Gravity Tractor Could Deflect Asteroids

Hugh Pickens writes "A new study at the Jet Propulsion Labs shows that weak gravitational pull of a "gravity tractor" could deflect an Earth-threatening asteroid if it was deployed when the asteroid was at least one orbit away from potential impact with Earth. First a spacecraft would be crashed directly into the asteroid, similar to the Deep Impact mission that impacted a comet in 2005. This would provide a big change of direction, but in a less controllable fashion that could push the path of the asteroid into a dangerous keyhole. But then a second spacecraft, the gravity tractor, would come into play, hovering about 150 meters away from the asteroid, to exert a gentle gravitational force, changing the asteroid's velocity by only 0.22 microns per second each day. Over a long enough time, that could steer it away from the keyhole. In the simulation, a simple control system kept the spacecraft in position, and a transponder on the asteroid helped monitor its position and thus determine its trajectory more precisely than would be possible otherwise. 'The gravity tractor is a wimp, but it's a precise wimp,' said astronaut Jack Schweickart. 'It can make very small, precise changes in orbit, and that's what you need to avoid a keyhole.'"

Comments

Re:If they ever do this...

[4D6963]

I hope their simulations use doubles, not floats!

I know you're joking but for just the speed values if they used time increments in the order of the second then the speed differences would be in the order of e-18, which is too small for a double's mantissa. I'd rather go with long doubles, or better (I think you can achieve something like that by using a number to store the closest representable value and another one to represent the tiny difference from what it should be).

Re:I hate to accuse JPL of forgetting something...

[trongey]

Who modded the parent "insightful". The answer is pretty simple, and is even illustrated in the article. The picture shows a craft with three thrusters all angled away from the asteroid. The resulting thrust is a vector normal to the target. Sure, it sacrifices efficiency, but it works.

Compensated summation

[DrYak]

(I think you can achieve something like that by using a number to store the closest representable value and another one to represent the tiny difference from what it should be).

Yup. It's the Kahan summation algorithm. It works as you describe it and it used to compensate the error that happens when doing very big sums of very small numbers (exactly the situation in the gravity tractor's problem)