The Indian Learning.
The use of AI is growing at an unparalleled rate in the field of space exploration. In space, literally, there are more stars than there are grains of sand on Earth, and each of those stars may harbour life or have a planet that is potentially inhabitable. We would run out of time even though all humans were to unite under one umbrella and research each of those stars. A quicker, safer, and more reliable solution needs to be available that can take care of all the dreary work, which appears to be none other than artificial intelligence.
This article is based on a careful analysis of research articles and internet-based resources combined with personal opinion to understand and appreciate the potential of AI in space exploration.
The Past and Present of AI in Space Exploration
AI (Artificial Intelligence) has always been a comrade of space research organisations like NASA, the European Space Agency, the CNSA (China National Space Administration) and Space X. The origin of AI and space exploration is older than anyone might think. A Rocket Booster technology was developed during World War II, allowing the first generation of spaceflight, with artificial satellites and interplanetary probes being launched by the Soviet Union and the United States using AI technologies. The journey had already begun in the mid-20th century. Towards the end of 20th century, SKICAT (Sky Image Cataloguing and Analysis Tool) detected what was beyond human competence. It classified roughly a thousand objects in low resolution during the second Palomar Sky Survey. Astronomers have used similar AI systems, Convolution Neural Network, in finding 56 new gravitational lenses.
With the commencement of the 21st century came another breakthrough in space exploration, the success of EO1 (Earth Observing 1). The EO-1 satellite had been efficacious in obtaining images of natural calamities. Even before the ground crew realised that the occurrence had taken place, the AI operating with it began to take pictures of the hazards. This was the first satellite—
· to map active space lava flows.
· to assess the methane leakage of a facility from space.
· to track re-growth from space in a partially logging Amazon forest.
NASA along with Google made 2017 the year of the scientific breakthrough by discovering two unidentified planets by the help of artificial intelligence viz. Kepler 90 (now Kepler 90i) and Kepler 80 (now Kepler 80g).
Why is AI Used in Space Exploration?
Exploring the space gives rise to massive quantities of information that cannot be processed by human intelligence. This is where applications of artificial intelligence count. AI can alter the course of space exploration by analysing and extracting the meaning of the data. Researchers may find life on new planets with the results. It will help to recognise and monitor trends that humans may not have made possible. It is also possible to identify planets that have the correct conditions to sustain human life.
The rovers presently exploring the surface of Mars are expected to make decisions without the mission control's explicit commands. AI applications are what really makes this possible. For example, the NASA Curiosity rover could very well move on its own, avoiding obstacles on the way and identifying the best route to take.
We get data in the form of pictures from space. However, the task is to decipher those images and extract the information required. Here, machine learning will help. NASA Frontier Development Lab and tech giants like IBM and Microsoft have unified to harness machine learning as a solution for detecting solar storm damage, measuring the atmosphere, and measuring the 'space weather' of a given planet by evaluating the magnetosphere and atmosphere.
Machine Learning, a derivative of Artificial Intelligence, played a major part in the successful landing of SpaceX Falcon 9 at Cape Canaveral Air Force Station in 2015. It determined the ideal way to land the rocket via real-time data that enables route prediction.
The geological composition and historical importance of a planet can be understood via AI applications. Not only this, but AI can also submit, evaluate, and classify images of the same and decide on the next appropriate move.
Deep Learning, a branch of Artificial Intelligence could be used for automatic landing, logical decision-making and fully automated systems.
The next generation spacecraft will be more independent, self-sufficient and autonomous thanks to the Artificial Intelligence applications. AI would go beyond human limitations to perceive observations and submit information back to Earth.
AI applications can augment planetary tracking systems, allow smart data transmission, and eliminate the risk of human error by way of predictive maintenance.
India’s feat in Space Exploration using AI
Indian Space Research Organisation had created a solar-powered robotic vehicle named Pragyan which would explore the lunar surface. This Pragyan was integrated with Chandrayaan-2 rover. Pragyan contained LIBS (Laser Induced Breakdown Spectroscope) from the Laboratory for Electro Optic Systems, Bengaluru which would identify elements present near the landing site as well as an APIXS (Alpha Particle Induced X-ray Spectroscope) from the Physical Research Laboratory in Ahmedabad which would enable inspection of composition of the elements identified by LIBS.
Artificial Intelligence enabled the Chandrayaan-2’s rover in numerous ways -The AI-powered Pragyan was able to communicate with the lander. It contained motion technology that was designed to help the rover travel and land on moon's surface. Moreover, the artificial intelligence algorithm could aid the rover spot signs of water and other minerals on the lunar surface. AI could well allow the rover to share pictures that would have been used for research and experimentation.
