Master of AI

Grypd · 2004-08-23 08:46:31

Think of this as the Robots X prize.

http://www.darpa.mil/grandchallenge/]Darpa Grand Challenge

The grand challenge is now in its second year. The first year no organisation was able to meet many of the goals that had been put down. Actually all the robots either crashed or where deactivated for safety reasons. Interestingly as we are at the beginning of the AI science it means one of the most astounding organisations at the first grand challenge was a couple of friends from the midwest, who due to the simplicity they had designed into there entry it resulted in one of the most effective and at a shoestring budget compared to the big teams. We even had a team who had made a motorbike entry but forgot to plug something in so it went in circles before being "fenced". So Darpa's real interest is to push robot science to the point we have a robot which can reach a point destroy it and return home without further human control. But it will aid all the rest of the science when we have totally automated vehicles

SpaceNut · 2004-08-24 10:37:56

Follow up story on Robonaut.
NASA’s Robonaut finally gets its ‘legs’
Second-generation android uses wheels or adapter
http://www.msnbc.msn.com/id/5798085/

040823_robonaut-segway.vlarge.jpg

John Creighton · 2004-08-24 10:52:01

I like BeamBots they have been used by Robot enthusiasts and industry for decades. And they get better each year. But the one thing they are not is intelligent, they may have terrain object avoidance but no real complicated logic problems if they get stuck they just shout for help.
Im currently working on one to mow my lawn it will follow a magnetic wire around my garden and do my most hated garden chore. I could program the dimensions of the garden in and let it go but I believe the way im doing will give the straight lines I want.

How would it cut the grass? Does it bit the grass? How many motors is your BEAMbot going to have for walking? How many legs? In http://www.newmars.com/cgi-bin/ikonboar … SF;f=16]My Mars Rovers : http://www.newmars.com/forums/viewtopic … Simulation of Intelligent Robotic Colony I am discussing a two motor walker that I believe could be modified so it can walk in a circle.

Grypd · 2004-08-24 11:17:37

My mower was/is using a rotating blade and for power it is connected to the grid by cable. The legs where a 3 blade rotating method. But it looks more and more like im going back to wheels. Easier to operate. Oh and I tried to use Nitrinol wires so as to make Elecric muscles but again not really too strong on the budget I had.

John Creighton · 2004-08-28 15:06:38

Here is an interesting distributined computing artificial intellegence project.

http://www.aiathome.com/]http://www.aiathome.com/

Grypd · 2004-08-28 15:58:59

My hope in using a lawn mower beam bot was to see if

a) it could be done but in this I had real trouble

B) One of the most effective mining technigues in space is to use a rotating brush to collect surface material up from the ground. Standard digging technigues used on Earth are a lot less effective on the Moon and Mars and objects that have a reduced gravity. Also the surface of these objects are covered in slightly compressed rubble or regolith caused by asteroid impact and we must dig through this to find if we want to non fractured strata. Using a beambot meant I could use legs as stabilisation.

It was my intention to use them like an ant colony would but I had too much trouble with weight issues/movement.

Unfortunatly I have now come to the Idea that until a form of stonger way to make muscles for robot legs and arms appears with a decent lightweight power source . Robots using wheels will have the advantage when being used. (it is what I now plan to turn my lawn mower robot into but with the advantage that we can use a petrol engine to provide power source). Also the reduced complexity of a Wheel based robot reduces maintenance problems.

John Creighton · 2004-08-28 16:40:48

Unfortunatly I have now come to the Idea that until a form of stonger way to make muscles for robot legs and arms appears with a decent lightweight power source .

Have you considered gears?

Grypd · 2004-08-28 17:06:41

I Did use gears to gain an increase in power to the legs, But they still struggled with any power plant I could put in and I really did not want to have it attached to the mains but in the end I did. And it still struggled to move in fact it barely did. I even used/attempted to a rotating leg system so the legs would flip if they got stuck(a commen problem). In the end I gave up as it had turned out so bad. You see to keep the Bot stable it had to have legs which where too heavy, the flipping legs just stopped the machine dead. sorry if im not explaining myself too well.

John Creighton · 2004-08-28 17:35:55

As for power try a fork lift battery or maybe somekind of genrator could be hooked up to a lawn mower engine. I don't know if the limit is in your power source or the ellectric motors you have. As far as stability goes move one leg at a time as was done in:

http://faq.solarbotics.net/patent/patent.html]Mark Tildens Nervous Net Patent

comstar03 · 2004-08-28 22:44:55

John, go to Honda robotics in japan they have all the information on their robots online including walking and motor movements.

SpaceNut · 2004-09-10 14:07:58

Quite the feat for these Engineer to Build a Robot That Walks on Water, The prototype shows how important lighter and stronger materials have been to robotics.
http://story.news.yahoo.com/news?tm....g_robot

SpaceNut · 2004-10-07 20:23:03

This is a first step done the road to AI.

