MONTREAL – At age 13, Artash Nath is already taking on the big questions puzzling scientists – climate change, planetary atmospheres and asteroid collisions.
Artash and his sister Arushi, who both live in the Toronto area, have made over 25 space projects since 2014. Back then, their project “Curious Bot” – an autonomous Mars rover built with Arduino – received one of the top five NASA Space Apps awards. They’ve picked up multiple awards for their efforts ever since. At the Montreal Space Symposium Thursday (Oct. 10), Artash discussed a project related to planetary defense.
Finding asteroids in space is no small problem. While NASA has found no imminent threats to Earth, the agency and many partner observatories remain on the hunt to continue tracking down members of the asteroid population.
Detecting an asteroid is a difficult challenge in itself, as they are small and (because many asteroids are carbonaceous) tend to be dark, reflecting little light for telescopes to detect. Once an asteroid is found, telescopes must track the path for enough time to generate an accurate orbit. And even once the orbit is known, the path can change due to outgassing on the asteroid, or interactions with other planets or small worlds.
Artash is interested in predicting the risk index of collision with near-Earth objects (NEO), particularly the class of comets and asteroids known as potentially hazardous objects. These worlds have a minimum orbit intersection distance with Earth of only 0.05 astronomical units (or sun-Earth distances). It’s the equivalent distance of 20 times beyond the orbit of the moon, which is close in astronomical terms.
The community tends to use two indexes to convey the risk of asteroid collisions: the Torino Scale (with a risk rated between 1 at the low end and 10 at the high end) and the logarithmic Palermo Technical Impact Hazard Scale used by NEO specialists, based on factors such as impact energy and probability.
Artash proposed using supervised machine learning models to go through large datasets of asteroids to better characterize which might be the most hazardous. In simple terms, the model uses a “neural network” – a series of calculation steps modelled on how the human brain works – to classify asteroids and encode patterns.
To do the calculations, the computer is first given a training dataset to learn how to perform the various steps. It then performs calculations autonomously using another dataset. The computer outputs the results, which are then reviewed by the programmer for accuracy and for the possibilities of making adjustments to the program inputs in future iterations.
Artash ran his own calculations using NASA Jet Propulsion Laboratory’s Sentry system, an automated collision monitoring system that regularly plumbs the catalogue of asteroids to figure out which ones are most likely to hit Earth in the next 100 years. He chose 942 members of this dataset to train the neural network on what parameters to look for.
Then to test the ability of the computer to train itself, he deployed the program in a dataset maintained by JPL’s Center for Near Earth Object Studies, which computes asteroid and comet orbits along with their chances of colliding with Earth.
Artash’s work showed that the collision risk of known asteroids is low. Of the NEOs that are classified as potentially hazardous, 90 percent of them have a collision risk that is 10 to the third less likely than a random background event.
But he warned that more research into asteroids is needed, especially in considering the geopolitical implications of whether countries have the right to use nuclear weapons in space to defend themselves, and figuring out which countries should be responsible for funding a mission.
There are several missions (either ongoing or planned) that directly relate to NASA’s understanding of asteroids to feed into planetary defense scenarios. One example is OSIRIS-REx (Origins, Spectral Interpretation, Resource Identification, Security, Regolith Explorer). It’s a mission at asteroid Bennu that includes a Canadian laser instrument to measure topography; the spacecraft will eventually return a sample to Earth for analysis.
Later in the 2020s, NASA plans to launch an asteroid impactor called DART (Double Asteroid Redirection Test). This will crash an impactor spacecraft into an asteroid moon to test how well current technology can deflect a threatening world incoming to Earth. The European Space Agency will examine the effects up close in a follow-up mission called Hera, while ground telescopes around the world would also monitor the orbit of the asteroid moon to see the results.