2001 fiction set to be science fact? NASA boffin mulls artificial intelligence to watch over the lunar Gateway If humans are to go beyond the Moon, they must rely less on ground control and more on AI systems to perform operations such as flying, and conduct scientific experiments more autonomously, according to a NASA paper [PDF] out this week. "Future missions to Mars will require crew and ground to operate independently from Earth- …
The Lunar Gateway will be largely made up of four components: a habitat for the explorers to live, an airlock system to let them in an out of the spacecraft, a power propulsion vessel to travel to the Moon, and a logistics hub to accept cargo shipments.
So, it's going to have Pod Bay Doors...what could possibly go wrong right?
@Stop_Forth
But they DID put a handle on the outside (of Discovery) - that's how Dave Bowman was able to re-enter after HAL refused to open the pod bay door (using the pods "extendable" arm to turn the handle to open the access door).
What would have been better was if Bowman had taken his helmet with him, when he took the pod to try and save Frank Poole. Health & Safety would take a dim view of his actions.
Yes, it's from Heinlein's novel. When I first started looking for it (around 1969) there was no Internet! It wasn't in any printed dictionary. My father (an engineer) had never heard the word, neither had anyone else. I haven't been looking for it every day of the last fifty years, but it has been bugging me that Clarke used a word that I didn't know and have never seen used anywhere else. Anthony Burgess works are full of words I don't know, Sci-Fi not so much. I have read some of Heinlein's stuff but not that particular work, which I was unaware of before today.
Poole wasn't really thinking at all how about how he could rescue Bowman & bring him inside the pod or go outside & do some form of "buddy breathing" without his helmet.
NFC - Helmet detection system that prevents the pod being launched without it (Maybe not so hot an idea if needing to use the pod as a emergency lifeboat).
Anyone else think the Discovery suit helmets look like a kids zippy or dinosaur helmet with stick on eyes (My sister had something like that as a kid).
https://vignette.wikia.nocookie.net/2001/images/0/07/5398A480-53F9-4F5C-9358-97B29FB5040D.gif/revision/latest?cb=20190116072934
well, the dependency on a single space suit's systems NOT to fail is always a bad idea.
Here's what I came up with: emergency space suits that are, for all practical purposes, balloons with straps that you tighten in specific places to make them somewhat manageable.
So here's what you do: your space craft or station has an explosive decompression event. You have about 30 seconds to put one of these on, +/- whatever condition you find yourself in, bleeding ears/eyes and so on. So you go to the wall where one is stowed, rip open the package [large handle, bright colored paint], step into this thing, and zip it up, then pull the O2 lever and it inflates.
Then you adjust a few straps around the legs, your torso, arms, and whatnot. VOILA, instant space suit!
I bet you could make them for under $1k and they'd be 100% reliable. Then you make them MANDATORY on ALL space flights, like O2 masks on airlines.
What's really needed are TWO "Gateways" - one in Earth orbit and the second in Lunar orbit.
Smaller rockets (taking up maybe Crew Dragon or even Soyuz capsules) can then be used to transport the humans up to the Earth Gateway, where they can then switch to another rocket which just travels from Earth to Moon (and hence Lunar Gateway) and back again.
Experiment equipment, and cargo (inc food, water, oxygen etc) can be sent to Lunar Gateway in advance of crew arrivals.
And a version of Crew Dragon, with some added rocket power, could be used as the Lunar Lander, as with 1/6th the gravity, a small re-useable craft that just needs to carry humans could be used easily enough.
Because the ISS is already in a high-inclination orbit (because Baikonur, which is fairly far north), poorly suited for trans-lunar injection. It'd cost a bit more delta-V both for a Florida/Texas launch to match ISS orbit, and to lose the north/south momentum to put you on course for the Moon. Although, now that I think about it, twice a month, the orbit would line up with the Moon's so that you could TLI from the northernmost or southernmost point in the orbit, and not be too wasteful in that regard. But there would still be the launch cost.
The current plan plan is to launch 4 astronauts in an SLS+Orion to LOPG. The SLS is between $1B and $2B to launch (+ the cost of an Orion). Everything else can go to LOPG on a commercial rocket (a Falcon Heavy launch costs $0.09B) but Orion was designed from the start to require SLS. Once at LOPG crew can transfer to a transfer + descent + ascent vehicle, go to low lunar orbit. Two astronauts take the descent + ascent vehicles to the Moon for seven days. They take the ascent vehicle back to the transfer vehicle and the transfer vehicle back to LOPG then everyone + 0.1T of cargo goes back to Earth in the Orion. The idea was to make the transfer vehicle reusable but that idea would be a challenge with the 2024 deadline. Orions should be able to make two trips but they have a hefty refurbishment cost. Throwing an SLS in the sea is so expensive that no-one expects enough budget to do that more than once per year.
A different plan is to launch a Starship on a Superheavy to Low Earth Orbit. The cost is expected to be under $10M and you get the Superheavy back. Next send up the tanker version of Starship on a Superheavy, refuel the first Starship and return the Superheavy+Tanker to Earth for a cost of another $10M. Repeat about 4 more times until the first Starship is full of fuel. Move the Starship to an elliptical orbit and do another fuel run then the Starship can go to the Moon, land, stay (probably until sunset), take off and return to Earth ready for another trip. Starship is intended to carry 150T up and 50T down (including crew). Probably a bit less on a Moon run as you cannot make methane fuel there unlike Mars.