NASA software finds satellite problems
NASA scientists recently corrupted a spacecraft's system and caught the glitches with artificial intelligence (AI) software in a test.

http://www.fcw.com/fcw/articles/2004/10 … -07-04.asp

John Creighton · 2005-01-12 17:35:34

Here is an interesting article on AI:

http://www.cs.washington.edu/homes/lazo … ra/ai.html

Autonomous vehicles: A DARPA-funded onboard computer system from Carnegie Mellon University drove a van all but 52 of the 2849 miles from Washington, DC to San Diego, averaging 63 miles per hour day and night, rain or shine;
Computer chess: Deep Blue, a chess computer built by IBM researchers, defeated world champion Gary Kasparov in a landmark performance;
Mathematical theorem proving: A computer system at Argonne National Laboratories proved a long-standing mathematical conjecture about algebra using a method that would be considered creative if done by humans;
Scientific classification: A NASA system learned to classify very faint signals as either stars or galaxies with superhuman accuracy, by studying examples classified by experts;
Advanced user interfaces: PEGASUS is a spoken language interface connected to the American Airlines EAASY SABRE reservation system, which allows subscribers to obtain flight information and make flight reservations via a large, on-line, dynamic database, accessed through their personal computer over the telephone.

SpaceNut · 2005-02-15 11:03:01

NASA Robotic & Human Exploration of Mars Roadmap Committee Meeting 4-5 Jan 2005 Presentation information

Lots of info:

Alexander Sheppard · 2005-02-16 22:17:37

It occurs to me, and this seems to be a rather obvious point, that you can't seriously talk about creating industry on Mars with beforehand artificial intelligence until you can build equivalent systems on Earth. I am curious what kinds of systems you are talking about building there using this technology.

What exactly do you mean by moving humans into the place "in relative luxury"? It seems to me that people on Mars will, physically, enjoy a great deal of luxury and probably lead very fulfilling lives compared to most of their cousins on Earth. Might you suggest terraforming the planet before humans put down a presence?

Personally I tend to think if you have the technology to create a full scale industrial apparatus on Mars without any human involvement, then it will only be a short time before technological transcension and then human colonization becomes less important compared with colonization by more developed forms of intelligence.

If you mean teleoperation, my reaction is that this would have more promise, but I think you might as well ship some humans over fairly soon anyway.

Martin_Tristar · 2005-02-17 01:10:29

You don't need AI to develop robotic systems for outer space, a rule based software environment can be used to build up a complex software environment for the robotic systems to function.

The current gaming, graphics and other media technologies can be used to create an effective software platform for robots to function in space and where the rules haven't got answer the robotic unit will request an update from command and a new rule will be supplied applying to those conditions.

We require a storage medium within the robotic units to process and apply the rules while construction or other related tasks. we could also develop a monitor environment in VR world and can override any robotic unit under supervision from within the VR environment.

While the robotic systems may seem to function independently they don't but act under supervision and lead by a human controller as team leader or supervisor.

SpaceNut · 2005-05-23 12:12:10

Well here is another twist on the uses for AI. Say that you want to live on forever but you can not have a human body, Death Could Be Averted By 'Downloaded Brains', British Futurologist Says

While the predictions might sound outlandish, they were merely the product of extrapolations drawn from the current rate at which computers are evolving,
If you draw the timelines, realistically by 2050 we would expect to be able to download your mind into a machine, so when you die it's not a major career problem,

Do you still pay taxes? ???

Rxke · 2005-05-23 15:27:13

Yeah, and it's all over Slashdot too.

Noone ever read Ray Kurzweil? Sigh. He predicted this years ago, using the same extrapolations, though his estimates are a bit more moderate.

The age of spiritual machines, 1999. Required reading. It'll change your view on the world forever.

Grypd · 2005-05-23 17:17:30

Interesting point to this if we are "downloaded" then what rights do we have, are we human and if we come up with a real AI what rights does this have compared to a "downloaded" human copy. And do you have to be dieing to have the "download" happen. Could there be more than one of you in a computer system.

Still a human brain controlling machinery is a very useful technological trick as this would allow machinery to work in places that are dangerous for a purely human brain or to allow a human mind to control machines capable of doing much more than pure flesh and blood can.

Still I think it will answer one thing, are what we do the result of our hormones like love, anger, hate or do we find that emotions will fade if we are no longer flesh and blood.

PurduesUSAFguy · 2005-05-23 17:22:33

Well I think you could put automate alot of simple resource utilization tasks up to a point.

For example, I think it would be very possible to land a completely automated construction materials factory. Engineer a couple of loader dump truck rovers and a central receiving center to crack it and produce cement and other by products and then feed them into a smelter to produce magnesium alloy and then cast it into rebar.

