Tuesday, June 01, 2010

BP Oil spill and Space Robots

If you have been a long time reader of this blog you probably have a geeky desire to see all the technical details. I've been watching the BP spill response with morbid fascination.
BP looks like it is being very open with the technical aspects of its response.
just take a look at all the videos on this page, the scale and scope of the operation are mind numbing. If that link does not work as a permalink, just look for the June 1 Videos on the LMRP, or go back and watch all of Kent Wells presentations.

I've been watching the the technical videos, diagrams and briefings for at least the last two weeks. I think this whole thing could be used as a pretty detailed response to those space scientists that say don't send humans, send Robots. With the BP spill we have lots of very sophisticated ROV's ,they have reasonable access to the surface for repair, adjustment and tool change out, a round trip operating delay of ~10 micro seconds and yet the whole process looks painfully hard. Trying to do any serious resource extraction or heavy construction remotely without direct onsite human intervention is currently significantly beyond state of the art. We need Humans on site.

We need humans in space. Go F9-Dragon!


Jon Goff said...

Well put, Paul. Robots are cool, but even adding a few people into the mix makes them a *lot* more productive.


Lee Valentine said...

The robots uber alles crowd have no experience whatsoever with robots. They are scientists who just want to measure something or politicians who want to use human space flight money to fund their own pet wasteful schemes.

Sebastian said...

Exactly, I recently had an discussion just about that. I asked simply how would robots just change wheel bearing in a rover (or do any kind of similar very uncomplicated repair requiring simple disassembly and reassembly, using crude tools). The answer was to make the bearing unbreakable (ROTFL!) by for example making it contact free magnetic bearing (ROTFLMAO!)

Laurent said...

Paul, I've been following your blog for a few years now.

I must admit I'm suprised by your comment.

Yes having robots do "stuff" in a very remote location is hard, but is the fact it is "hard" enough to deter making attempts...

You do not seem to run away from "hard" problems don't you? :-)

Paul Breed said...

I'm not afraid of hard, I also believe that you need a realistic view of your technology readiness.

Remote operation with zero onsite human intervention is hard, not impossible, just harder than most people think. They have Hollywood visions of robotics, and the BP spill offers a realistic industrial view of robots.

Laurent Gauthier said...

Well I fully agree with you when you say "Remote operation with zero onsite human intervention is hard, not impossible, just harder than most people think.".

But as the Darpa Grand Challenge and X-Prize LLC experiments have shown a lot of the basic technology needed for these "hard" problems is well understood and a lot of the problem lies in the integration of systems.

The DARPA Urban Challenge has been an amazing demonstration of that, with robotic drivers navigating through human traffic.

This problem was thought to be very hard to crack, yet they made it...

Laurent Gauthier said...
This comment has been removed by the author.
Laurent Gauthier said...

And of course to demonstrate your point here's what happens now: The saw is stuck... :-)

Itokawa said...

I disagree with the author's conclusions entirely. All that these videos show is that it takes time and effort to do things with ROVs. So? We could land ROVs on the Moon and take our time to develop an industrial infrastructure. Despite the fact that it might take 3x as long to complete an operation with an ROV as with a human, sending a human there might cost 100x as much.

Coastal Ron said...

I agree with Itokawa. Using the BP work underwater in extreme environments, and extrapolating that experience to off-world exploration is absurd.

If you want to look at the future of robotic lunar exploration, you should start with the robotic Mars missions, which we continue to expand. Those are hampered by communications distance, and the relatively small payload size we're able to land currently.

For the Moon, we are much closer for communication, and we can land significantly more payload per launch. Instead of one rover exploring, I would imagine two or more working in teams would expand our abilities even more, and provide opportunities for one to help the others when you run into problems.

On Mars, the lack of a simple arm with a brush limits the effectiveness of the solar panels. With a larger & more capable rover, you can add more manipulators, and more sophisticated ones. The rovers could clean each other off as needed.

In regards to changing a wheel bearing, that is a great question, and for now the only answer is to build the rovers with replaceable sub-assemblies that are plug & play. These are the kinds of problems that both robotic and human explorers will experience, and if you can solve them for the robots, then the humans will have a much easier time.

Paul Breed said...

One of the Mars Rovers is stuck in a sand dune. The other rover has traveled ~20Km or 0.0092km per day.

The human driven rover on Apollo 17 drove 39.5km on a 3 day mission.
(Only 4.5 hrs of actual driving) Even using the 3 day measure it was 1400 times more effective using technology 30 years older.

I'm not arguing that robot missions aren't amazing, I'm arguing that manned missions would be so much more capable.

Anonymous said...

Mmm, well, I don't know. Given a user interface technology sufficiently well integrated with the human nervous system I think the distinction may be moot. The technology may not even require implants, it may just require good enough feedback so that the nervous system gets the right cues. That said, a lag time of more than a few microseconds is known not to work. Haptic systems over the Internet (where lag times run upwards of 80 ms for local or regional RTTs) typically don't work very well.

heroineworshipper said...