The Way Ahead
It is a widely accepted fact that artificial intelligence is the key to unlock further developments in almost every arena. With respect to space exploration, there are a number of ways in which artificial intelligence can help.
1. Astronaut Assistants: Virtual assistants can possibly detect any hazards in long space flights, such as disturbances in the spacecraft atmosphere or medical assistance to astronauts who get sick due to zero-gravity conditions. AI can help condition humans prior to their long-distance space journey and can be particularly useful in conducting operations in deep space or on another planet when the reporting system fails to communicate. With deep learning techniques applied to speech recognition and facial recognition, AI could also hold a two-way conversation with astronauts and learn from their conversations.
One such space exploration camaraderie between a human and a machine began with CIMON (Crew Interactive Mobile Companion) on 29 June 2018, CIMON (a compact AI-endowed football-shaped robot) was launched on a 2-day mission to the ISS on Space X Dragon Cargo Capsule. Cimon can be used to alleviate the anxiety of astronauts by completing tasks they require it to do. NASA is also designing a companion for astronauts on board the ISS, called Robonaut, who will work closely astronauts or undertake the tasks that are too hazardous for them. Another noteworthy example is the Japanese Space Agency, which has developed an intelligent system—Int-Ball by JAXA for the ISS to take pictures of observations in the Japanese module. This autonomous, self-propelled and navigable ball camera used current drone technology and was intended to support astronauts with on-board challenges and exploration missions.
2. Planning and designing missions: Conventionally, fresh space missions depend on the information gathered in previous studies. However, this information may often be limited or not readily accessible. There is a need for a smart system that can respond to researcher's queries in real-time. Researchers are working on the concept of a Design Engineering Assistant to minimise the time needed for initial mission design, which would otherwise take several hours of human work. "Daphne" is an example of an intelligent assistant in the design of Earth observation satellite systems. Daphne is used by system engineers in satellite design teams. It supports their work by providing access to pertinent information including reviews and answers to specific problems.
3. Satellite Data Processing: Although there have been several crowdsourcing projects intended at satellite imagery analysis but only on a limited scale. For detailed analysis, artificial intelligence can prove to be a boon. In a recent research, scientists have tested various AI techniques for remote satellite health monitoring systems. This is effective for analysing data received from satellites to spot any problems, assess satellite health performance, and present a visual image for strategic planning.
4. Tackle Space Debris: According to European Space Agency, there are almost 34,000 objects greater than 10cm that pose significant risks to current space infrastructure. Artificial intelligence might solve this problem. In a recent study, researchers have developed a framework for developing ‘collision avoidance manoeuvres’ using machine learning (ML) approaches. One way to ensure the safety of space flights has recently been suggested by using already competent networks onboard the spacecraft. This allows for more versatility in the design of satellites while at the same time minimising the danger of collisions in space.
5. Effective Navigation in Space: NASA Frontier Development Lab has been working on an AI programme that functions like a GPS in space and makes it easy to reach Titan, Mars or even the Moon. By use of GPS as well as other GNSS systems in Medium Earth Orbit (MEO), Geostationary Orbit (GEO) and beyond, is "an emerging capability," as per Miller (Positioning Navigation and Timing (PNT) policy chief for the NASA Goddard Space Flight Center).
It is evident that artificial intelligence has the much-needed potential to facilitate advanced space exploration programmes and possibly aid in discovering more exoplanets which would not have been possible with traditional technologies and human intelligence.
Ronald Von Loon, How AI is Transforming Space Exploration, LinkedIn (Feb 8, 2021, 2:30 pm), https://www.linkedin.com/pulse/how-ai-transforming-space-exploration-ronald-van-loon Artificial Intelligence in Space Exploration – Importance of AI in Space Exploration, AIlabs ( Feb 8, 2021, 2:52pm),https://ailabs.academy/artificial-intelligence-in-space-exploration-importance-of-ai-in-space-exploration/#:~:text=The%20history%20of%20AI%20and,than%20anyone%20could%20possibly%20think.&text=AI%20has%20also%20helped%20in,2%2C545%20light%2Dyears%20from%20earth. Sakshi Gupta, AI Applications in Space Exploration: NASA, Chandrayaan2 and others, Springboard (Feb 8, 2021, 3:02 pm), https://in.springboard.com/blog/ai-applications-in-space-exploration-nasa-chandrayaan2-and-others/#:~:text=AI%20in%20Indian%20Moon%20Mission%20%E2%80%93%20Chandrayaan2&text=And%20that%20was%20the%20integration,surface%20on%20its%20six%20wheels.  Ibid.  Five ways artificial intelligence can help space exploration, The Conversation (Feb 9, 2021, 2:28 pm), https://theconversation.com/five-ways-artificial-intelligence-can-help-space-exploration-153664.