Does space science or exploration require people? For any research project?

[joshualevy]

Can someone tell me what scientific research which tells us something about space (any location), is better done with people than with robotic spacecraft?

This is a serious question. I know that if we are to colonize space, we need to send people. And if we need to learn about people in space, we need to send people to space. But if the goal is to learn about space, or to explore space, what research project is better done with people? By "better" I mean a combination of cheaper, more accurate, and safer.

If someone says "analyze moon rocks in 100 different ways", well obviously, a lunar soil return mission is vastly cheaper, better, and safer (all three) then sending people.

I've heard the argument that people are needed to repair broken equipment (like Hubble), but I've also heard that sending 5 Hubbles into orbit would have been cheaper (and obviously better, and obviously safer) than sending one Hubble and having it serviced by people five different times (which is what we did).

Any counter examples, given current technology?

Joshua

 

[Catastrophe]

One point is that, when great distances are concerned, a decision cannot be relied on in time from distant humans. Hence humans must be on site.
I have in mind the 2 day gap in communicating with Voyager.
This may actually just be a subsidiary factor - coming back to the question asking what space tasks require distant human participation.

I suppose one example might be qualitative assessment, such as some research involving microscopy (bugs in space?)

Cat

 

[joshualevy]

One point is that, when great distances are concerned, a decision cannot be relied on in time from distant humans. Hence humans must be on site.

I understand the theory, but I've never seen an actual example. What specific decisions would a human need to make for which specific scientific research or exploration? Mars rovers have roamed all over Mars doing experiments. Obviously, the combination of robots on site and remote humans works very well for that. In the future that is going to be more true not less true.

I suppose one example might be qualitative assessment, such as some research involving microscopy (bugs in space?)

Couldn't the robot do the qualitative assessment? What sort of qualitative assessment are you referring to? The robot could send pictures or videos home, so the only issue would be if a quick decision were needed. But even then, I would expect the robot to do the assessment.

But in any case, I'm asking for a specific, real world example. Do you have one in mind?

Joshua

 

[Phillip Huggan]

Some health research. And anything with a WMD potential. If you can destroy one city no, but ten is maybe enough an attack. If you can blot out the sun with mylar, it can be burned through. But if you have lots of silver reflectors you might "succeed". If the goal is a colony, the research for it along the way benefits being manned.
I image hacking eventually becomes enough a risk near the end of the century it is nearly all manned. If you don't get lots of water to an asteroid outpost it may need to be unmanned is the converse. There are some things humans are smarter at. Whether to collect rock samples or frozen slush after reaching a destination and observing the geology, might use reasoning too hard for 2030 AI. If I'm preparing radar against Saturn space pirates, the AI won't be able to configure the best defenses not knowing the limits of imagination. For space diseases, I can't imagine ever using AI to even determine if we should J.Toews ourselves and distance from new Earthlings.
Addenum: if the goal is research done by a future Triton or brown dwarf ice moon colony, every colony along the way that isn't at an imminent risk of natural space death, will benefit the goal by trying manned research. The Armageddon actor who got space dementia would take over certain colonies, and not others even with lots of medical imaging. If you do your research manned along the way, you extinguish the risk. If you use algorithms instead, it takes longer to get to the goal of improving space society, as you have to learn to research as humans regardless if you do it along the way or have safe software do it. It would be like hacking your way into nuclear technology rather than learning it yourself: India didn't have medical gases for Covid.

 

[Catastrophe]

An example might be microscopic examination of previously unknown microorganisms. Could a robot decide whether it was already known, or be able to say that it was previously unknown?

Cat

 

[joshualevy]

An example might be microscopic examination of previously unknown microorganisms. Could a robot decide whether it was already known, or be able to say that it was previously unknown?

Yes, in three different ways:
1. By sending the images back to earth for comparison by humans or by automated image recognition.
2. By sending the whole sample back to earth for comparison.
3. By using image recognition to identify the organism on the probe.
The first two are already much cheaper, faster, and safer than sending a person. The third is a little more experimental, but I suspect it would turn out much cheaper, faster and safer as well.

