity explore the final frontier, be it scientifically, theoretically, or even physically. Learn how it powers state-of-the-art telescopes, aids mission planning, and improves satellites.
Humanity has witnessed incredible technological progress since the first man set his foot on the Moon just over fifty years ago. Take the smartphones we carry around in our pockets every day: in terms of complexity, they have far superseded the computers onboard the Apollo Lunar Module Eagle, which flew Neil Armstrong and Buzz Aldrin safely to the Moon and back.
Artificial Intelligence (AI) and machine learning have great potential in propelling this progress even further. In November 2020, a multidisciplinary panel of experts addressed a wide range of issues on the intersection between AI and space, exploring topics such as data processing, satellite imaging, mission planning, and the importance of global discourse on the future of space governance.
While the virtual panel organised by prg.ai and Startup Disrupt discussed many applications of AI in space exploration, the speakers agreed on one thing: humankind must focus on quality education to ensure we can seize the opportunities that such technologies have to offer.
Let machines crunch the numbers
The Atacama Large Millimeter/submillimeter Array (ALMA) is a state-of-the-art telescope located in the Atacama Desert of northern Chile. It’s used to study some of the coldest and oldest objects in the universe and produces one terabyte of scientific data daily coming from the most energetic and cataclysmic events of the cosmos, contributing to our knowledge of the origins of life itself.
Not only is the sheer amount of data that we obtain from space a challenge, but in cosmic physics, the fact is that data is often ‘noisy’ — meaning that much of what is collected is meaningless. “In the vastness of the universe, specific phenomena often occur very sparsely. Artificial intelligence can help by doing the data processing work while allowing researchers to focus their efforts on observation and analysis,” explained Jan Lukačevič, who works as a cosmic engineer at the Czech Academy of Sciences.
Sorting through and making sense of large amounts of data is an area where AI and data science provide enormous potential, noted ALMA’s data analyst Ignacio Toledo during the discussion.
“Another field where many scientists believe that data science will be beneficial for astronomy is in discovering new scientific knowledge. On the one hand, discovering gravitational laws required a certain level of human abstraction that is not possible simply by correlating data. But we have all this information, what could happen if we let it pass through a deep learning network? Could we discover new laws of physics?” Toledo added.
From risk assessment to autonomous navigation
Another important and incredibly complex aspect of space exploration is mission planning. As an example, the mission planning of the Solar Orbiter began in 1998 only to be launched in February 2020. “In most cases, teams only have one chance. Cosmic technologies require double checks, and everything needs to be extra reliable,” Lukačevič explained during the debate.
How does AI come into play here? Machine learning algorithms are able to calculate trajectories, plan routes, and devise optimal schedules while also assessing the risks and priorities involved in the operation. AI can also help predict weather patterns when scheduling safe launch windows. By automating these processes and alleviating the need for human resources, the amount of time needed to schedule spaceflight is being dramatically cut down.
For instance, the Perseverance rover that is roaming the surface of Mars since February 2021 relied on a range of AI technologies in mission planning and scheduling, including in deploying its supersonic parachute which unfolded autonomously during its landing phase.
AI-powered technologies are also leveraged for the navigation of the rover on the Red Planet’s surface. While it receives daily instructions on the direction in which to go, two navigational cameras and six hazard-detection cameras enable it to recognise obstacles and risky paths and steer around them instead of stopping to wait for further instructions.
Sprucing up the planet — and the skies, too
But let’s return to Earth for a while.
Monitoring our planet from space using satellites is not a new practice — in fact, the first one was launched to space by the Soviet Union in 1957. Yet, satellites have since found much broader applications for governments and businesses, for example, in conservation by tracking changes on the Earth’s surface such as in species occurrence, fires, and sea cover.
AI is helping us employ this technology more efficiently and with increased precision by helping sort through large amounts of redundant data collected by satellites. “If the U.S. National Geospatial-Intelligence Agency was to manually process all of its satellite data for the next twenty years, it would have to recruit eight million human analysts,” said Jakub Brož, who is responsible for product and programme management at SpaceKnow.
Moreover, AI can potentially improve the satellite manufacturing process, extend mission and battery life, and aid in the prevention of satellite collisions. Space is getting increasingly crowded, and deep learning presents an opportunity to minimise the risks associated with space debris polluting the Earth’s orbit.
On an upward trajectory
There are countless promising AI applications in space, and we are only just beginning to unlock its potential. What are its most inspiring uses according to the AI & Space panellists? And what should we focus on to continue driving progress forward?
A point that several speakers emphasised was increased transnational cooperation. “We are not talking about artificial intelligence on a sufficiently global scale,” stressed Petr Boháček, an expert on transatlantic relations, space governance, and politics. “No global institutions are supporting and unifying research and political efforts, and global issues such as that of space are still viewed primarily through the paradigm of nation states.”
“If we are to travel as far as Jupiter and beyond, artificial intelligence will be crucial in enabling us as biological organisms to survive such a journey.”– Tomáš Mikolov, CIIRC CTU
Furthermore, effective science communication is key to engaging future scientists, researchers, and engineers, but also the general public, added X-Challenge co-founder Miky Škoda and programme director at the Brno Observatory and Planetarium Josef Forman. They both lauded the quality of technical education, which, in Czechia, is nothing short of impressive. But going forward, it will be crucial to provide the public with accessible and easy-to-digest information. “It’s the only way to help counter growing disinformation and simplify the issues for everyone,” Škoda stated.
It wouldn’t be a genuine discussion on all this AI and space without a bit of sci-fi to spark our imaginations. “If we are to travel as far as Jupiter and beyond in perhaps a few hundred years, artificial intelligence will be crucial in enabling us as biological organisms to survive such a journey,” prophesied Tomáš Mikolov, a senior AI researcher at CIIRC CTU, who is spearheading research into artificial general intelligence.
To learn about what else AI has in store for us regarding space exploration, head to YouTube and watch a recording of the AI & Space event now.
Next up from prg.ai
Typical Prague AI firm is young, self-sufficient, and export oriented, shows our new comprehensive study
130 companies, 11 interviews, 9 business topics. Explore all that and more in the unique study authored by prg.ai, which contains an overview of last year's most notable events on the local AI scene or articles on the future of AI or gender equality in research.
prg.ai newsletter #30: Start of the year marked by strategic partnerships
On the Prague AI scene, February brought new strategic partnerships, educational projects, and academic achievements. This month’s prg.ai newsletter will keep you in the loop, and thanks to its overview of upcoming events you can keep up with the heart of action next month, too.
ChatGPT: Well trained but inaccurate
The GPT-3 algorithms and the ChatGPT chatbot are all over the place. And rightly so. OpenAI's long-in-the-making family of machine learning models has reached its current peak with the public release of the ChatGPT chatbot. However, like any new technology, it comes with certain caveats.
prg.ai newsletter #29: Successes of the Prague AI scene in 2022
What were the highlights that our local AI ecosystem encountered over the last twelve months? As we welcome the new year and await the exciting challenges it has to bring, let’s take some time to reflect on the successes we accomplished in 2022.