One week ago, the B612 Foundation, a nonprofit group founded by former NASA astronaut Edward Tsang Lu, announced the discovery of more than 100 new asteroids. Better known as Ed Lu, the electrical engineer with a doctorate in applied physics has revealed that space rocks are defined using an innovative tool.
As the newspaper points out The New York Times, asteroid discoveries are reported all the time by observers around the world. This includes amateurs with backyard telescopes and robotic surveys that regularly monitor the night sky.
However, of the approximately 25,000 near-Earth asteroids that are at least 140 meters wide (minimum size to classify them as “potentially dangerous”), only about 40% have been identified. The remaining 60% – about 15,000 – remain hidden.
Experts say that in a possible collision with Earth, each of these rocks in space would have the potential to release energy equivalent to hundreds of millions of tons of TNT – a chemical used as an explosive by the military and in demolitions.
If that happened, the world would surely be destroyed or at least, millions of lives would be wiped out on the planet. In other words: initial identification is essential to try to avoid the impact of one of these “killer” asteroids.
The algorithm mimics the orbits of potential asteroids
Scientists at the University of Washington’s Institute for Data-Intensive Astrophysics and Cosmology, funded by the B612 Foundation, have developed a revolutionary system that will help avert this doomsday scenario.
This innovative computing resource has been applied to approximately 412,000 archival images from the National Optical-Infrared Astronomical Research Laboratory (NOIRLab), to “hunt” asteroids at 68 billion points of cosmic light captured by the entity. telescopes operated by the US government.
According to its developers, the algorithm is able not only to identify bright spots that could be asteroids, but also to find out which spots in the photos taken on different nights are actually the same body. Importantly, the researchers found a way to see what was seen but not noticed.
Named Trackless Heliocentric Orbit Recovery, or THOR, the algorithm produces a test orbit that matches the observed area of light, assuming a specific distance and speed. It then calculates where the asteroid was in the next and previous nights. If the algorithm can gather five or six observations in a few weeks, the object is a promising candidate for a new asteroid.
In principle, there are an infinite number of possible orbit tests to be evaluated, but doing so would require an infinity of impractical calculations. Since asteroids are clustered in several orbits, the algorithm must consider “only” a few thousand carefully selected possibilities.
However, calculating thousands of test orbits for thousands of potential asteroids is a daunting task. However, the advent of cloud computing makes this possible. According to the scientists involved, Google helped optimize time using its Google Cloud Platform.
This is just the beginning
Currently, the algorithm is set to search only major-belt asteroids, those that orbit between Mars and Jupiter, and not near Earth asteroids that may have collided with our planet.
Identifying near-Earth asteroids is more difficult because they move faster, and the algorithm needs to do more division calculations to make time and distance connections, Lu says.
The NOIRLab file contains seven years of data, indicating that there are tens of thousands of asteroids located. So far, just over 12% of the data from one month, September 2013, has been analyzed, in which THOR has found 1,354 possible asteroids.