Newsgroups: comp.ai.philosophy
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!cs.utexas.edu!news.unt.edu!hermes.oc.com!internet.spss.com!markrose
From: markrose@spss.com (Mark Rosenfelder)
Subject: AIs in traffic (was: Minsky's new article)
Message-ID: <Cz4q18.KCG@spss.com>
Sender: news@spss.com
Organization: SPSS Inc
References: <39vo5u$3fv@trog.dra.hmg.gb> <3a0rpq$6dl@cantaloupe.srv.cs.cmu.edu>
Date: Sat, 12 Nov 1994 00:43:55 GMT
Lines: 20

In article <3a0rpq$6dl@cantaloupe.srv.cs.cmu.edu>,
Hans Moravec <hpm@cs.cmu.edu> wrote:
>And, in the more ant-like realm of behavior, last year CMU's Navlab II
>drove itself 150 km from Pittsburgh to the Ohio border at full legal
>speeds, in traffic, driving on main roads, highways of various widths,
>taking on and off ramps.  It was controlled overall by a program that
>used a road map and a GPS system, but it stayed on the roads using a
>bank of neural nets, each for a different kind of road, each trained
>to imitate the steering commands of a human driver, given a low res
>road image.  The system switched automatically between the nets based
>on their outputs, which was a nice bell curve when a net is operating
>within its training range, and deteriorates from that shape otherwise.

Interesting; but sounds a bit dangerous.  How did the system read
road signs?  How did it deal with people crossing the road?  How about
road construction, with its special signs, flagmen, unusual road
surfaces, etc.?  How about rain, snow, ice, night, and fog?  Did it respond
correctly to flashing Mars lights, car horns, and the sound of a flat tire?
Did it correctly distinguish between policemen and hitchhikers waving at it?  