It should be pretty obvious that for any question that can be answered with a sample, that sample return is cheaper, faster, and safer than a crewed mission. We already have sample return missions to more places than we have crewed missions.

Joshua

 

[Catastrophe]

What if an immediate answer is required? The human would have to be there.
Image recognition would not be sufficiently accurate? Shape won't work. No tiger stripes or leopard spots.
If the robot can't identify it, does that mean it is unknown. I am no biochemist (although I have a science degree) - perhaps you can suggest identifying characteristics?

So the answer is not pretty obvious if you are 10 light years away. (This is obviously a hypothetical question)

Cat

 

[joshualevy]

What if an immediate answer is required? The human would have to be there.
Image recognition would not be sufficiently accurate? Shape won't work. No tiger stripes or leopard spots.

My answer 3. By using image recognition to identify the organism on the probe. Covers that situation perfectly. We already have facial recognition which can identify one face out of millions. Obviously the same technology could identify one microorganism right now. But even if not, we could spend a billions dollars to develop that technology, and it would still be much cheaper, much faster, and much safer, then sending a person.

As a secondary question, my sort of real world research or exploration requires responses that fast? Right now, it is common to drive on Mars, do some research, decide where to go (back on earth), and then drive there. What sort of real world research would require that sort of decision making?

So the answer is not pretty obvious if you are 10 light years away. (This is obviously a hypothetical question)

We can't send people 10 light years away, so this is a meaningless comment. Remember my original question: what scientific research or exploration can be done better with people. We can not send people 10 light years away, so obviously people can not be better than robots for that work. (In fact, we can't send robots, either, so for that work, they are equal.)

Joshua

 

[Catastrophe]

Fine. You asked a question, I tried to think of a circumstance that might fit.
No harm done. EOS.

 

[FromBelow]

If someone says "analyze moon rocks in 100 different ways", well obviously, a lunar soil return mission is vastly cheaper, better, and safer (all three) then sending people.

Look at the troubles they had with the Preserverence rover early on just getting the sample into the tube.

With current technology a manned mission can cover more ground faster and acquire more interesting samples than any robotic mission could. You can do a lot of great science with robots, but if you want to do serious excavation you really need a person there running the machines.

 

[Catastrophe]

Does space science or exploration require people? For any research project?

Now this may be a little tricky, but in the final analysis:

As of now (state of robotic development), the answer must be yes . . . . . . because people will be required to analyse the data provided by these distant robots, or to (eventually?) program the robots to analyse.

Apologies for the sleight of hand (or thought).

Cat

 

[Ken Fabian]

Can someone tell me what scientific research which tells us something about space (any location), is better done with people than with robotic spacecraft?

This is a serious question. I know that if we are to colonize space, we need to send people. And if we need to learn about people in space, we need to send people to space. But if the goal is to learn about space, or to explore space, what research project is better done with people? By "better" I mean a combination of cheaper, more accurate, and safer.

If someone says "analyze moon rocks in 100 different ways", well obviously, a lunar soil return mission is vastly cheaper, better, and safer (all three) then sending people.

I've heard the argument that people are needed to repair broken equipment (like Hubble), but I've also heard that sending 5 Hubbles into orbit would have been cheaper (and obviously better, and obviously safer) than sending one Hubble and having it serviced by people five different times (which is what we did).

Any counter examples, given current technology?

Joshua

Crewed missions are almost entirely done for the sake of having crewed missions - with developing the capability to do so often a stated, specific objective of space agencies. I think this is about maintaining the public support required to get government support; whilst probes to the far reaches of the solar get popular interest the crewed missions get a lot more, especially when presented as about national pride and that flows through in funding. I don't think there is anything outside of self referential objectives like "demonstrate the feasibility of crewed missions" where having people present actually does anything better.

