Slashdot Mirror

← Back to Stories (view on slashdot.org)

NASA Offers Bounty For Improved Asteroid Detection Algorithms

Posted by Unknown on from the all-power-to-sensors dept.

Hugh Pickens DOT Com writes "Dara Kerr reports at CNET that NASA is launching an 'Asteroid Data Hunter' contest to inspire the creation of algorithms that identify asteroids in images captured by ground-based telescopes. ... The winning solution must increase the detection sensitivity, minimize the number of false positives, ignore imperfections in the data, and run effectively on all computer systems. 'Current asteroid detection initiatives are only tracking one percent of the estimated objects that orbit the Sun,' says Chris Lewicki. 'We are excited to partner with NASA in this contest to help increase the quantity and knowledge about asteroids that are potential threats, human destinations, or resource rich.' NASA's goal is to discover those unknown asteroids and then track and characterize them. For the contest, citizen scientists will be allowed to study images taken from ground-based telescopes to see if they can develop improved algorithms for identifying asteroids. If dangerous asteroids are found, NASA could determine if they'd be viable for a re-direction into a lunar orbit. 'For the past three years, NASA has been learning and advancing the ability to leverage distributed algorithm and coding skills through the NASA Tournament Lab to solve tough problems,' said Jason Crusan, NASA Tournament Lab director. 'We are now applying our experience with algorithm contests to helping protect the planet from asteroid threats through image analysis.'"

38 comments

  1. $35,000 split x ways? by Anonymous Coward · · Score: 2, Insightful

    > NASA will give away $35,000 in awards to competition winners

    Big spenders! This really motivates me to get started.

    1. Re:$35,000 split x ways? by something_wicked_thi · · Score: 5, Funny

      I think they're spending too much. Everyone knows this is just another trumped up threat by scientists to get government to pay for scientists' extravagant lifestyles. But asteroids are by no means settled science. Just because 99% of astronomers agree that a large asteroid hitting the earth would result in devastating consequences for the human race, I know this chemist who saw an asteroid enter the earth's atmosphere and burn up harmlessly. Why, just the other day, I saw one shoot across the sky harmlessly.

      And don't try to tell me the dinosaurs were wiped out by an asteroid. Everyone knows that volcanoes kill more dinosaurs every year than all the asteroids combined.

      These asteroid alarmists need to get a life and calm down.

    2. Re:$35,000 split x ways? by Anonymous Coward · · Score: 0

      Fear of asteroids (or alien invasion) is the next logical step after this whole fear of terrorism thing. Just another way to suck money from taxpayers.

  2. Am I the only one... ? by Noryungi · · Score: 3, Insightful

    ... Who thinks this whole article is written like a freaking marketing PR announcement?

    I mean: "We are excited to partner with NASA" [...] "NASA has been learning and advancing the ability to leverage distributed algorithm and coding skills" [...] etc.

    Don't misunderstand me: the idea is great and, if they can detect more asteroids, faster, and with a better precision, we will all be better off in the long term. But I am just tired of these shockingly stupid buzzwords ("excited", "advancing", "leveraging", "coding skills", yadda, yadda, yadda).

    And get off my lawn!

    --
    The right to offend is far more important than the right not to be offended. (Rowan Atkinson)
    1. Re:Am I the only one... ? by wonkey_monkey · · Score: 1

      And get off my lawn!

      Vacate my grass-encoated outer-building frontage!

      --
      systemd is Roko's Basilisk.
  3. Stop crowdsourcing work for free by Anonymous Coward · · Score: 2, Interesting

    Find intelligent people and HIRE THEM to come up with algorithms.

    1. Re:Stop crowdsourcing work for free by Anonymous Coward · · Score: 0

      $35,000 payout. Similar to a professional poker players life.

    2. Re:Stop crowdsourcing work for free by plover · · Score: 4, Insightful

      I think they have intelligent people. What they're looking for is some outside perspective.

      When you've been staring at your own solution for years and years, it's good to have someone make you question it once in a while. They will no doubt get plenty of rookie and novice suggestions, the easy stuff they've long ago solved. They may even get some of the suggestions that took them a long time to understand and develop. What they're hoping for is something completely different. Maybe some grad student working on a new form of video compression will spot similarities that can be applied. Who knows?

      --
      John
    3. Re:Stop crowdsourcing work for free by Anonymous Coward · · Score: 0

      So basically, crowdsourcing work for free

    4. Re:Stop crowdsourcing work for free by Anonymous Coward · · Score: 0

      But they do have intelligent experts, in the field of algorithm contest development. It's like hiring an ambulance driver instead of a doctor for every street corner. He doesn't solve the problem, but he can take the problem to the people who can solve it.