A fully funded program to return humans to the moon in 10 years would have been better off using humans & a lot more inspiring, but now we're looking at 50-100 years before anything replaces Soyuz. In that timescale you're better off with robots. Humans in space are going the way of fighter pilots in 100 years. The human species is not going to spread among the stars. In the next 1000 years human bodies will be replaced by machines implanted with our thoughts & the machines will leave Earth.

Coastal Ron said...

Paul Breed said:

"One of the Mars Rovers is stuck in a sand dune. The other rover has traveled ~20Km or 0.0092km per day.

The human driven rover on Apollo 17 drove 39.5km on a 3 day mission."

Spirit and Opportunity have been working on Mars for over six years, while only guaranteed for 90 days. Constellation would have a hard time keeping astronauts on the Moon for 30 days. The Mars low mileage/day includes many planned stops for tests and observations, as well as powering down for the winter. Mileage doesn't count - only results.

No doubt humans are more capable, but my point is that with the Moon closer, and with the latest robotic technology that is available today (not tomorrow), robots can leave for the Moon on Atlas or Delta within 5 years, and be prospecting full-time while we work out whatever HSF plans get approved.

I'm not anti-human Moon exploitation, I'm anti-waiting for the perfect plan. Robots are cheaper to build and improve - you can build, get feedback and launch an upgraded replacement in far less time than you can plan one human mission, and the original robot could still be performing valuable work.

The Japanese are planning a robotic Moon base, and I hope they beat us to it, because that will cause a new rush to the Moon. That's when NewSpace companies will be truly successful.

Ed said...

The whole manned-versus-unmanned debate misses the whole point. Sending robots anywhere only makes sense if we also plan on later sending human beings - we could send a thousand robots to Mars but if we don't send people then those robots are mostly wasted. And, if we are sending human beings anywhere then they will almost certainly be taking robots along with them.

Hobart said...

You can put people a mile below the surface of the ocean. I wonder why they don't? I mean, it's a business proposition. Either those people offer value at that site, or they don't. I guess they don't.

The same may be said for off-world humans. They may offer value, and they may not. Whether they do depends a lot on how much they cost. They cost a LOT. You can talk about how much faster humans can drive around, but for the price of putting a human behind a wheel on Mars we could put robots all over the planet, covering vastly more territory than humans ever could.

Putting humans "into the mix" offers value? You bet. In what way is virtual presence not putting humans "into the mix"?

Finally, be aware that Moore's Law doesn't work for humans. Robots (both teleoperated and autonomous) are getting rapidly more capable. (Rapidly on a timescale of our human space flight progress.) If you measure future capabilities of robots by past performance, you're making a big mistake.

Paul Breed said...

They don't put humans into the mix 1Mile under water, because they can't.
There is no benefit having a human on the other side of a tiny porthole if he can't touch anything. The deepest dive ever done was about 1000 ft, the deepest a 1 ATM hard suit can go is 2K ft. These ROV's are at 5K ft.

Anonymous said...

Look at BP's stock price, which is down 40%. If you knew when the well was going to be fixed, you could make a killing. In fact if you delayed the fix, you could make even more money.
Maybe it's not as hard as it seems.

Anonymous said...

@Coastal Ron ... Steven Squyres, head of robotic research on Mars, stated that what the robots had done on Mars in six years could have been done by humans in a week and half.

Coastal Ron said...

LupusSolus said...

"@Coastal Ron ... Steven Squyres, head of robotic research on Mars, stated that what the robots had done on Mars in six years could have been done by humans in a week and half."

Yes, but we have robots on Mars, and no humans. We're getting lots of research done because it's cheaper to send robots than humans.

As I've stated before, I'm not anti-human exploration, I'm anti-waiting around for humans to get to the Moon sometime next decade.

We can have multiple large rover/robots on the Moon in 5 years, and follow up with more as quickly as we can build & improve them. Why not use them to do the initial grunt work, and tell us the things that will make human exploration a lot easier?

We have the technology, so there's no reason we can't get started on it now. And as I understand it, NASA does have robotic missions planned for the Moon in their proposed budget.

Hobart said...

"They don't put humans into the mix 1Mile under water, because they can't."

Huh? Alvin goes to 4400 meters with people in it. Oh, you mean humans in "suits"? How does not wearing a suit keep you out of the mix? What kind of "mix" are we talkin' about here?

But really, saying that humans are necessary because robots are challenged at the spill site is daft. I was in Target the other day, and a checker there screwed up the bill. Ah! That checker needs to be replaced with a robot. Obviously a human wasn't up to the task.

You can make arguments for humans, but pointing at the troubles at Deepwater aren't those arguments.

reader said...

Ok, ive been in the robotics, past tense. I know the challenges. I also know about a couple different schools of thoughts about how do to robotics.
Lets just say that im thoroughly unimpressed by the stuff that they use at the BP site, as compared to the real state of the art, available elsewhere in the world.

Take a trip to one of the asian robotics trade shows.

Burnsie said...

Mining is becoming fully automated because it is too hard to keep killing people. In this age of TV and instant news, mining companies are moving to tele centres and fully automated mines.

Even a telecentre on the moon is a lot cheaper than miners in lunars mines.