There is nothing that would require humans to be present or would be done better than using machines and in almost every case crews would be a liability compared to uncrewed; the priorities of any mission shift from any exploration or scientific objective to ensuring the safety and comfort of the astronauts, with very large payload requirements that could have been used for probes and instruments.

For the most part crewed missions could not even reach most of the places that uncrewed probes have already gone let alone do it better. Machines don't need any return capability and using them to destruction comes with no moral dilemmas. Where it is worthwhile having return capability is a lot simpler and less expensive without people. Good communications gives access to a wider arrange of experts than can ever be included as crew in any mission - and I think crewed missions will be more reliant on the availability of expertise back home to know how to respond to the unexpected than ever would be the case for uncrewed probes/landers/rovers.

 

[Catastrophe]

Does space science or exploration require people? For any research project?

Ken,

There is nothing that would require humans to be present or would be done better than using machines and in almost every case crews would be a liability compared to uncrewed

I agree with you, which is why suggested that there is still a need for people, if you include evaluation as part of the research project.

Cat

 

[Clovis]

Look at the troubles they had with the Preserverence rover early on just getting the sample into the tube.

With current technology a manned mission can cover more ground faster and acquire more interesting samples than any robotic mission could. You can do a lot of great science with robots, but if you want to do serious excavation you really need a person there running the machines.
[

 

[Clovis]

Not just yet but AI is working on it . Introducing the Androids.

 

[Ken Fabian]

Look at the troubles they had with the Preserverence rover early on just getting the sample into the tube.

With current technology a manned mission can cover more ground faster and acquire more interesting samples than any robotic mission could. You can do a lot of great science with robots, but if you want to do serious excavation you really need a person there running the machines.

No, a crewed mission could not cover more ground faster and no astronaut could've fixed the problems were they on Mars. If we want crewed missions by preference... we'd still be waiting. It is not just that machines do it better - only machines are capable of doing it at all.

Nothing is made easier by including people. Quite the opposite... enormously bigger missions are needed for a start and we are not yet capable of any crewed missions to Mars, let alone crewed ones that can bring along SUV's with long range capability and a well stocked and fitted repair shop. I imagine doing delicate dismantling and reassembly in a pressure suit with thick gloves in a dusty environment would be challenging too - could be better done with robots?.

I expect any crewed missions (if they happen) will still rely on rovers for the ground exploration - and accept the loss of rovers if they go wrong, because (going places people can't) they won't be able to get anywhere near them to them to fix them.

If it is important for a robotic rover to work reliably, how much more so when there there is a lot more equipment and there are lives depending on it? At that distance using bespoke equipment of high complexity reliability isn't optional, it is a fundamental requirement and any repairs, where those are possible at all, are likely to be done in consultation with experts back on Earth.

 

[Catastrophe]

Ken, I agree with this

Nothing is made easier by including people. Quite the opposite... enormously bigger missions are needed for a start

, which is why I pointed out that people were involved on Earth to evaluate the information collected by robots. I still think some decisions as to what to examine might not be within robot capability, but such might come in the future.

Cat

 

[billslugg]

In terms of scientific exploration, no sense in sending humans to far off places until robots have milked them for all the robots are worth. At the rate AI is advancing it may never come time to send humans for scientific work.

 

[Catastrophe]

billslugg, I certainly agree with this:

In terms of scientific exploration, no sense in sending humans to far off places until robots have milked them for all the robots are worth.

. . . . . . and I think it will be a very long (if ever) time before robots will decide "what to do next" when confronted with a possibility they have not been pre-programmed to respond to.