    5. Re:Stop crowdsourcing work for free by Anonymous Coward · · Score: 0

      Most who have the expertise in the field of algorithm development or signal processing are smart enough to realize a poker player's life of algorithm development is not going to work out as a career. That being said, it's a fine hobby if you treat it that way. If I'm looking around for an interesting project to stave off boredom, a $35,000 reward for success would make me a lot more likely to pick your project.

    6. Re:Stop crowdsourcing work for free by Anonymous Coward · · Score: 0

      But the contest format is stupid. You have to join up to some stupid bounty-hunter website, and then they'll drip-feed you data, and make you compete for small prizes ($500-$1000) on each of several arbitrarily-sized 4 day tasks.

      They could just put the data up, and then say they'll pay $35,000 in 6 months for the solution that works best on some data they didn't put up. Now that would be interesting.

      Instead they've done it in some stupid corporate way - which is, of course, the exact way in which they've failed to solve this problem in the past.

    7. Re:Stop crowdsourcing work for free by Anonymous Coward · · Score: 0

      I would much have preferred the several month timeline instead of a bunch of small tasks with deadlines. I have a day job that involves similar algorithms and wouldn't mind messing around with a different situation and a chance at some extra money. But I can't rearrange my schedule and have things done every couple of days for a side project.

    8. Re:Stop crowdsourcing work for free by Anonymous Coward · · Score: 0

      I agree, that's really dumb. I really wanted that $35k, and I believe I could have a good crack at it, but with this structure I'm just not interested any more.

      And, http://www.topcoder.com/astero... is not available at the moment. I guess it's been slashdotted.

  4. It's not like it's brain surgery... by Anonymous Coward · · Score: 0

    Oh.. wait
    https://www.youtube.com/watch?v=iGCMtk695Cg

  5. Wheres the data by Anonymous Coward · · Score: 0

    If NASA wants an even playing field for asteroid detection, then wheres the data on current comet positions and the asteroid belts?
    It seems like a big data job using parallel processing and supercomputers etc.

  6. Additional superbounty : $ 10 million by invictusvoyd · · Score: 1

    * Determine if the asteroid's orbit collides with earth's
    * Launch a nuclear tipped rocket to disintegrate or deflect the asteroid

    IF not ,

    * Check if there are any valuable minerals on the asteroid .
    * Mine and launch the vehicle back to earth

  7. Trajectory by Anonymous Coward · · Score: 0

    "If dangerous asteroids are found, NASA could determine if they'd be viable for a re-direction into a lunar orbit."

    Or at least making sure they land in Russia.

    1. Re:Trajectory by wagnerrp · · Score: 1

      If an asteroid is found that is legitimately dangerous, we don't have the technology to capture it in Lunar orbit.

    2. Re:Trajectory by cavreader · · Score: 1

      Using a legitimately dangerous asteroid will increase the motivation factor and throw budget restrictions out the window. Then we would see whether what we are truly capable of when it comes to technology development.

    3. Re:Trajectory by wagnerrp · · Score: 2

      There's a big difference between slightly changing an orbit so it does not kill us all, and performing a Lunar entry burn for shits and giggles. The asteroid that detonated over Russia last year was around 12Gg, with a velocity of roughly 19km/s. The detonation as only about half a MT, so even a direct impact would not have been that bad outside the local region. It's a long way off from cataclysmic.

      So let's intercept this thing and slow it down to Lunar orbit. Assuming a good chemical rocket at 450s... no, lets go all out and give it something nuclear at ~1000s, you're looking at some 70Gg in fuel. Now remember, you actually have to intercept it with that much fuel first. Even assuming you get lucky with several years of preparation and a fortuitous gravitational assist, you're still looking at a good 25km/s. You're up close to a full Tg in launch mass. The Saturn V was only around 3Gg. If you used that more traditional chemical rocket, you would be looking at 100x that requirement. Rocketry is very sensitive to specific impulse.

      The requirements to capture even a small asteroid are just so absolutely staggering, trying to do so with an extinction-scale asteroid is not even remotely in the realm of possibilities, regardless of how much money we threw at it.

  8. Before or after? by Dan+East · · Score: 5, Funny

    Does the algorithm need to detect asteroids before or after they impact the earth?

    --
    Better known as 318230.
    1. Re:Before or after? by pr0fessor · · Score: 1

      Why am I imagining this as your answer?

      if(AstroidBoundingSphere.Intersects(EarthBoundingSphere)) LaunchCounterMeasures();

  9. The prize is just $35K by Anonymous Coward · · Score: 0

    Which is *much* cheaper than hiring people...

