As I said, I remain highly skeptical that curvature measurements can be made that precise. If onboard video is really all one has to work with, my approach would be to measure the angular size of ground features directly below and comparing with simulations for various altitudes. Much more sensitive to altitude than fuzzy horizon curvature.
But above all, probably integrating accelerometer data and trying to show that in the worst case of it overestimating acceleration and underestimating air drag deceleration would still produce over 100,000 ft. Their accelerometer data quality looks good so the only question is if one accepts it as legitimate.
Quote from: ugordan on 10/09/2011 02:34 pmAs I said, I remain highly skeptical that curvature measurements can be made that precise. If onboard video is really all one has to work with, my approach would be to measure the angular size of ground features directly below and comparing with simulations for various altitudes. Much more sensitive to altitude than fuzzy horizon curvature. Hmmm, that is perhaps a better approach. Still, I think getting an estimate +/- 20% of altitude is feasible.
(Though your method may be better... it'd also be a little more complicated since you are proving you're x meters away from feature A, but that's not enough... you also have to show you're y meters from feature B and probably z meters from feature C for good measure... all at about the same time.)
QuoteBut above all, probably integrating accelerometer data and trying to show that in the worst case of it overestimating acceleration and underestimating air drag deceleration would still produce over 100,000 ft. Their accelerometer data quality looks good so the only question is if one accepts it as legitimate.Accelerometer data can be screwed up by vibration, bias, etc. Since you're integrating a very small number (twice, actually...), a tiny amount of bias can add up to a huge difference in altitude.
...Quote from: Robotbeat on 10/10/2011 12:43 am(Though your method may be better... it'd also be a little more complicated since you are proving you're x meters away from feature A, but that's not enough... you also have to show you're y meters from feature B and probably z meters from feature C for good measure... all at about the same time.)Which can be "easily" done by reprojecting the landscape below to different projections with different altitude and comparing to the one the camera sees (linearized, of course) and using a best fit approach.The size of features below is very sensitive to altitude difference, the curvature of the horizon much less so (when delta h << R).
Quote from: Robotbeat on 10/10/2011 12:43 amQuoteBut above all, probably integrating accelerometer data and trying to show that in the worst case of it overestimating acceleration and underestimating air drag deceleration would still produce over 100,000 ft. Their accelerometer data quality looks good so the only question is if one accepts it as legitimate.Accelerometer data can be screwed up by vibration, bias, etc. Since you're integrating a very small number (twice, actually...), a tiny amount of bias can add up to a huge difference in altitude.15G acceleration is not really a small number. I said the *worst case* scenario should be used where it is assumed all accelerometer outputs are overestimates of actual acceleration due to measurement artifacts. There are finite error bars to any measurement.
Recent update on this from aRocket, prize is still available, most announced attempts seem to be aiming for launching at BALLS in September. I wonder what will happen if there are multiple successful ones ? ( highly doubt that's likely though )I think someone should simply reinstate CATS prize with perhaps Kickstarter funding at some point.
On September 11, 2012, a team of experienced hobbyists reached a milestone on the journey toward amateur space flight: a GPS documented flight to 104k’ AGL with full recovery on only 21000 N*s of propellant. This report reviews that project.
See http://anewdomain.net/2013/05/20/carmack-prize-video-aeropac/The winner's project page has disappeared by archive.org still has it:https://web.archive.org/web/20130514184403/http://www.aeropac.org/100k/It includes the text:QuoteOn September 11, 2012, a team of experienced hobbyists reached a milestone on the journey toward amateur space flight: a GPS documented flight to 104k’ AGL with full recovery on only 21000 N*s of propellant. This report reviews that project.I think the announcement of John paying up was on the Armadillo website, which is also down now (it's like it's 1998 around here.)
