Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!wuarchive!udel!rochester!pt.cs.cmu.edu!o.gp.cs.cmu.edu!andrew.cmu.edu!hm02+
From: hm02+@andrew.cmu.edu (Hans P. Moravec)
Newsgroups: comp.ai
Subject: Re: Shooting pigeons
Message-ID: <0buf84m00WBN43NUwc@andrew.cmu.edu>
Date: 23 Mar 91 01:40:20 GMT
Organization: Carnegie Mellon, Pittsburgh, PA
Lines: 22

> loren@ingrid.llnl.gov (Loren Petrich) writes:

>	You might want to check out alternatives to traditional AI,
> such as Neural Nets, Fuzzy Logic, and the kind of robotics that Brooks
> has used in his mechanical insects. I make these plugs because these
> are techniques that have produced _results_, and that's what one is
> supposed to get, right?

Expert systems are not used much in robot vision systems, but "traditional"
methods such as high pass, low pass and specialized filters,
correlation, frequency transforms, and specialized operators are in
practical use in industrial robots and such weapons as the Tomahawk
cruise missile and fire and forget TV guided bombs and missles, where
millions of pixels of image data are
processed per second.  The neo-cybernetic techniques have
yet to prove themselves in such computationally difficult domains.  Fuzzy
logic, neural nets and reactive organizations have so far been used only in
relatively simple contexts, like camera autofocus and train speed
control, and a few mobile robots, where the input is limited to a few
dozen measurements at a time.
 
                                                -- Hans Moravec