Newsgroups: comp.robotics
Path: brunix!sgiblab!swrinde!ihnp4.ucsd.edu!library.ucla.edu!csulb.edu!csus.edu!netcom.com!nagle
From: nagle@netcom.com (John Nagle)
Subject: Re: PROGRAMMABLE MATERIALS / Shape Changing Robotics Technology
Message-ID: <nagleCpIHrz.I0H@netcom.com>
Organization: NETCOM On-line Communication Services (408 241-9760 guest)
References: <768426734snz@stellar.demon.co.uk>
Date: Mon, 9 May 1994 01:35:59 GMT
Lines: 77

Joe@stellar.demon.co.uk (Joseph Michael) writes:
>                     INTERNATIONAL PARTNERS REQUIRED
>_______________________________________________________________________________
>	A recent patent application has been filed for a SHAPE CHANGING ROBOT.
>If anyone is iterested in commercial partnership in developing these robots
>please email me - Joe@stellar.demon.co.uk.

      Anybody ever heard of this guy?  This sounds like somebody who just
read Michael Crighton's "Demon Seed".

>	The product is called PROGRAMMABLE MATERIALS. Under computer control,
>you instruct programmable materials to for example turn into a walking machine.
>Alternatively, you can turn into a wall or a flat surface etc.
>It can usefully deform around objects e.g. you can take several tons of
>programmable materials up the stairs and through a narrow door entrance
>into a failed nuclear power station to prop up the ceilings, erect lead
>walls, pump out dust and smoke, install lighting and cameras etc.

      There are a few things already working along these lines, but the
existing technology isn't very good.  Mark Yin at Stanford just built
a cute little machine made up of a number of 2"x2" cubes, each of which
has two actuators that can deform it.  By assembling enough of these cubes,
you can make some interesting gadgets.  Each has a CPU, and they form a
network.  

      Another approach is materials that deform when electric or magnetic
fields are applied.  A number of such systems exist, magnetic-particle
clutches being the most useful.  There are the shape-memory alloys, which
are slow, magnostrictive materials, which have a small dimensional change,
and a few other schemes.  I recall reading about some approaches involving
organic fluids (starch?) but don't recall hearing about that in recent
years.  

      This sort of thing is a standard claim for nanotechnology, but the
state of the art there is nowhere near doing this.

      But with the exception of magnetic-particle clutches, none of this
stuff is useful enough to show up in "Machine Design" magazine, which is
the trade journal for actuators and such.

      If this stuff is for real, make and market a simple actuator based
on it, and see if it goes anywhere.

>	In military applications ...
>	In big civil engineering projects, ...
>	Programmable materials are highly fault tolerant robots. ...

       A bit ambitious for phase I.

>	If you going to post a proposal, please remember that it will take
>about a year to put the hardware and software together. A more realistic time
>scale is three years for commercial products because the operating system
>I am proposing require complete rewrites from normal commercial products.

       Nah.  Concentrate on the actuator for now.
        
>The electronics would be cubersome using normal CPU technology and for that
>reason, I am also proposing a new RISC CPU and ASIC glue logic to get the
>job done. The initial costs of bringing the technology to market is around
>$3 million per year for three years. That timescale can be brought down to
>1.5 years with investment running at $5 million per year. The costs are
>recovered quickly through access to publicity, demonstations, seminars
>and sales of prototype kit. Huge orders from the Nuclear Industry are
>expected because there are no competing systems, and the patent is fresh.

       I wouldn't expect "huge orders from the nuclear industry".  There
are in fact a number of companies making robots for the nuclear
industry.  Remotec, in Oak Ridge, TN, makes a rather nice machine.
Some others come from the Field Robotics Lab at Carnegie-Mellon.  But
the market is tiny.  

       If you're serious about this, make an appointment with Arthur Rock,
at 3000 Sand Hill Road, Menlo Park, CA, who is generally considered to
be one of the smarter venture capitalists.  If Rock
will put money in, others will follow.

						John Nagle
