Foresight Nanotech Institute Logo
Image of nano

Robots

There was some objection to my post Is Robo Habilis a gateway to Intelligence? to the effect that it might take a lot of extra time to build the robots, and that would lengthen the time necessary to develop AI. That might certainly be true of the garage experimenter, but in the world at large, the robots are already here. The kind of robots I’m thinking of are bolted-to-the-table torsobots in the tradition of Cog:

Cog picture

The reason is that as of even date, you just can’t put enough processing in a mobile robot to be doing the kind of processing a human is doing with its sensory and motor streams.

There’s also of course Kismet:

kismet

and various other research robots ranging from the simple –

Babybot

babybot

Ludwig

Ludwig

They get more complex, e.g. Berti:

Berti,Berti

to the extremely anthrobiomimetic CRONOS2:

Cronos2

to some serious engineering:

Domo:

Domo

iCub:

iCub

Saika:

Saika

Nasa’s Robonaut:

Robonaut

These use commercial arms and manipulators:

The UMASS Torso robot:

UMASS torso

The Iowa torsobot:

Iowa bot

And finally, these appear to be commercially available:

Hiro:

Hiro

Motoman SDA10:

Motoman

and the Meka (a descendent of Domo in commercial form):

headarm

so … the bodies are there.

5 Responses to “Robots”

  1. Tim Tyler Says:

    To backtrack – there were two points under discussion – one was whether testing for intelligence using robot controllers made sense – the existence of robots seems irrelevant to that.

    The other point was whether we would see machine intelligence without it going through a series of mechanical robot phases. I think the most important bodies for intelligent machines are probably largely virtual ones – and humans. Real physical robot bodies have relatively little significance today – due to their poor capabilities, and poor penetration into society. Yes, there are some in factories and warehouses – but those are mostly behind firewalls and are practically inaccessible from the wider internet. Assuming that artificial intelligence will spark on a big server somewhere, robots are no good to it unless they are connected up.

    The *real* robots – from the point of view of intelligent machine development – are the billions of mobile phones and the keyboard-mouse-screen combos that populate offices throughout the land. Those robots make extensive (albeit second-hand) use of molecular nanotechnology in the form of human beings – greatly extending their power and reach. Their numbers, biomass, etc dwarf that of pure-mechanical robots.

    Maybe in the future we will see some more mechanical robots – and maybe then they will have some relevance to intelligent machines. However, it looks a lot as though we will get machine intelligence before we get much nanotechnology or very competitive robotics. As far as the path to machine intelligence goes, mechanical robots look like an irrelevance.

    Of course there’s a tradition that thinks that embodiment is important. It is true that you must have sensors and actuators – but they can open directly out onto the internet – which is quite a rich and stimulating environment, thank you.

    Not building mechanically-actuated bodies is cheaper, faster, and allows machine minds to utilise the existing internet networks and human infrastructure – which represents a huge leg-up for an intelligent machine.

    Loud speaker and LCD-monitor actuators work quite well enough. There will be little or no need for servos and motors in the development of intelligent machines – which is a good job – since those actuators and their controllers are at a primitive stage of development.

  2. Tim Tyler Says:

    I see you presented on a related topic recently:

    The Robotics Path to AGI Using Servo Stacks – http://www.vimeo.com/7260012

    Machine development has taken place in an environment where many of the most easily accessible niches are those where the machines compliment human intelligence, rather than competing with it.

    As a result, machines are excellent at mental arithmetic and remembering things – areas where we are weak. However, there is *relatively* little incentive for them to get good at things where brains are already good – since humans can already do those tasks.

    That situation seems likely to persist for a while. As a result, machine intelligence will evolve in an ecosystem where complimenting human intelligence is what is important. So – they probably won’t start off with an intelligence that is much like ours.

    Consider what happens if Google gets smart, for example. Will it be human-like? No human can remember the whole net. No human can answer millions of questions simultaneously. The first intelligent machines are likely to be nothing like humans.

  3. Tim Tyler Says:

    Oops – I meant to say “complEmenting human intelligence” ;-)

  4. Glyn Moody (glynmoody) 's status on Sunday, 08-Nov-09 14:05:51 UTC - Identi.ca Says:

    [...] http://www.foresight.org/nanodot/?p=3477 a few seconds ago from Gwibber [...]

  5. Dave Wyland Says:

    Re: Tim Tyler
    This says that the key to robotics is “intelligence” (or its equivalent) and that it can be developed independent of embodiment. A reasonable hypothesis to test. We have been testing it for 50+ years with little to show for it in any practical sense. Factory robots do not require it, and (almost) all field robots (iRobot in Iraq, UAVs, Mars Rovers, etc.) are tele-operated.

    The mechanics of robots has proceeded at a good pace. Robots in factories do their jobs quite well. We have running and grasping robots, ect., but they only work well when a human is supporting them with some level of tele-operation. This supports the point that the mechanics are good (or at least good enough), but their control is missing – and still has to be supplied by a human for practical applications.

    I suggest another direction. Factory robots we have. Field robots we need. Harvest Automation (http://www.harvestautomation.com/) has made a start with field robots that autonomously move plants around in a nursery. A simple – but not necessarily easy – task. There is a huge range of applications for robots that can move things around without knowing before hand exactly where they are, as is required in factory robots. This is tough engineering, as opposed to R&D. But it is a path that can be developed to yield increasingly useful stuff.

Leave a Reply