  10. I know!!! by Anonymous Coward · · Score: 0

    Use Kerbal Space Program to calculate it.

  11. Maybe they would... by Anonymous Coward · · Score: 0

    ...if they had the budget to do so.

  12. so many haters...on the humanity by schlachter · · Score: 2

    think about it, this is possibly one of the most important things that we as a society can do...protect our selves from the large number of inevitable asteroid strikes which will wipe out millions of people instantly.

    now it's being crowdsourced to the world, as it should be, and people bitch and moan. go save the world if you're the type and don't bitch about $35K not being enough to motivate you to save the world if you're a selfish jackass.

    --
    My God can beat up your God. Just kidding...don't take offense. I know there's no God.
    1. Re:so many haters...on the humanity by Anonymous Coward · · Score: 0

      think about it, this is possibly one of the most important things that we as a society can do...protect our selves from the large number of inevitable asteroid strikes which will wipe out millions of people instantly.

      now it's being crowdsourced to the world, as it should be, and people bitch and moan. go save the world if you're the type and don't bitch about $35K not being enough to motivate you to save the world if you're a selfish jackass.

      You may be young enough to not remember when people were actually paid to do science... $35k is not enough money, so it's insulting. Better to have it be 'for fun' or a student contest which should award the prize as a scholarship to a single person.

    2. Re:so many haters...on the humanity by number17 · · Score: 1

      think about it, this is possibly one of the most important things that we as a society can do...protect our selves from the large number of inevitable asteroid strikes which will wipe out millions of people instantly.

      My car analogy is that this project is improving the detection of tractor trailers hitting park cars. Call me when the car can do something about it.

    3. Re:so many haters...on the humanity by schlachter · · Score: 1

      Posting as AC huh? Well, I am paid to do science. And I find this challenge pretty exciting. I might work it in my free time. I might pair up with a few other scientists who are interested in doing the same. Or if I really want to work this full time, I might write up a grant so I do get "paid". I don't really care about the $35K. And it's not pay. It's just incentive. Like a reward.

      --
      My God can beat up your God. Just kidding...don't take offense. I know there's no God.
    4. Re:so many haters...on the humanity by Anonymous Coward · · Score: 0

      Different AC, but someone else who is paid to do science for a day job. While I would be interested in the big picture goal of the contest, the format makes it nearly untenable for me since it is broken up into a bunch of one week or so small pieces with small cash prizes each. Scheduling around that when some weeks involve a lot more work than others at my day job makes it near impossible unless I dropped about every other priority, and did it on my own without trying to line up schedules with friends and coworkers.

  13. For reference, here's one of the current systems by StupendousMan · · Score: 4, Informative

    If you're interested in the current state of the art, read this article from the Publications of the Astronomical Society of the Pacific (April 2013). It describes the hardware and software used by the Pan-STARRS team to detect asteroids automatically in data taken with their 1.8-meter telescope on Hawaii and its 1.4-gigapixel CCD camera.

    http://arxiv.org/abs/1302.7281

    --
    Michael Richmond "This is the heart that broke my finger."
    mwrsps@rit.edu http://stupendous.rit.edu
  14. Lunar orbit? by braindrainbahrain · · Score: 1

    Why move a dangerous asteroid into lunar orbit? That's a little too close to home should something go wrong.
    If they want to study it, just send a probe to bring back pieces of it.

    1. Re:Lunar orbit? by neo-mkrey · · Score: 1

      I was thinking the same thing, except I don't even want them that close. Fling them towards the sun and be done with it.

  15. Would not want that responsibility, thank you by FilmedInNoir · · Score: 1

    So I created this algorithm for NASA, and ummm, I guess I missed a decimal point somewhere cause it didn't pick up this big asteroid heading towards us.
    Anyway... probably want to spend some time with your loved ones today. I'm going to go empty a liquor store myself.

    --
    Sig. Sig. Sputnik
  16. Re:For reference, here's one of the current system by abramjablonski · · Score: 2

    MOPS is a very cool system and seems to be well-engineered, but it isn't quite state of the art - according to the performance numbers in section 3.12, it takes months to run simulations (with 100k or so objects), and it's a pipelined system (using Linux clusters), so there are multiple fairly beefy machines crunching the numbers. I know an astrophysicist that has a program that can do that in a night, on his laptop (not source, but... http://derastrodynamics.com/do...). That being said, MOPs is (AFAIK) free, has data interchange/ API capabilities, and is in production, while Dr. Der wants to get paid really well for his algorithms - slightly more than the $35k that NASA's offering :D

  17. I have a simple Solution. by Anonymous Coward · · Score: 0

    When it hits earth. it is detected simple. and zero chance of failure