Cat

 

[Phillip Huggan]

Colbert had a skit with a tin foil robotic gas station in space. We can't do that one. My AI will be future hardware running LISP basically; just looking for an output (footsteps, phrases). Running a multi-product factory isn't there.
Around people, they will be Ph.D or somewhat higher in conversation, but they will still need to have a military escort to avoid someone throwing that info in their hardware with a USB-style device.
Setting up a warehouse of crates on a Mars moon, sure. But setting up an outpost deep in space may be complicated enough, 100 simple AI probes are needed for one of them to do it right. Interstellar had a centrifuge disintegrate. It would be hard to know which pieces to circular Travelling Salesman retrieve, but too good an AI may learn manufacturing.
For simple existing space plans, no prob. For things like navigating at a fraction of c, it will take time to safely have an AI capable of doing complex tasks, we might just wait for the database lookup type. I assume the movie with Russian and USA AI teeming up to North Korea humans, made real AI R+D slow. It was the reference last decade when two AIs started spitting out code at eachother fast, and the research was terminated.

 

[Catastrophe]

Would it not be easier just to have an orbiter with personnel? At least they would be 'there' for decision making rather than back home on Earth?

Cat

 

[FromBelow]

Nothing is made easier by including people. Quite the opposite...

I never implied that it was, nor was it the subject of the discussion.

The Apollo expeditions recovered far more (and more varied) material than all robotic sample return missions combined, and I don't think you could achieve the same results by simply throwing the same amount of money at robotics with the current state of technology.

 

[Ken Fabian]

The Apollo expeditions recovered far more (and more varied) material than all robotic sample return missions combined, and I don't think you could achieve the same results by simply throwing the same amount of money at robotics with the current state of technology.

Each one of the moon landings were much bigger scale missions than all the robotic missions you are comparing to put together and a lot more expensive. Had the robotic missions started with similar payload capacity (and funding) they could have returned a lot more sample material, from much more varied locations. Bigger, better rovers, more on-board capability to do high quality analysis on the spot as well as capability to collect sample material for return to Earth is all easier and cheaper than using astronauts. The Moon Buggy took astronauts a few kilometers from their landing site, they made a few stops and picked up some samples but robotic rovers just keep on going and going.

In any case I think it may be more important to do thorough surveys of everything, everyplace rather than pick and choose what appears interesting; that picking and choosing what appears interesting to the human eye is a consequence of lacking the resources to be thorough. Instrumentation can deliver very high quality images including down to microscopic and make other observations beyond the range of human senses - and have them assessed by the world's leading experts rather than a generalist astronaut.

Would it not be easier just to have an orbiter with personnel? At least they would be 'there' for decision making rather than back home on Earth?

I don't see how that really adds anything; the expertise you can include will always be limited. Back home you can access every top expert. The communications time is probably not going to add much to the time to make a decision. I am struggling to think of any emergency situation involving a robotic rover or other remote operated equipment where 20 minutes would be not only crucial but people in orbit would be better able to respond.

The payload for putting those people in orbit, supporting them and returning them is payload that could be used for multiple rovers, better instruments, more capabilities.

 

[Catastrophe]

Ken, my point was that, whether people are in orbit or staying at home, they are necessary for evaluation and decisions. In orbit is just quicker, but, I agree, much more involved. No matter. Humans are still necessary.

OK, one can envisage all sorts of SF scenarios where robots understand and plan better than humans, but we are not there yet.

Cat

 

[Phillip Huggan]

There is still a blueprint risk even if the AI is made safe.
I figure a real-time response to a novel space risk via machine learning takes 10 minutes if I'm 150 and keep reading. That is me programming it, not itself as an agent. It took me 1.5 hours to figure out a paragraph on how to get water to a cooling pond. Ignoring the Terminator risk and the hacked blueprints risk, how do we have an AI make a good value judgment? We have superior wetware to world model. The AI doesn't have qualia, so it would take Eras to know whether to save the doctors or the engineers in the centrifuge without some updating procedure. It means the AI has to be continually made or it is already obsolete once real time events happen.
The AI starts off writing UK papers. Then German ones in two years. Then Chinese in five years. Then Bangladesh. Then screwball journals with weak physics. Why would we ever use it assuming an active dynamic space economy?