Asteroid Detection by AI: Future Possibility or Reality?
Spotting Hidden Asteroid Paths with Artificial Intelligence
In recent years, the rapid advancement of modern technology has positioned artificial intelligence (AI) as a cornerstone in the technological world. The establishment of various artificial intelligence-based companies such as Sam Altman’s OpenAI as well as Jensen Huang’s NVIDIA has caused the accelerated use of AI across various fields.
Asteroids pose a significant risk to life on Earth, which is why their detection is pivotal to ensure humanity’s safety as well as to advance research in astrophysics and astronomy. Traditionally, asteroid detection relied on calculations made by humans by utilising ground-based telescopes as well as human observation alone.
By observing an asteroid from a distance, and by obtaining information about it such as its velocity, the direction it is moving in etc., we can figure out its trajectory and determine where it will be in the near future. However, we cannot calculate the long-term trajectory of a certain asteroid due to the three-body problem, which states that the position of any body in a system that has 3 or more bodies cannot be calculated and remains in chaos. Because of this, calculating the trajectory of an asteroid over a long-term distance is rarely completely accurate. Both AI and long-term asteroid detection seem to share a problem — accuracy. However, when put together, they truly produce some miracles.
How AI Enhances Asteroid Detection
Data Processing and Analysis
AI algorithms can process huge amounts of data in very small amounts of time, which means if the correct astronomical data about the asteroid is input into the AI model, it can calculate the short-term trajectory immediately. Modern Machine Learning Models can search through images captured by to identify potential asteroid threats that human observation and identification simply cannot. By scanning each and every pixel in an image, the AI can identify asteroids that humans would have never noticed.
Pattern Recognition
Modern Deep Learning Models excel at pattern recognition, making them excellent for identifying subtle changes in brightness or other factors that could indicate the orbit of an asteroid around a body or even the asteroid itself.
Automation
AI does not need to eat, drink, or sleep. It can operate 24/7 without any disruption, which allows it to identify asteroids at midnight, at early morning, at high noon, and at any time of day whatsoever! No shifts necessary at all!
Recent Applications
Recently, scientists from the Asteroid Institute trained an AI named THOR to identify asteroids in images. The team then provided THOR with 400,000 images from the institute’s archive and surprisingly, the AI discovered 27,500 previously unknown asteroids!
The European Space Agency(ESA) has also integrated AI into their NEO Coordination Centre to detect asteroids, which has helped the centre process large amounts of data collected from their telescopes.
In 2022, the Vera C. Rubin Observatory developed the HelioLinc3D program, which was an AI Algorithm built to detect asteroids. The algorithm managed to detect its first asteroid successfully in mid-2022, which was a 180-meter wide asteroid, named 2022 SF289, which is expected to approach Earth within a distance of 384,400 kilometers, a relatively close distance. But don’t worry, its trajectory is still hundreds of thousands of kilometers away from Earth and probably won’t impact our planet.
Gemma Parreño, a data scientist, has accepted NASA’s challenge to help them track and find Asteroids using AI. She created a program named DeepAsteroid using Google’s open-source AI platform. DeepAsteroid was used to assist scientists in detecting, tracking, and classifying asteroids.
Challenges
AI models rely on high-quality data, so disruptions or disturbances in the images given or data presented can affect the results and identification quality of the AI.
Another challenge is the process of training the AI Model or creating an algorithm to perform asteroid detection, which would require tons of testing, trial and error, and of course, a huge amount of data processing, requiring a significant amount of computational power.
Conclusion & The Future for AI
While Artificial Intelligence has a promising future in asteroid detection and already a promising start, for now, it remains more of an assistant to humans in the field of asteroid detection. Though, future advances in technology and increased computational power are likely to enhance AI’s abilities. Paving the way for AI to play a crucial role in protecting our beloved planet from asteroids.
Acknowledgements
https://www.space.com/ai-finds-first-potentially-dangerous-asteroid