Orion-like test articles have done a pad abort, gone to orbit on a Delta IV Heavy and returned to Earth and launched on mini rocket for an in-flight abort test. SLS is supposed to have its first test flight no earlier than November 2020. A prototype Starship might fly to 2km this year. I am not sure which launch date is less believable.
Stopping at LOPG costs fuel. If SLS could take Orion to low lunar orbit that is where LOPG would be. Instead it orbits around L1 (and will make astronauts really regret missing the launch window from the moon). Stopping at LEO on the way back is not an obvious choice. You can use loads of propellant or for much less mass you can use a heat shield and aerobraking. If you have a heat shield you can probably go to Earth more easily than LEO. (Crew Dragon's parachute testing has demonstrated the mathematical models for parachutes are badly wrong. This should be a problem for Orion but Starship and Superheavy do not use parachutes.)
One LOPG is a really expensive regular waste of propellant. No-one has had the nerve to request extra budget for second one.
No matter how many times I read "Orion", I can't get the Project Orion spaceship out of my head. I just expect it to be a massive, nuclear-power juggernaut.
"An Earth to Moon transfer and Moon to Earth transfer are very small amounts of Delta-V compared to getting into orbit."
Yes, but you have to get that fuel up there in the first place. The tyranny of the rocket equation and all that jazz.
If we could make the fuel on the Moon, that makes things interesting. It also influenced the fuel choice for Starship (for Mars)
"What's really needed are TWO "Gateways" - one in Earth orbit and the second in Lunar orbit."
I was thinking that too. Since it looks decidedly as though NASA are adding more and more bells and whistles onto a design intended to be launched and in use within 5 years, it seems to me as though they are deliberately dragging their heels to make sure the 2024 deadline can't be made, probably on the assumption that it's an unrealistic and politically motivated goal leading to failure, so might as well "shoot for the moon", eh?
I was asking myself the same question - they're talking about programmed automated control systems, not AI where you try to teach an initially "dumb" system to program itself from experience.
Programs can (and often) present unexpected outputs, because the underlying logic isn't correct and the particular combination and/or sequence of inputs received wasn't anticipated. AI won't protect against the unexpected and, when that occurs, it's probably harder to address because the actual logic is only known within the computer. Knowing the logic and the NI (Naturally Intelligent) meat (in situ or back on the ground) has a chance to manage the unexpected. HAL was fiction but, like a lot of ACC's ideas, was well thought through and should serve as a cautionary tale.
But, as you said, Buzzword bingo gets more funding dollars...
PS I had the same physics teacher as ACC, my claim to fame (along with several hundred other boys who were taught at that school over several decades).
I was asking myself the same question - they're talking about programmed automated control systems, not AI where you try to teach an initially "dumb" system to program itself from experience. ..... Struggling to find a different name
Is not AI currently presently trying to teach an initially "dumb" system to program itself from experience in its dealings with humans?
Do you suppose that to be both practically and virtually impossible rather than most likely probably certain.
It is surely impossible to deny now that machinery conveys and instructs humans to enact in actions for media presentation/universal advertising ...... which are further machines energised into Support ACTivIT with NEUKlearer HyperRadioProACTive AI and IT Systems.
For a Fundamentally Different Perspective with Views Extraordinarily Rendered via All the Help Provided from Future Stored Provisions in the Virtual Nature of United Stately Matters.
I Kid U Not ....... and do you know where you are then? And what needs next doing? :-)
An AI finds a piece of text which most fully answers a question, in contexts and subtexts of both the question and the piece. Looking for the piece the AI establishes its cause-and-effect relationships, selecting one that best fits the question. And finally in the piece the AI finds the answer, as a phrase, sentence or paragraph, modifying it appropriately.
If the AI can't find the right piece of text,it tries to detect it wherever it can, which is called Machine Learning.
Howdy, Il.Gel.
You might like to realise that things have moved on quite surprisingly and into SMARTR Machine Learning Fieldcraft.
Nice first post, though. There's a lot to learn from El Regers here who just keep on giving. And things are getting real lively in these and those aforementioned fields above.
Indeed. There's nothing in those requirements that looks impossible to do without AI, and do bear in mind that:
* All the Apollo computational heavy lifting was done in Houston because the onboard computational power in both LM and CM was severely limited by available technology, space and power requirements.
* In 1997 Michael Foale was able to figure out the moves needed to stop Mir tumbling, after it had been hit by a Progress supply vessel, by using only his thumb against a viewport as a stellar sighting aid and his (well-trained) brain.
* Neil Armstrong had done something similar in Gemini 8 when the OAMS attitude control system malfunctioned and he manually stopped an uncommanded roll using the RCS.
Rather than picking a solution and attempting to shoehorn it into all the problems?
Yes, machine learning might be applicable for some tasks (although how do you train it for a moon mission?), but good old fashioned software written to do a specific job must surely be the correct choice for many, if not most things. Take the flight control software as an example : Surely this is just a variant on autopilot software.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
