Asteroid 2018 CB is set to , around 63,000 kilometres above our heads. But if a deadly asteroid was set to hit the planet, humanity would have to scramble to conduct the world鈥檚 riskiest experiment. Since we have never attempted to shove a space rock off its course, any defence efforts would be an untested prospect.
It鈥檚 not that people haven鈥檛 come up with ideas 鈥� like chucking a whole bunch of nuclear weapons at an incoming threat 鈥� but with limited budgets devoted to asteroid defence, which should we pursue? A team of researchers have come up with an algorithm called the Deflector Selector to help us decide.
To be clear, , formerly at Carnegie Institution for Science at Washington, D.C, and her team are not suggesting we give a computer free reign over a bunch of nuclear missiles 鈥� after all, they鈥檝e seen the movies. 鈥淲e are absolutely not advocating putting the algorithm in charge of asteroid defence,鈥� she says.
Advertisement
Instead, their machine learning algorithm can be used to study a given population of potentially hazardous objects, then determine which technology has the best chance of deflecting them from Earth鈥檚 path.
Boom, smash or tug
To build their algorithm, Nesvold and colleagues created a simulation of six million hypothetical objects with the potential to hit the Earth. For each, they simulated how far in advance of collision it could be detected, and the velocity change needed to knock it off course. With this information, they looked at the effects of three technologies: nuclear weapons, kinetic impactors, and gravity tractors.
While nuclear weapons release an explosive force, a kinetic impactor is an attempt to essentially shoot a bullet out of the air with a much smaller bullet. Gravity tractors are more subtle. The idea here is to hover a spacecraft near an asteroid, and let its gravitational pull ever so slowly tug the space rock off course. The concept has never been tested, and is probably the least well-developed of the three ideas.
The team simulated the capabilities of each of these three technologies, assuming they were launched on a Delta IV Heavy rocket 鈥� the most powerful in operation, until the recent launch of SpaceX鈥檚 Falcon Heavy.
Quick decisions
All this number crunching required a cluster of 100 computer cores running for around 40 hours, which is why the team then turned to machine learning. Using the simulation as training data, they taught an algorithm to study a given population of objects and then decide which deflection tech had the best chance of success.
鈥淭he benefit of the machine learning algorithm is that, once trained, it can give us an answer in seconds instead of hours,鈥� says Nesvold.
The team tried their trained algorithm on three different populations: hazardous near-Earth asteroids, comets, and rubble piles 鈥� loose collections of material, rather than solid objects. In each case, nuclear weapons could tackle about 50 per cent of the objects. Kinetic impactors and gravity tractors had lower success rates.
That might suggest we should invest in nuclear weapons as our asteroid defence of choice, but clearly there are sensitive issues when it comes to mounting these on rockets. 鈥淭he risks created by pursuing a nuclear asteroid defence programme would far outweigh the actual risks posed by asteroid impacts,鈥� says , who heads the Catalina Sky Survey, an asteroid survey programme that discovered 2018 CB earlier this week.
As an alternative, Nesvold points out that the gravity tractor success rates improve when they have longer lead times to work with. 鈥淭hat points to the need for better asteroid monitoring and detection surveys,鈥� she says. 鈥淓arlier detection of potential impactors will provide more lead time to plan and execute a mission,鈥� agrees Christensen.
Reference:
Read more: We got a good look at the interstellar asteroid and it鈥檚 weird
