Re: Humanoid Robots on the Mass Market

From: Eugene Leitl (eugene.leitl@lrz.uni-muenchen.de)
Date: Mon May 22 2000 - 21:21:27 MDT


Technotranscendence writes:

> I kind of agree, though I think such floor cleaner robots could be built
> cheaply now. If you just need a vacuum cleaner that can guide itself, one

Not household-usable ones. The amount of crunch and the code for
building maps/navigating is very nontrivial. I trust Moravec's
expertise on this. However, I'm placing such devices 2010, while he
does it iirc 2005.

> could borrow Rodney Brooks model of robotics. That is build simple,
> basically "stupid" robots that can only do not very complex things like not
> bump into a wall.

There is too much that can go wrong with a robot in a house. Not
falling down the stair, and not messing up stuff on the floor is very
nontrivial.
 
> Since this seems so easy (from my computer desk chair:), I wonder why
> someone hasn't done it. After all, there are a lot of big lobbies -- Grand

Because it stops looking so easy if you go out, and try to do it.

> Central Station, NYC, for example -- and such that could cleaned by such
> robots. (I think the average house or apartment would be a bit too much
> here, but I could be wrong. Not my area, you know?)
 
Anything which is not cluttered, and changes only very slowly, and can
pay a lot for a single installation, is relatively easy.

> Also, I believe making robots look humanoid would mostly be a waste.

Of course.

> Designs should be based on cost (including design, manufacturing, and
> maintenance) versus functional efficiency. Even an all around robot
> probably would not be humanoid. The only reasons I can see for making some

A box with an arm or two would do just fine.

> consumer robots humanoid is for esthetics and for marketing. Like tail fins
> on cars.

Sony Aibo.



This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:11:32 MDT