Post on 21-Apr-2017
transcript
It's not often that NASA asks you to come to Silicon Valley to save the world using AI…
But that’s exactly what happened last Summer to 12 researchers from around the world.
ABOUT FDL
GOAL: CLOSE KNOWLEDGE GAPS BY MATCHING DOCTORAL-LEVEL TALENT FROM THE PLANETARY SCIENCES WITH PEERS FROM
THE MACHINE LEARNING COMMUNITY.
Asteroids are a statistically significant threat. We get hit by a large object capable of loss of life every 50-100 years, on average.
3)’What is the best choice of technology to make a successful deflection of an Asteroid that poses a threat to Earth?’
KEY RESULTS (1)
“Only 31 meteorites have ever been found that can be linked to a source orbit”
‘What is it made of?’
KEY RESULTS (1)
•Meteorite searches in target regions can require hundreds of hours to possibly find one meteorite.
•The specific material and structure of fresh meteorites, linked to source orbits, make them nearly as valuable as asteroid sample returns, such as OSIRIS-REx.
‘What is it made of?’
KEY RESULTS (1)
Meteor Rapid Response Drone. •The team built 25,000 training images of meteorites
•six deep learning models, along with a 15 million image library.
•Result: an automatic meteorite detection system, all driven by a user-friendly app for use in the field.
This approach is close to providing a rapid response system to search a target area within days of a fresh meteorite fall.
•Asteroid shapes are critical for asteroid deflection techniques - as any mitigation plan needs to know the center of mass.
•Should an object be too close to shift, shape is critical for understanding the potential for damage and planning effective disaster response.
KEY RESULTS (2)
KEY RESULTS (2)KEY RESULTS (2)
FOUR WEEKS
‘What shape is it?’
This is currently a laborious manual process that takes a trained practitioner around four weeks.
Enhanced Shape Modeling of NEOs
KEY RESULTS (2)
“They reduced the search for the asteroid spin axis and shape to a few hours of computing, achieving better results than 4 weeks of work by one of the world’s experts.”
KEY RESULTS (2)
• Used a 3D-VAE (Variational Auto-encoder) to generate 3D voxel shapes of asteroids conditioned by the radar images (as input).
• This approach may enable a rapid understanding of the shape of an asteroid - while it is still being tracked by radar, even as an incoming object.
• Automatically discover the spin axis angles
3)’What is the best choice of technology to make a successful deflection of an Asteroid that poses a threat to Earth?’
KEY RESULTS (3)
‘What is the best choice of technology to make a successful deflection of Asteroid that poses a threat to Earth?’
NuclearDevice
?
KEY RESULTS (3)
“The Deflector Selector”
A TRAINING SET BASED ON 1.5 MILLION ORBITAL SIMULATIONS OF
THREE DIFFERENT KINDS OF DEFLECTION TECHNOLOGY.
“There isn’t a tool of this sophistication available to the Planetary Defense community.”Astronomer, JL Galache from the IAU’s Minor Planet Center (& FDL Mentor)
KEY RESULTS (3)
1.5 million orbital simulations were used train a decision tree to select a set of effective technologies (Nuclear Device, Gravity Tractor or Kinetic Impactor) for a given hazardous object.
Once trained, it had an accuracy of 98% for determining which technology would produce a successful deflection.
A Machine Learning Decision Tree…
KEY RESULTS (3)
•The most effective technology predicted by the decision tree is the nuclear explosive, due to the high ΔVs it can impart and its instantaneous effect.
•Future improvements to the model will use simulated hazardous object populations to further improve accuracy. Nuclear
Device
KEY RESULTS (3)
•This important work will help inform strategic decisions which deflection technologies should be prioritized, and what asteroid characteristics are the most important to be known in advance of taking action.
THANK YOU
www.frontierdevelopmentlab.orgjames@frontierdevelopmentlab.org