The crew can then land and bring along compressors and air bladders and then will have all the materials to produce a series of monolithic domes. I forgot what the construction technique was called but it involeves laying a foundation, and then setting an air bladder on top of it. then a web of rebar and concrete (and I would guess in the case of mars heating elements) is laid up on top of the bladder, which is then inflated producing a dome. You can then reuse the bladder and build another dome, and another and so on...

At least that's how I can envision producing habititats on Mars using completely endigenous(sp?) materials.

Grypd · 2005-05-23 17:52:30

Well I think you could put automate alot of simple resource utilization tasks up to a point.
For example, I think it would be very possible to land a completely automated construction materials factory. Engineer a couple of loader dump truck rovers and a central receiving center to crack it and produce cement and other by products and then feed them into a smelter to produce magnesium alloy and then cast it into rebar.
The crew can then land and bring along compressors and air bladders and then will have all the materials to produce a series of monolithic domes. I forgot what the construction technique was called but it involeves laying a foundation, and then setting an air bladder on top of it. then a web of rebar and concrete (and I would guess in the case of mars heating elements) is laid up on top of the bladder, which is then inflated producing a dome. You can then reuse the bladder and build another dome, and another and so on...
At least that's how I can envision producing habititats on Mars using completely endigenous(sp?) materials.

We can automate a lot of processes but in the case of robots we have found that the capability of a robot is let down by its thought processes. Robots have more or less cracked the ability to have motion, power, ability to utilise items and vision. The stumbling block is there brain we have to program them and it makes them want to follow step one then step two processes and this limits there effectiveness. Of course Telerobotics where simple avoidance control mated with a master command by a human operator can makes for a much more effective system for use in construction and in resource utilisation.

The problem with this becomes communication and control lag. In the case for the Moon this is not too much of a problem but with mars distance it means that effectiveness will be much more a problem. Still NASA has been experimenting with these systems as a robot does not need 3 hours of pre breathing to be ready for a space walk and in an emergency a "Robonaut" will be able to get to the problem site a lot faster. And the Robonaut is a telerobotic device which acts and uses tools like a human can just better (in space anyway)

SpaceNut · 2005-06-09 11:32:13

Then again an AI bot would require the sense of touch. Not straight from the star trek movie were the borg graft skin onto Data's exo skeleton but close.

Goddard Technologist Proposes Sensitive High-tech Robot Skin

High-Tech Robot Skin Goddard Technologist Proposes Sensitive Skin Covering for Robots

SpaceNut · 2005-06-24 08:52:44

Japanese robots to guard shops and offices Will alert human guards of intruders, fires, water leaks

050623_japrobot_hmed_7a.hmedium.jpg

A few more steps along the way towards a robo cop.

Grypd · 2005-06-24 09:34:07

you beat me to it,

Rxke · 2005-06-24 14:02:30

Sigh. And me tinkering here on my first semi-robust robot...
Made out of old CD/HD parts, diskdrives, pieces of old scanners, and assorted olden-days hi-tech debris

But mine will be able to walk around in the grass, through the dirt, hah!
That one looks like it will stumble over a loose brick in a flat pavement

New Mars Forums

Announcement

#26 2004-08-23 08:46:31

Re: Master of AI

#27 2004-08-24 10:37:56

Re: Master of AI

#28 2004-08-24 10:52:01

Re: Master of AI

#29 2004-08-24 11:17:37

Re: Master of AI

#30 2004-08-28 15:06:38

Re: Master of AI

#31 2004-08-28 15:58:59

Re: Master of AI

#32 2004-08-28 16:40:48

Re: Master of AI

#33 2004-08-28 17:06:41

Re: Master of AI

#34 2004-08-28 17:35:55

Re: Master of AI

#35 2004-08-28 22:44:55

Re: Master of AI

#36 2004-09-10 14:07:58

Re: Master of AI

#37 2004-10-07 20:23:03

Re: Master of AI

#38 2005-01-12 17:35:34

Re: Master of AI

#39 2005-02-15 11:03:01

Re: Master of AI

#40 2005-02-16 22:17:37

Re: Master of AI

#41 2005-02-17 01:10:29

Re: Master of AI

#42 2005-05-23 12:12:10

Re: Master of AI

#43 2005-05-23 15:27:13

Re: Master of AI

#44 2005-05-23 17:17:30

Re: Master of AI

#45 2005-05-23 17:22:33

Re: Master of AI

#46 2005-05-23 17:52:30

Re: Master of AI

#47 2005-06-09 11:32:13

Re: Master of AI

#48 2005-06-24 08:52:44

Re: Master of AI

#49 2005-06-24 09:34:07

Re: Master of AI

#50 2005-06-24 14:02:30

Re: Master of AI

Board footer