AI for Space Exploration: How the Space Industry Uses Machine Learning and AI Tech
17th Jul 2024
Artificial intelligence (AI) and machine learning (ML) are having outsized impacts on space exploration. Orbital Today will be exploring some of the ways in which this is happening, because more automated decision making is happening on the industry’s constructions in space and on the ground than ever before.
AI and ML are not the same thing, and this distinction should be kept in mind. AI is essentially human-like thinking exhibited by machines, allowing them to solve complicated problems. Machine learning can be seen as a segment of AI that usually refers to computer systems utilising large datasets to grasp new insights or to pull such insights out in an otherwise cumbersome manner.
From major space agencies like NASA to startups and academic institutions, organisations and space experts around the world are using AI and ML technologies to monitor spacecraft for faults, search for life beyond Earth, find new planets and galaxies, leverage earth observation data to tackle issues like climate change and for many other purposes.
Space agencies and AI
The world’s leading space agencies are investing into AI and ML to support their missions and operations. NASA, for example, over 30 AI and ML applications in its AI for space exploration inventory.
NASA’s usage includes systems for organising planetary rover missions; detecting aircraft and ships from satellite images; predicting the intensity of hurricanes; mapping Earth’s surface water; predicting algal blooms; identifying flight anomalies; streamlining image searches; determining methane clouds on Saturn’s largest moon, Titan; creating pedestrian safety corridors for drone testing; estimating the position of spacecraft landing on Mars, and many more.
To support its continued use of AI technologies, NASA appointed David Salvagnini as its first-ever Chief Artificial Intelligence Officer in May. Salvagnini, who is also the agency’s Chief Data Officer, is tasked with planning NASA’s AI usage and deployments while considering and mitigating ethical concerns and risks. As part of his new role, he’ll work closely with partners in government, academia and industry to help the space agency remain at the forefront of AI innovation.
In Britain, the UK Space Agency has also shown an active interest in AI technology in recent times. In 2023, it provided over £500,000 of funding to ten projects using AI, Earth Observation, satellites and big data to help businesses understand environmental harm and become more sustainable. One of these companies included environmental consultancy GeoSmart Information, which received £42,544 to research how tech can aid water resource management.
AI is listed as one of the 11 critical technologies in the UK Space Agency’s Space Exploration Technology Roadmap. In this report, the agency describes AI as an important tool for “enhancing the capability and efficiency of missions”.
Another major space agency utilising AI solutions is the Indian Space Research Organisation (ISRO). During its Chandrayaan-3 mission, the agency used AI to help find a suitable site for a soft landing. The six-wheeled Pragyan rover, a vital part of the Chandrayaan-2 mission, is powered by AI technology. Similarly to NASA, ISRO also develops AI tools for climate-related goals like monitoring deforestation.
Startups at the cutting edge of AI innovation
While large space agencies like NASA are continuing to dedicate large parts of their budgets to AI adoption, startups also play a leading role in the development of innovative AI and machine learning technologies for space exploration.
American startup Slingshot Aerospace, for instance, offers AI-powered software for identifying anomalous behaviour in spacecraft, creating naturalistic simulations for trainees, generating realistic constellation visualisations, and gaining actionable insights from space data. Its solutions are used by the likes of NASA, United States Space Force, U.S. Air Force, U.S. Navy, the UK Space Agency, Hiscox and many other global organisations, illustrating the growing mainstream adoption of AI space tools.
Based in Finland, Kuva Space is another high-growth AI space startup. Through the biggest hyperspectral satellite constellation and artificial technology, it hopes to tackle some of the world’s most prevailing issues — including climate change, food security and national security — by providing data insights from space. As reported by Tech Crunch, the company is currently embarking on a mission to send 100 satellites to space by 2030. Last November, it raised €16.6 million in a Series A investment round.
In the UK, Earth-i is combining the power of earth observation services and AI technology to help businesses and organisations make more informed decisions across areas like natural disaster response, global stockpile management, border and civilian security, climate change and more. The company has previously received funding from the UK Space Agency.
Trinity Space Technologies, another British space technology startup, also sees a bright future for AI and space exploration. Tom Love, co-founder of Trinity Space Technologies, tells Orbital Today that the startup is increasingly deploying AI technology across its space systems and customer support.
“Advanced systems engineering and space logistics tools enable the streamlining of critical processes, ensuring mission reliability and safety while allowing more time and resources to be allocated to advancing organisations down the Technology Readiness Level (TRL) scale,” he says.
As part of an ongoing research project, Love says his team has implemented AI technology in an in-orbit object tracking system. This allows them to conduct close-proximity operations “safely and effectively”.
He adds: “This is crucial for reducing the risk of catastrophic collisions between spacecraft, which could terminate the mission and generate potentially hundreds or thousands of debris fragments.”
Space AI in academia
In academia, scientists worldwide are leveraging AI to learn more about planets, galaxies and life beyond Earth. For example, American academics have created an AI system for finding aliens. It can determine the biological or non-biological nature of samples – take, for instance, rock remnants from Mars – with an impressive accuracy rate of 90%.
Another groundbreaking study saw scientists use deep learning to improve the search for planets like ours. They successfully tested the system on three stars: The Sun, Alpha Centauri B (HD128621) and Tau ceti (HD10700).
Elsewhere, scientists have found ways AI can manage space rubbish, search for asteroids, model outbursts of energy in black holes, discover new galaxies, detect space plasmoids and so much more.
AI isn’t perfect
Although AI is proving to be a practical space exploration tool, it’s not without challenges. Dr Shirin Dora — a lecturer at Loughborough University in the UK — points out that an issue with artificial neural networks is that they can be energy-intensive and, based on his research, believes that spiking neural networks could be a more sustainable alternative.
“Existing research has focused on adapting existing algorithms for ANNs [artificial neural networks] to train SNNs [spiking neural networks] with the goal of using their inherently energy-efficient computing capabilities,” he says.
“My research goes beyond this by focusing on developing spike-level algorithms to build SNNs that use as few spikes as possible to perform any task, thereby significantly lowering the energy needs of trained SNNs. The algorithms developed through my research could easily be adapted to a broad variety of problems in space like satellite communication and navigation.”
Exploring space isn’t just about finding aliens or new planets; it greatly benefits humanity by tackling problems such as climate change, natural disasters and food poverty. What’s clear is that AI is making these things even easier, and with new AI space use cases constantly emerging, the possibilities seem limitless.
Thank you for your comment! It will be visible on the site after moderation.